1 00:00:06,760 --> 00:00:08,520 >> WELCOME TO THE 2 00:00:08,520 --> 00:00:11,520 ANNUAL KUAN-TEH JEANG 3 00:00:11,520 --> 00:00:11,880 LECTURE. 4 00:00:11,880 --> 00:00:17,120 MY I AM THE ACTING DIRECTOR OF E 5 00:00:17,120 --> 00:00:19,160 NIH OFFICE OF EQUITY, DIVERSITY 6 00:00:19,160 --> 00:00:20,840 AND INCLUSION. 7 00:00:20,840 --> 00:00:22,440 EDI IS PLEASED TO PARTNER WITH 8 00:00:22,440 --> 00:00:25,560 THE OFFICE OF INTRAMURAL 9 00:00:25,560 --> 00:00:26,800 RESEARCH TO HOST THIS VERY 10 00:00:26,800 --> 00:00:29,720 IMPORTANT LECTURE AS WE 11 00:00:29,720 --> 00:00:31,960 CELEBRATE ASIAN AMERICAN, NATIVE 12 00:00:31,960 --> 00:00:33,840 HAWAIIAN AND PACIFIC ISLANDER 13 00:00:33,840 --> 00:00:34,200 HERITAGE MONTH. 14 00:00:34,200 --> 00:00:36,520 THANK YOU FOR TAKING TIME TO 15 00:00:36,520 --> 00:00:38,120 PAUSE AND HONOR THE LIFE AND 16 00:00:38,120 --> 00:00:40,480 LEGACY OF AN INDIVIDUAL WHO 17 00:00:40,480 --> 00:00:42,880 MEANT SO MUCH TO SO MANY. 18 00:00:42,880 --> 00:00:45,320 I'M HONORED TO HAVE PERSONALLY 19 00:00:45,320 --> 00:00:48,800 KNOWN AND WORKED WITH K.T. 20 00:00:48,800 --> 00:00:50,520 THOSE WHO KNEW K.T. KNEW HE WAS 21 00:00:50,520 --> 00:00:52,440 A BRILLIANT SCIENTIST WHO WAS 22 00:00:52,440 --> 00:00:54,480 PASSIONATE ABOUT DIVERSITY, 23 00:00:54,480 --> 00:00:56,160 EQUITY, INCLUSION AND 24 00:00:56,160 --> 00:00:58,240 ACCESSIBILITY. 25 00:00:58,240 --> 00:01:00,080 I HAD THE PLEASURE OF WORKING 26 00:01:00,080 --> 00:01:02,480 WITH K.T. IN THE DEIA SPACE AND 27 00:01:02,480 --> 00:01:05,080 TO HAVE EXPERIENCED HIS STAUNCH 28 00:01:05,080 --> 00:01:06,920 ADVOCACY FOR THE CAREER 29 00:01:06,920 --> 00:01:08,320 DEVELOPMENT OF ASIAN AMERICAN 30 00:01:08,320 --> 00:01:09,600 SCIENTISTS. 31 00:01:09,600 --> 00:01:11,760 I KNOW K.T. WOULD BE PROUD OF 32 00:01:11,760 --> 00:01:13,560 THE FEDERAL GOVERNMENT'S CURRENT 33 00:01:13,560 --> 00:01:16,640 FOCUS ON DEIA, AND HE WOULD 34 00:01:16,640 --> 00:01:19,120 SURELY BE ACTIVELY ENGAGED IN 35 00:01:19,120 --> 00:01:22,640 NIH'S EFFORT TO DEVELOP ITS 36 00:01:22,640 --> 00:01:25,240 FIRST DEIA STRATEGIC PLAN. 37 00:01:25,240 --> 00:01:28,680 I ENCOURAGE YOU TO VISIT EDI'S 38 00:01:28,680 --> 00:01:30,720 WEBSITE TO TAKE PART IN THE 39 00:01:30,720 --> 00:01:32,360 CURRENT ASIAN AMERICAN NATIVE 40 00:01:32,360 --> 00:01:34,440 HAWAIIAN AND PACIFIC ISLANDER 41 00:01:34,440 --> 00:01:35,480 HERITAGE MONTH CAMPAIGN. 42 00:01:35,480 --> 00:01:38,320 THIS YEAR'S THEME IS "WE'RE NOT 43 00:01:38,320 --> 00:01:40,600 A MONO LITH," AND IT FOCUSES ON 44 00:01:40,600 --> 00:01:42,440 THE VASTLY DIVERSE LANGUAGES, 45 00:01:42,440 --> 00:01:44,360 CULTURES, IDENTITIES, RELIGIONS, 46 00:01:44,360 --> 00:01:47,840 AND CUSTOMS FROM THE MAIN 47 00:01:47,840 --> 00:01:55,920 COUNTRIES UNDER THE AANHPI 48 00:01:55,920 --> 00:01:56,440 UMBRELLA. 49 00:01:56,440 --> 00:01:58,560 LET HIS LIFE'S WORK INSPIRE US 50 00:01:58,560 --> 00:02:00,240 TO STRIVE FOR PRO TETIONAL 51 00:02:00,240 --> 00:02:01,440 EXCELLENCE, RAISE OUR VOICES TO 52 00:02:01,440 --> 00:02:03,520 DRAW ATTENTION TO INEQUITIES, 53 00:02:03,520 --> 00:02:05,880 AND COMMIT TO FINDING SOLUTIONS. 54 00:02:05,880 --> 00:02:09,120 NOW WE WILL HEAR FROM DR. ROLAND 55 00:02:09,120 --> 00:02:11,840 OWENS, DIRECTOR OF RESEARCH 56 00:02:11,840 --> 00:02:13,040 WORKFORCE DIVERSITY IN THE 57 00:02:13,040 --> 00:02:14,680 OFFICE OF INTRAMURAL RESEARCH, 58 00:02:14,680 --> 00:02:17,560 AND HE WILL PAY TRIBUTE TO 59 00:02:17,560 --> 00:02:18,160 DR. JEANG AND ACKNOWLEDGE HIS 60 00:02:18,160 --> 00:02:23,200 FAMILY. 61 00:02:23,200 --> 00:02:24,360 >> THANK YOU. 62 00:02:24,360 --> 00:02:29,440 SO I KNEW TEH IN TWO CAPACITIES. 63 00:02:29,440 --> 00:02:31,600 BEFORE I TOOK OVER MY CURRENT 64 00:02:31,600 --> 00:02:33,680 POSITION, I WAS A VIROLOGIST IN 65 00:02:33,680 --> 00:02:36,920 THE DIABETES INSTITUTE AND SO I 66 00:02:36,920 --> 00:02:39,440 KNEW HIM AS PART OF THE VIROLOGY 67 00:02:39,440 --> 00:02:39,680 COMMUNITY. 68 00:02:39,680 --> 00:02:41,680 AND THEN WHEN I TOOK OVER AS 69 00:02:41,680 --> 00:02:44,920 DIRECTOR OF RESEARCH WORKFORCE 70 00:02:44,920 --> 00:02:46,760 DEVELOPMENT, WE ENGAGED 71 00:02:46,760 --> 00:02:48,600 FREQUENTLY IN ISSUES OF 72 00:02:48,600 --> 00:02:50,280 DIVERSITY AND INCLUSION. 73 00:02:50,280 --> 00:02:54,880 SO I'D LIKE TO SHARE SOME OF MY 74 00:02:54,880 --> 00:03:04,840 REMEMBRANCES OF DR. JEANG. 75 00:03:04,840 --> 00:03:11,120 HE WAS BORN IN 1958 IN TAIWAN, 76 00:03:11,120 --> 00:03:13,680 AND THEN MOVED TO LIBYA TO SPEND 77 00:03:13,680 --> 00:03:15,720 MOST OF HIS CHILDHOOD. 78 00:03:15,720 --> 00:03:17,720 THEN HE CAME TO THE UNITED 79 00:03:17,720 --> 00:03:21,520 STATES IN 1970. 80 00:03:21,520 --> 00:03:23,640 THIS IS ACTUALLY MORE COMMON A 81 00:03:23,640 --> 00:03:25,840 STORY AT NIH THAN YOU WOULD 82 00:03:25,840 --> 00:03:28,320 THINK OF PEOPLE BEING WORLD 83 00:03:28,320 --> 00:03:30,960 TRAVELERS TO FINALLY GET TO NIH. 84 00:03:30,960 --> 00:03:35,400 HE ATTENDED THE MASSACHUSETTS 85 00:03:35,400 --> 00:03:36,000 INSTITUTE OF TECHNOLOGY, AND 86 00:03:36,000 --> 00:03:39,400 THEN EARNED HIS MD PH.D. FROM 87 00:03:39,400 --> 00:03:44,680 JOHNS HOPKINS AT AGE 25. 88 00:03:44,680 --> 00:03:46,880 AND STUDIED REGULATION OF GENE 89 00:03:46,880 --> 00:03:48,760 EXPRESSION IN CYTOMEGLOVIRUS 90 00:03:48,760 --> 00:03:52,720 WITH GARY HAYWARD. 91 00:03:52,720 --> 00:03:55,120 THEN HE CAME TO THE NIH IN 1985, 92 00:03:55,120 --> 00:03:59,160 ACTUALLY THE SAME YEAR I DID. 93 00:03:59,160 --> 00:04:00,600 AND DID A POSTDOCTORAL 94 00:04:00,600 --> 00:04:02,960 FELLOWSHIP WITH THE LATE 95 00:04:02,960 --> 00:04:05,200 DR. GEORGE KORY IN THE NATIONAL 96 00:04:05,200 --> 00:04:06,040 CANCER INSTITUTE. 97 00:04:06,040 --> 00:04:07,760 HE THEN MOVED TO THE NATIONAL 98 00:04:07,760 --> 00:04:08,800 INSTITUTE OF ALLERGY AND 99 00:04:08,800 --> 00:04:10,240 INFECTIOUS DISEASES AND BECAME 100 00:04:10,240 --> 00:04:14,440 CHIEF OF THE MOLECULAR VIROLOGY 101 00:04:14,440 --> 00:04:16,640 SECTION IN THE LABORATORY OF 102 00:04:16,640 --> 00:04:21,400 MOLECULAR MICROBIOLOGY, WE HE WE 103 00:04:21,400 --> 00:04:24,120 STUDIED HUMAN IMMUNODEFICIENCY 104 00:04:24,120 --> 00:04:28,600 VIRUS HIV-1 AND A RELATED VIRUS, 105 00:04:28,600 --> 00:04:29,120 HTLV-1. 106 00:04:29,120 --> 00:04:31,200 HE WENT ON TO BECOME 107 00:04:31,200 --> 00:04:32,720 EDITOR-IN-CHIEF OF RETRO 108 00:04:32,720 --> 00:04:37,360 VIROLOGY, AND HE WAS A LEADER 109 00:04:37,360 --> 00:04:38,560 WITHIN THE VIROLOGY INTEREST 110 00:04:38,560 --> 00:04:38,800 GROUP. 111 00:04:38,800 --> 00:04:40,440 HE WAS ALSO A LEADER WITHIN THE 112 00:04:40,440 --> 00:04:41,920 SEW 113 00:04:41,920 --> 00:04:43,360 SOCIETY OF CHAI KNEE 114 00:04:43,360 --> 00:04:47,000 BIOSCIENTISTS IN AMERICA AND AN 115 00:04:47,000 --> 00:04:49,360 ADVOCATE FOR RACIAL MINORITY 116 00:04:49,360 --> 00:04:52,520 GROUPS AND PERSONS WITH 117 00:04:52,520 --> 00:04:55,520 DISABILITIES. 118 00:04:55,520 --> 00:04:56,520 JUST AS A PERSONAL ASIDE, WE 119 00:04:56,520 --> 00:04:57,840 USED TO PARK IN THE SAME PARKING 120 00:04:57,840 --> 00:04:59,440 LOT FREQUENTLY AND, YOU KNOW, 121 00:04:59,440 --> 00:05:01,000 ABOUT ONCE A MONTH, HE WOULD 122 00:05:01,000 --> 00:05:03,440 COME UP TO ME AND SAY, HEY, 123 00:05:03,440 --> 00:05:05,680 ROLAND, OKAY, SO HAVE YOU MADE 124 00:05:05,680 --> 00:05:06,680 ANY PROGRESS THIS MONTH? 125 00:05:06,680 --> 00:05:09,600 AND IT'S REALLY IMPORTANT TO 126 00:05:09,600 --> 00:05:13,400 HAVE PEOPLE LIKE THAT WHO ARE 127 00:05:13,400 --> 00:05:15,320 WILLING TO KEEP TALKING ABOUT 128 00:05:15,320 --> 00:05:16,720 THINGS AND KEEP HOLDING OUR FEET 129 00:05:16,720 --> 00:05:21,080 TO THE FIRE. 130 00:05:21,080 --> 00:05:22,560 AND WE CELEBRATE HIS LIFE AND 131 00:05:22,560 --> 00:05:27,880 LEGACY IN MANY WAYS. 132 00:05:27,880 --> 00:05:29,080 ONE, THIS LECTURESHIP. 133 00:05:29,080 --> 00:05:33,280 WE ALSO HAVE A SPECIAL 134 00:05:33,280 --> 00:05:35,200 RECOGNITION FOR FAIR AWARD 135 00:05:35,200 --> 00:05:36,960 WINNERS WHO SPECIALIZE IN 136 00:05:36,960 --> 00:05:38,920 VIROLOGY, AND IF YOU'VE EVER 137 00:05:38,920 --> 00:05:46,720 BEEN BETWEEN BUILDING 31 AND -- 138 00:05:46,720 --> 00:05:49,840 OUTSIDE BUILDING 31, THERE'S A 139 00:05:49,840 --> 00:05:52,160 LITTLE MEMORIAL TO DR. JEANG. 140 00:05:52,160 --> 00:05:54,280 WHICH IS SHOWN HERE. 141 00:05:54,280 --> 00:05:56,560 AND THIS ALSO REMINDS ME TO 142 00:05:56,560 --> 00:05:59,040 RECOGNIZE THAT WE DO HAVE 143 00:05:59,040 --> 00:06:02,160 MEMBERS OF THE FAMILY PRESENT ON 144 00:06:02,160 --> 00:06:07,360 THIS -- AT THIS EVENT. 145 00:06:07,360 --> 00:06:08,920 I WOULD LIKE TO THANK HIS WIFE, 146 00:06:08,920 --> 00:06:10,520 HIS BROTHERS, HIS CHILDREN, AND 147 00:06:10,520 --> 00:06:12,640 ALL THE FAMILY MEMBERS FOR 148 00:06:12,640 --> 00:06:17,680 SHARING HIS LIFE WITH US. 149 00:06:17,680 --> 00:06:22,600 NOW I'LL HAND IT OVER TO 150 00:06:22,600 --> 00:06:23,800 DR. MICHAEL GOTTESMAN, THE 151 00:06:23,800 --> 00:06:25,640 DEPUTY DIRECTOR FOR INTRAMURAL 152 00:06:25,640 --> 00:06:26,840 RESEARCH, WHO WILL INTRODUCE 153 00:06:26,840 --> 00:06:28,080 TODAY'S SPEAKER. 154 00:06:28,080 --> 00:06:29,800 >> THANK YOU, ROLAND. 155 00:06:29,800 --> 00:06:30,880 GOOD AFTERNOON, EVERYONE. 156 00:06:30,880 --> 00:06:36,760 LET ME ADD MY WELCOME TO THIS, 157 00:06:36,760 --> 00:06:37,640 THE 2022 KUAN-TEH JEANG MEMORIAL 158 00:06:37,640 --> 00:06:38,560 LECTURE. 159 00:06:38,560 --> 00:06:40,560 I'M MICHAEL GOTTESMAN, AS YOU'VE 160 00:06:40,560 --> 00:06:41,520 HEARD, DEPUTY DIRECTOR FOR 161 00:06:41,520 --> 00:06:43,600 INTRAMURAL RESEARCH AT THE NIH. 162 00:06:43,600 --> 00:06:47,640 AS YOU ALSO JUST HEARD, TEH 163 00:06:47,640 --> 00:06:49,040 JEANG, ASIDE FROM BEING A 164 00:06:49,040 --> 00:06:50,640 REMARKABLE SCIENTIST, WAS A 165 00:06:50,640 --> 00:06:51,960 STAUNCH ADVOCATE FOR DIVERSITY 166 00:06:51,960 --> 00:06:53,800 AND INCLUSION OF MINORITIES. 167 00:06:53,800 --> 00:06:55,640 HE SUBSCRIBED TO THE MOTTO THAT 168 00:06:55,640 --> 00:06:58,080 TALENT KNOWS NO BORDERS. 169 00:06:58,080 --> 00:06:59,440 FOR TEH, THIS APPLIED NOT ONLY 170 00:06:59,440 --> 00:07:01,360 TO THE BIOMEDICAL RESEARCH 171 00:07:01,360 --> 00:07:02,960 WORKFORCE AS A WHOLE, BUT TO 172 00:07:02,960 --> 00:07:05,080 LEADERSHIP POSITIONS TOO. 173 00:07:05,080 --> 00:07:07,080 AND WE CERTAINLY HAVE FOUND 174 00:07:07,080 --> 00:07:08,480 REMARKABLE TALENT IN TODAY'S 175 00:07:08,480 --> 00:07:10,040 SPEAKER, DR. MICHAEL CHIANG, WHO 176 00:07:10,040 --> 00:07:12,280 IS ONLY THE SECOND CHINESE 177 00:07:12,280 --> 00:07:13,520 AMERICAN TO DIRECT AN NIH 178 00:07:13,520 --> 00:07:15,120 INSTITUTE. 179 00:07:15,120 --> 00:07:16,320 DR. CHIANG IS THE DIRECTOR OF 180 00:07:16,320 --> 00:07:18,600 THE NATIONAL EYE INSTITUTE, 181 00:07:18,600 --> 00:07:19,800 RECRUITED JUST TWO YEARS AGO, 182 00:07:19,800 --> 00:07:24,080 JUST BEFORE THE PANDEMIC IN 183 00:07:24,080 --> 00:07:24,280 2020. 184 00:07:24,280 --> 00:07:25,840 DR. CHIANG IS AN OPHTHALMOLOGIST 185 00:07:25,840 --> 00:07:27,000 WITH EXTENSIVE EXPERIENCE AS A 186 00:07:27,000 --> 00:07:28,600 CLINICIAN, RESEARCHER, AND 187 00:07:28,600 --> 00:07:29,480 EDUCATOR. 188 00:07:29,480 --> 00:07:31,560 HIS RESEARCH HAS INVOLVED 189 00:07:31,560 --> 00:07:34,040 TELEMEDICINE AND ARTIFICIAL 190 00:07:34,040 --> 00:07:36,680 INTELLIGENCE FOR DIAGNOSIS OF 191 00:07:36,680 --> 00:07:38,760 RETINOPATHY OF PREMATURITY AND 192 00:07:38,760 --> 00:07:40,160 OTHER OPHTHALMIC DISEASES. 193 00:07:40,160 --> 00:07:41,760 IMPLEMENTATION AND EVALUATION OF 194 00:07:41,760 --> 00:07:43,200 ELECTRONIC HEALTH RECORD 195 00:07:43,200 --> 00:07:44,760 SYSTEMS, MODELING OF CLINICAL 196 00:07:44,760 --> 00:07:48,520 WORKFLOW, AND DATA ANALYTICS. 197 00:07:48,520 --> 00:07:49,880 THESE TALENTS WHICH ARE CUTTING 198 00:07:49,880 --> 00:07:52,080 EDGE HAVE REALLY IMPROVED THE 199 00:07:52,080 --> 00:07:53,240 ENVIRONMENT AT THE NIH 200 00:07:53,240 --> 00:07:54,760 SCIENTIFICALLY. 201 00:07:54,760 --> 00:07:56,320 DR. CHIANG EARNED HIS 202 00:07:56,320 --> 00:07:56,880 UNDERGRADUATE DEGREE IN 203 00:07:56,880 --> 00:07:59,840 ELECTRICAL ENGINEERING AND 204 00:07:59,840 --> 00:08:00,320 BIOLOGY FROM STANFORD 205 00:08:00,320 --> 00:08:00,960 UNIVERSITY. 206 00:08:00,960 --> 00:08:02,280 HIS MASTER'S DEGREE IN 207 00:08:02,280 --> 00:08:04,440 BIOMEDICAL INFORMATICS FROM 208 00:08:04,440 --> 00:08:05,520 COLUMBIA UNIVERSITY COLLEGE OF 209 00:08:05,520 --> 00:08:08,160 PHYSICIANS AND SURGEONS, AND HIS 210 00:08:08,160 --> 00:08:10,120 M.D. AND ADDITIONAL MASTERS IN 211 00:08:10,120 --> 00:08:11,680 MEDICAL SCIENCE FROM THE HARVARD 212 00:08:11,680 --> 00:08:14,160 MEDICAL SCHOOL AND THE HARVARD 213 00:08:14,160 --> 00:08:15,080 M.I.T. DIVISION OF HEALTH 214 00:08:15,080 --> 00:08:18,120 SCIENCES AND TECHNOLOGY. 215 00:08:18,120 --> 00:08:20,640 PRIOR TO BEING NAMED DIRECTOR OF 216 00:08:20,640 --> 00:08:21,600 THE NATIONAL EYE INSTITUTE, 217 00:08:21,600 --> 00:08:22,800 DR. CHIANG WAS THE KNOWLES 218 00:08:22,800 --> 00:08:24,960 PROFESSOR OF OPHTHALMOLOGY AND 219 00:08:24,960 --> 00:08:27,560 MEDICAL INFORMATICS AND CLINICAL 220 00:08:27,560 --> 00:08:29,040 EPIDEMIOLOGY AT OREGON HEALTH 221 00:08:29,040 --> 00:08:30,840 AND SCIENCE UNIVERSITY, ALSO 222 00:08:30,840 --> 00:08:34,120 KNOWN AS OHSU, IN PORTLAND, 223 00:08:34,120 --> 00:08:36,080 OREGON, AND ASSOCIATE DIRECTOR 224 00:08:36,080 --> 00:08:38,880 OF THE OHSU CASEY EYE INSTITUTE. 225 00:08:38,880 --> 00:08:40,520 HE WAS A PRINCIPAL INVESTIGATOR 226 00:08:40,520 --> 00:08:43,280 ON SEVERAL NIH GRANTS SINCE 227 00:08:43,280 --> 00:08:45,760 2003, AND HE AND HIS RESEARCH 228 00:08:45,760 --> 00:08:48,080 GROUP PUBLISHED MORE THAN 200 229 00:08:48,080 --> 00:08:49,680 PEER REVIEWED JOURNAL PAPERS. 230 00:08:49,680 --> 00:08:50,920 HIS CLINICAL PRACTICE IS FOCUSED 231 00:08:50,920 --> 00:08:53,880 ON PEDIATRIC OPHTHALMOLOGY AND 232 00:08:53,880 --> 00:08:55,360 ADULT STRABISMUS. 233 00:08:55,360 --> 00:08:57,720 AS NEI DIRECTOR, DR. CHIANG 234 00:08:57,720 --> 00:09:00,960 OVERSEES AN ANNUAL BUDGET OF 235 00:09:00,960 --> 00:09:01,600 APPROXIMATELY $825 MILLION, THE 236 00:09:01,600 --> 00:09:03,440 MAJORITY OF WHICH SUPPORTS 237 00:09:03,440 --> 00:09:05,160 VISION RESEARCH THROUGH NEARLY 238 00:09:05,160 --> 00:09:06,880 1600 RESEARCH GRANTS AND 239 00:09:06,880 --> 00:09:08,520 TRAINING AWARDS MADE TO 240 00:09:08,520 --> 00:09:10,480 SCIENTISTS AT MORE THAN 241 00:09:10,480 --> 00:09:11,360 250 MEDICAL CENTERS, 242 00:09:11,360 --> 00:09:13,120 UNIVERSITIES AND OTHER 243 00:09:13,120 --> 00:09:14,240 INSTITUTIONS. 244 00:09:14,240 --> 00:09:16,120 NEI, OF COURSE, ALSO HAS A 245 00:09:16,120 --> 00:09:17,520 SIGNIFICANT AND SUCCESSFUL 246 00:09:17,520 --> 00:09:19,680 INTRAMURAL RESEARCH PROGRAM THAT 247 00:09:19,680 --> 00:09:21,440 INCLUDES CLINICAL RESEARCH. 248 00:09:21,440 --> 00:09:25,440 THE TITLE OF DR. CHIANG'S TALK 249 00:09:25,440 --> 00:09:30,360 TODAY IS ARTIFICIAL 250 00:09:30,360 --> 00:09:31,960 INTELLIGENCE, OPHTHALMOLOGY AND 251 00:09:31,960 --> 00:09:33,920 BEING AN ASIAN AMERICAN CLINICAL 252 00:09:33,920 --> 00:09:35,440 SCIENTIST. 253 00:09:35,440 --> 00:09:36,800 MICHAEL, I'M THRILLED YOU 254 00:09:36,800 --> 00:09:38,440 ACCEPTED THE OFFER TO DELIVER 255 00:09:38,440 --> 00:09:40,200 THIS YEAR'S KUAN-TEH JEANG 256 00:09:40,200 --> 00:09:41,480 MEMORIAL LECTURE. 257 00:09:41,480 --> 00:09:43,640 TEH WAS A GOOD FRIEND AND 258 00:09:43,640 --> 00:09:44,840 COLLEAGUE VERY SPECIAL TO ME, 259 00:09:44,840 --> 00:09:46,040 AND HE, TOO, WOULD BE THRILLED 260 00:09:46,040 --> 00:09:48,240 TO KNOW SOMEONE OF YOUR CALIBER 261 00:09:48,240 --> 00:09:49,760 HAS BEEN RECRUITED TO LEAD THE 262 00:09:49,760 --> 00:09:50,520 NEI. 263 00:09:50,520 --> 00:09:51,560 MICHAEL, YOU'RE ON. 264 00:09:51,560 --> 00:09:53,200 >> MICHAEL, THANK YOU VERY MUCH 265 00:09:53,200 --> 00:09:55,000 FOR THE INTRODUCTION. 266 00:09:55,000 --> 00:09:59,120 AND TO ROLAND AND SHELMA, THANK 267 00:09:59,120 --> 00:10:00,480 YOU FOR YOUR WORK IN PUTTING 268 00:10:00,480 --> 00:10:01,120 THIS EVENT TOGETHER. 269 00:10:01,120 --> 00:10:02,760 YOU KNOW, I AM ENORMOUSLY 270 00:10:02,760 --> 00:10:05,360 PRIVILEGED TO BE ABLE TO GIVE 271 00:10:05,360 --> 00:10:06,920 THIS KUAN-TEH JEANG MEMORIAL 272 00:10:06,920 --> 00:10:10,680 LECTURE BASICALLY IN RECOGNITION 273 00:10:10,680 --> 00:10:12,960 OF ASIAN AMERICAN NATIVE 274 00:10:12,960 --> 00:10:14,000 HAWAIIAN PACIFIC ISLANDER 275 00:10:14,000 --> 00:10:14,600 HERITAGE MONTH IN MAY. 276 00:10:14,600 --> 00:10:17,200 I REALLY HOPE TO BE ABLE TO DO 277 00:10:17,200 --> 00:10:19,680 JUSTICE TO DR. JEANG'S LEGACY 278 00:10:19,680 --> 00:10:20,400 HERE. 279 00:10:20,400 --> 00:10:21,600 I NEVER HAD AN OPPORTUNITY TO 280 00:10:21,600 --> 00:10:25,680 MEET HIM BECAUSE I CAME ONLY IN 281 00:10:25,680 --> 00:10:26,520 NOVEMBER 2020. 282 00:10:26,520 --> 00:10:27,880 AND -- BUT WHAT I HAD A CHANCE 283 00:10:27,880 --> 00:10:32,120 TO DO IS TO BE IN TOUCH WITH HIS 284 00:10:32,120 --> 00:10:34,280 WIFE DIANE AND I SPOKE WITH HIS 285 00:10:34,280 --> 00:10:36,680 BROTHER MING JUST TO LEARN A 286 00:10:36,680 --> 00:10:37,760 LITTLE ABOUT HIM AND ABOUT THE 287 00:10:37,760 --> 00:10:38,560 FAMILY AND I THOUGHT I WOULD 288 00:10:38,560 --> 00:10:40,080 SHARE SOME OF WHAT I LEARNED. 289 00:10:40,080 --> 00:10:42,800 JUST TO BUILD ON WHAT SOME OF 290 00:10:42,800 --> 00:10:46,760 ROLAND AND MICHAEL MENTIONED, HE 291 00:10:46,760 --> 00:10:48,400 WAS BORN IN TAIWAN AND WAS THE 292 00:10:48,400 --> 00:10:49,800 YOUNGEST OUT OF THREE BROTHERS. 293 00:10:49,800 --> 00:10:51,120 BECAUSE HIS FATHER WAS A CIVIL 294 00:10:51,120 --> 00:10:52,560 ENGINEER, HE WAS SENT BY THE 295 00:10:52,560 --> 00:10:53,400 GOVERNMENT AND THAT'S HOW THE 296 00:10:53,400 --> 00:10:55,920 FAMILY ENDED UP IN AFRICA. 297 00:10:55,920 --> 00:10:59,080 AND HE WAS KNOWN BY K.T., OR 298 00:10:59,080 --> 00:11:02,280 SIMPLY T., AND K.T. WAS EDUCATED 299 00:11:02,280 --> 00:11:05,760 IN THE BRITISH SCHOOLS. 300 00:11:05,760 --> 00:11:06,640 MING MENTIONED THE OLDER 301 00:11:06,640 --> 00:11:08,000 BROTHERS WERE SENT TO BOARDING 302 00:11:08,000 --> 00:11:09,920 SCHOOLS IN IRELAND, SO THE OLDER 303 00:11:09,920 --> 00:11:12,000 BROTHERS HAD IRISH ACCENTS AND 304 00:11:12,000 --> 00:11:14,360 K.T. GREW UP WITH A BRITISH 305 00:11:14,360 --> 00:11:15,160 ACCENT. 306 00:11:15,160 --> 00:11:17,400 BUT THE FAMILY MOVED TO THE U.S. 307 00:11:17,400 --> 00:11:19,600 WHEN K.T. WAS 12, AND HE REALLY 308 00:11:19,600 --> 00:11:21,080 WENT TO SCHOOL IN ARIZONA AND 309 00:11:21,080 --> 00:11:21,760 THEN HOUSTON. 310 00:11:21,760 --> 00:11:23,760 AND WHEN HE WAS IN HIGH SCHOOL, 311 00:11:23,760 --> 00:11:25,840 SAID, I WANT TO GO TO M.I.T., 312 00:11:25,840 --> 00:11:28,360 AND THE FAMILY REACTED BY 313 00:11:28,360 --> 00:11:29,560 WONDERING HOW ARE WE GOING TO 314 00:11:29,560 --> 00:11:31,600 PAY FOR M.I.T., IT'S VERY 315 00:11:31,600 --> 00:11:32,280 EXPENSIVE. 316 00:11:32,280 --> 00:11:33,800 SO IT WAS K.T. WHO LOOKED INTO 317 00:11:33,800 --> 00:11:34,960 THE FACT THAT YOU COULD GET 318 00:11:34,960 --> 00:11:36,080 SCHOLARSHIPS TO GO THERE, AND 319 00:11:36,080 --> 00:11:37,240 THAT'S HOW HE ENDED UP STUDYING 320 00:11:37,240 --> 00:11:43,160 AT M.I.T., WHERES WEIGH ON E 321 00:11:43,160 --> 00:11:44,080 DEBATE TEAM. 322 00:11:44,080 --> 00:11:46,360 HIS COACH EVENTUALLY BECAME 323 00:11:46,360 --> 00:11:47,240 SECRETARY OF THE TREASURY AND 324 00:11:47,240 --> 00:11:48,000 PRESIDENT OF HARVARD. 325 00:11:48,000 --> 00:11:48,800 THIS IS A PICTURE OF THE FAMILY 326 00:11:48,800 --> 00:11:50,160 I RECEIVED. 327 00:11:50,160 --> 00:11:51,040 HERE'S K.T. AND HERE'S A PICTURE 328 00:11:51,040 --> 00:11:52,200 OF HIM BACK WHEN HE WAS A 329 00:11:52,200 --> 00:11:56,120 STUDENT AT M.I.T. 330 00:11:56,120 --> 00:11:58,400 SO AS ROLAND HAD MENTIONED, WHEN 331 00:11:58,400 --> 00:12:00,720 HE WAS 18, HE DECIDED THAT HE 332 00:12:00,720 --> 00:12:03,640 WANTED TO ENROLL IN AN M.D. PH 333 00:12:03,640 --> 00:12:05,480 D. PROGRAM AND WAS ACCEPTED AT 334 00:12:05,480 --> 00:12:06,880 JOHNS HOPKINS, SO HE GOT 335 00:12:06,880 --> 00:12:08,200 HIS DEGREES THERE, DID AN 336 00:12:08,200 --> 00:12:09,000 INTERNSHIP IN MEDICINE AND 337 00:12:09,000 --> 00:12:10,320 DECIDED THAT HE REALLY WANTED TO 338 00:12:10,320 --> 00:12:11,640 DO A RESEARCH FELLOWSHIP AND 339 00:12:11,640 --> 00:12:14,320 BUILD A CAREER IN ACADEMICS. 340 00:12:14,320 --> 00:12:16,520 SO HE CAME HERE TO NIH. 341 00:12:16,520 --> 00:12:18,400 AND EVENTUALLY HE BECAME CHIEF 342 00:12:18,400 --> 00:12:20,560 AS ROLAND SAID OF THE MOLECULAR 343 00:12:20,560 --> 00:12:22,880 VIROLOGY SECTION OF THE NIAID 344 00:12:22,880 --> 00:12:24,400 LABORATORY OF MEDICAL 345 00:12:24,400 --> 00:12:24,920 MICROBIOLOGY. 346 00:12:24,920 --> 00:12:28,480 HE BECAME THE FOUNDING EDITOR OF 347 00:12:28,480 --> 00:12:29,600 RETROVIROLOGY, AND WAS A 348 00:12:29,600 --> 00:12:31,000 PROPONENT OF OPEN ACCESS 349 00:12:31,000 --> 00:12:32,080 PUBLICATION LONG BEFORE IT 350 00:12:32,080 --> 00:12:34,040 BECAME HIP, AND POPULAR, AND WAS 351 00:12:34,040 --> 00:12:37,800 A REAL ADVOCATE FOR CAREER 352 00:12:37,800 --> 00:12:39,680 DEVELOPMENT OF A YOUNGER 353 00:12:39,680 --> 00:12:40,600 SCIENTIST AND MID CAREER 354 00:12:40,600 --> 00:12:41,160 SCIENTISTS. 355 00:12:41,160 --> 00:12:42,480 THIS PICTURE I GOT FROM DIANE, 356 00:12:42,480 --> 00:12:44,320 HIS WIFE, WHO'S A VETERINARIAN 357 00:12:44,320 --> 00:12:46,000 AT THE FDA, AND THE TWO OF THEM 358 00:12:46,000 --> 00:12:47,800 MET AT JOHNS HOPKINS, AND HIS 359 00:12:47,800 --> 00:12:49,920 CHILDREN ARE NOW ADULTS. 360 00:12:49,920 --> 00:12:52,680 DAVID, JOHN AND DIANA. 361 00:12:52,680 --> 00:12:55,800 AND JUST GOING BEYOND THAT, HE 362 00:12:55,800 --> 00:12:57,880 WAS INTERESTED IN A LOT OF 363 00:12:57,880 --> 00:12:58,520 SERVICE ACTIVITIES. 364 00:12:58,520 --> 00:13:00,360 AS ROLAND MENTIONED, HE WAS 365 00:13:00,360 --> 00:13:01,960 PRESIDENT OF THE SOCIETY OF 366 00:13:01,960 --> 00:13:03,280 CHINESE BIOSCIENTISTS IN 367 00:13:03,280 --> 00:13:07,240 AMERICA, AND HERE'S A PICTURE OF 368 00:13:07,240 --> 00:13:08,560 THAT CREW, HE WAS AN EDITOR AT 369 00:13:08,560 --> 00:13:11,840 CELL AND BIOSCIENCE, WHICH IS 370 00:13:11,840 --> 00:13:14,360 THE SBCA'S JOURNAL AND ASSOCIATE 371 00:13:14,360 --> 00:13:15,360 EDITOR OF CANCER RESEARCH. 372 00:13:15,360 --> 00:13:19,320 HE WAS A MEMBER OF ACADEMIA 373 00:13:19,320 --> 00:13:20,640 SINICA, TAIWAN, ACADEMY OF 374 00:13:20,640 --> 00:13:21,000 SCIENCE. 375 00:13:21,000 --> 00:13:23,040 THIS IS HIS NAMEPLATE FROM THAT. 376 00:13:23,040 --> 00:13:25,040 THE THING THAT WE REALIZED IS 377 00:13:25,040 --> 00:13:27,640 THE CHINESE CHARACTER FOR THE 378 00:13:27,640 --> 00:13:29,040 NAME JEANG IS THE SAME AS MY 379 00:13:29,040 --> 00:13:29,720 LAST NAME. 380 00:13:29,720 --> 00:13:31,360 WE ROMANIZE IT DIFFERENTLY SO MY 381 00:13:31,360 --> 00:13:35,080 NAME IS SPELLED C, H, I, A, N, 382 00:13:35,080 --> 00:13:35,440 G. 383 00:13:35,440 --> 00:13:37,480 BUT THEN RECEIVED AWARDS AT 384 00:13:37,480 --> 00:13:38,440 JOHNS HOPKINS FOR PUBLIC 385 00:13:38,440 --> 00:13:38,760 SERVICE. 386 00:13:38,760 --> 00:13:40,760 ALSO THIS IS A PICTURE OF KT AND 387 00:13:40,760 --> 00:13:42,040 MICHAEL GOTTESMAN THAT WAS SHOWN 388 00:13:42,040 --> 00:13:45,680 BEFORE WHEN HE GAVE THE GEORGE 389 00:13:45,680 --> 00:13:47,080 KHOURY LECTURE IN 2012. 390 00:13:47,080 --> 00:13:49,040 AS SAID BEFORE, HE WAS A REAL 391 00:13:49,040 --> 00:13:49,920 ADVOCATE FOR LEADERSHIP 392 00:13:49,920 --> 00:13:52,520 OPPORTUNITIES FOR SCIENTISTS OF 393 00:13:52,520 --> 00:13:54,120 ASIAN DESCENT AND TESTIFIED 394 00:13:54,120 --> 00:13:56,720 BEFORE THE U.S. EQUAL EMPLOYMENT 395 00:13:56,720 --> 00:13:57,560 OPPORTUNITY COMMISSION IN 2008, 396 00:13:57,560 --> 00:14:00,920 AND EVENTUALLY GOT AN AWARD FROM 397 00:14:00,920 --> 00:14:05,280 THE NIH EEO, HE GOT THE NIH EEO 398 00:14:05,280 --> 00:14:09,120 AWARD POST HUMESLY POSTH. 399 00:14:09,120 --> 00:14:10,520 SO I THOUGHT A LOT ABOUT THIS 400 00:14:10,520 --> 00:14:11,560 AND IN TRYING TO PUT THIS 401 00:14:11,560 --> 00:14:13,400 TOGETHER TO HONOR DR. JEANG'S 402 00:14:13,400 --> 00:14:14,880 LEGACY, I THOUGHT I WOULD TALK 403 00:14:14,880 --> 00:14:17,040 ABOUT NUMBER ONE ARTIFICIAL 404 00:14:17,040 --> 00:14:18,400 INTELLIGENCE AND OPHTHALMOLOGY 405 00:14:18,400 --> 00:14:19,920 BECAUSE THAT'S WHAT I DO 406 00:14:19,920 --> 00:14:20,720 PROFESSIONALLY, BUT I ALSO 407 00:14:20,720 --> 00:14:22,200 THOUGHT I WOULD SPEND ABOUT HALF 408 00:14:22,200 --> 00:14:23,680 THE TIME TALKING ABOUT SOME OF 409 00:14:23,680 --> 00:14:26,480 MY OWN THOUGHTS ABOUT BEING AN 410 00:14:26,480 --> 00:14:27,240 ASIAN AMERICAN 411 00:14:27,240 --> 00:14:27,960 CLINICIAN-SCIENTIST, BECAUSE I 412 00:14:27,960 --> 00:14:29,960 FELT THAT THAT WOULD BE AN 413 00:14:29,960 --> 00:14:32,800 APPROPRIATE WAY TO REFLECT ON 414 00:14:32,800 --> 00:14:35,560 WHAT DR. JEANG'S LEGACY IS HERE. 415 00:14:35,560 --> 00:14:37,560 SO I THOUGHT I WOULD JUST START 416 00:14:37,560 --> 00:14:39,480 BY SAYING A LITTLE ABOUT WHAT MY 417 00:14:39,480 --> 00:14:42,000 BACKGROUND IS. 418 00:14:42,000 --> 00:14:44,120 SO LIKE K.T.'S FAMILY, MY 419 00:14:44,120 --> 00:14:45,400 PARENTS WERE FROM TAIWAN. 420 00:14:45,400 --> 00:14:47,840 THEY CAME TO THE U.S. IN THE 421 00:14:47,840 --> 00:14:50,320 1960s, AND ORIGINALLY SETTLED 422 00:14:50,320 --> 00:14:51,640 IN SOUTH CAROLINA AT A TIME WHEN 423 00:14:51,640 --> 00:14:53,200 THE SOUTH AND THE U.S. WAS 424 00:14:53,200 --> 00:14:54,320 SEGREGATED. 425 00:14:54,320 --> 00:14:56,560 SO IT WAS A VERY INTERESTING AND 426 00:14:56,560 --> 00:14:57,680 ODD INTRODUCTION TO THIS COUNTRY 427 00:14:57,680 --> 00:14:59,880 FOR THEM. 428 00:14:59,880 --> 00:15:01,400 WITHIN A FEW YEARS, THEY MOVED 429 00:15:01,400 --> 00:15:03,920 TO PITTSBURGH, PENNSYLVANIA, 430 00:15:03,920 --> 00:15:05,680 BECAUSE MY DAD WAS A GRADUATE 431 00:15:05,680 --> 00:15:07,440 STUDENT AT CARNEGIE MELLON 432 00:15:07,440 --> 00:15:08,280 UNIVERSITY AND THAT'S WHERE I 433 00:15:08,280 --> 00:15:09,920 WAS BORN, IN PITTSBURGH. 434 00:15:09,920 --> 00:15:11,680 SO WITHIN THE FIRST FEW YEARS OF 435 00:15:11,680 --> 00:15:13,800 MY LIFE, WE MOVED TO DETROIT, 436 00:15:13,800 --> 00:15:15,040 MICHIGAN, WHICH IS WHAT I WOULD 437 00:15:15,040 --> 00:15:17,000 CONSIDER MY HOMETOWN, AND THE 438 00:15:17,000 --> 00:15:20,600 REASON FOR THAT IS THAT MY 439 00:15:20,600 --> 00:15:23,480 FATHER GOT A JOB AT FORD MOTOR 440 00:15:23,480 --> 00:15:24,560 COMPANY, AND THE U.S. AUTO 441 00:15:24,560 --> 00:15:26,400 INDUSTRY IS ALL IN DETROIT, 442 00:15:26,400 --> 00:15:26,680 MICHIGAN. 443 00:15:26,680 --> 00:15:28,600 SO I WOULD SAY THAT I HAD A VERY 444 00:15:28,600 --> 00:15:32,480 CLASSIC MIDWESTERN UPBRINGING. 445 00:15:32,480 --> 00:15:33,720 AMERICAN CULTURE BUILT ON THE 446 00:15:33,720 --> 00:15:35,320 AUTOMOBILE AND IN MANY WAYS A 447 00:15:35,320 --> 00:15:38,360 REALLY WARM AND FAMILY-ORIENTED 448 00:15:38,360 --> 00:15:40,960 ATMOSPHERE IN THE MIDWEST. 449 00:15:40,960 --> 00:15:42,760 NOW, IT TURNS OUT THAT WHEN I 450 00:15:42,760 --> 00:15:48,640 WAS PROBABLY 12 YEARS OLD, I 451 00:15:48,640 --> 00:15:50,360 REMEMBER MY DAD COMING HOME A 452 00:15:50,360 --> 00:15:52,440 LITTLE BIT GRUMPY FROM WORK. 453 00:15:52,440 --> 00:15:53,960 IT TURNED OUT HE HAD A 454 00:15:53,960 --> 00:15:55,000 PERFORMANCE REVIEW, AND WHAT HIS 455 00:15:55,000 --> 00:15:56,400 SUPERVISOR SAID AT THE TIME IS 456 00:15:56,400 --> 00:15:59,440 THAT, YOU ARE REALLY AN 457 00:15:59,440 --> 00:16:00,760 OUTSTANDING ENGINEER BUT YOU 458 00:16:00,760 --> 00:16:01,880 HAVE TO IMPROVE YOUR 459 00:16:01,880 --> 00:16:02,880 COMMUNICATION IN ENGLISH IF 460 00:16:02,880 --> 00:16:04,520 YOU'RE GOING TO ADVANCE IN THE 461 00:16:04,520 --> 00:16:06,200 ORGANIZATION AT FORD MOTOR 462 00:16:06,200 --> 00:16:06,520 COMPANY. 463 00:16:06,520 --> 00:16:08,120 AND HE WAS REALLY FRUSTRATED 464 00:16:08,120 --> 00:16:09,240 ABOUT THAT. 465 00:16:09,240 --> 00:16:11,800 BUT WE DIDN'T HAVE THE SORT OF 466 00:16:11,800 --> 00:16:13,560 FAMILY WHERE, YOU KNOW, OUR 467 00:16:13,560 --> 00:16:15,360 PARENTS WOULD TALK TO US ABOUT 468 00:16:15,360 --> 00:16:17,000 THEIR PROBLEMS. 469 00:16:17,000 --> 00:16:18,600 SO IT TURNED OUT THAT PROBABLY A 470 00:16:18,600 --> 00:16:21,120 YEAR OR TWO LATER, I WAS IN THE 471 00:16:21,120 --> 00:16:23,320 FAMILY CAR, AND THERE WAS A 472 00:16:23,320 --> 00:16:25,040 CASSETTE TAPE HERE, AND I PUT 473 00:16:25,040 --> 00:16:28,040 THE CASSETTE TAPE INTO THE RADIO 474 00:16:28,040 --> 00:16:30,000 PLAYER, AND INSTEAD OF MUSIC, 475 00:16:30,000 --> 00:16:32,800 WHAT CAME ON WAS A VOICE THAT I 476 00:16:32,800 --> 00:16:34,280 RECOGNIZE AS WALTER CRONKITE'S 477 00:16:34,280 --> 00:16:34,920 VOICE. 478 00:16:34,920 --> 00:16:36,320 NOW, FOR THOSE OF YOU WHO DON'T 479 00:16:36,320 --> 00:16:38,080 KNOW, WALTER CRONKITE WAS 480 00:16:38,080 --> 00:16:39,760 PROBABLY THE MOST WELL-KNOWN 481 00:16:39,760 --> 00:16:41,480 U.S. NEWS BROADCASTER IN THE 482 00:16:41,480 --> 00:16:44,120 1970s AND 1980s, AND HE WAS 483 00:16:44,120 --> 00:16:45,720 FOR A WHILE CONSIDERED THE MOST 484 00:16:45,720 --> 00:16:51,000 TRUSTED PERSON IN AMERICA. 485 00:16:51,000 --> 00:16:52,600 I DIDN'T REALIZE THIS UNTIL 10 486 00:16:52,600 --> 00:16:54,560 YEARS LATER BUT THE REASON 487 00:16:54,560 --> 00:16:56,400 WALTER CRONKITE'S VOICE WAS ON 488 00:16:56,400 --> 00:16:57,280 THIS CASSETTE TAPE WAS BECAUSE 489 00:16:57,280 --> 00:17:00,160 MY FATHER WAS TRYING TO LEARN TO 490 00:17:00,160 --> 00:17:02,040 COMMUNICATE BETTER IN ENGLISH 491 00:17:02,040 --> 00:17:03,240 AND HE THOUGHT WALLER IT 492 00:17:03,240 --> 00:17:04,400 CRONKITE WAS THE BEST 493 00:17:04,400 --> 00:17:06,000 COMMUNICATOR HE KNEW SO HE WOULD 494 00:17:06,000 --> 00:17:08,320 TRY TO PATTERN HIS SPEECH AFTER 495 00:17:08,320 --> 00:17:09,160 WALTER CRONKITE'S. 496 00:17:09,160 --> 00:17:10,280 SO I'M GOING TO COME BACK AND 497 00:17:10,280 --> 00:17:10,800 TALK ABOUT THAT. 498 00:17:10,800 --> 00:17:12,160 SO THOSE ARE ALL SOME OF THE -- 499 00:17:12,160 --> 00:17:13,520 WHAT I'D SAY, YOU KNOW, WARM 500 00:17:13,520 --> 00:17:17,240 THINGS ABOUT MY CHILDHOOD. 501 00:17:17,240 --> 00:17:18,560 NOW, ONE OF THE OTHER THINGS I 502 00:17:18,560 --> 00:17:19,920 WANT TO MENTION, THOUGH, IS THAT 503 00:17:19,920 --> 00:17:22,000 THE DETROIT SUBURBS, IN THE 504 00:17:22,000 --> 00:17:25,080 1980s, WERE VIRTUALLY 100% 505 00:17:25,080 --> 00:17:26,520 WHITE. 506 00:17:26,520 --> 00:17:28,000 IT'S AN AUTO TOWN. 507 00:17:28,000 --> 00:17:30,120 IN THE 1980s, THERE WAS 508 00:17:30,120 --> 00:17:30,920 SIGNIFICANT ANTI-ASIAN SENTIMENT 509 00:17:30,920 --> 00:17:32,240 IN DETROIT, AND THE REASON FOR 510 00:17:32,240 --> 00:17:33,760 THAT WAS JAPANESE COMPETITION 511 00:17:33,760 --> 00:17:35,840 WITH THE U.S. AUTO INDUSTRY. 512 00:17:35,840 --> 00:17:40,760 AND I PERSONALLY SAW SCENES LIKE 513 00:17:40,760 --> 00:17:42,120 THIS, WHERE THERE WERE AUTO 514 00:17:42,120 --> 00:17:43,600 WORKERS HERE DESTROYING A 515 00:17:43,600 --> 00:17:45,520 JAPANESE CAR TO TAKE OUT THEIR 516 00:17:45,520 --> 00:17:46,040 FRUSTRATIONS. 517 00:17:46,040 --> 00:17:47,880 SO I THINK IN MANY WAYS, IT WAS 518 00:17:47,880 --> 00:17:48,840 ONE OF PROBABLY THE MORE 519 00:17:48,840 --> 00:17:50,840 DIFFICULT PLACES TO BE AN ASIAN 520 00:17:50,840 --> 00:17:53,000 IN THIS COUNTRY BACK IN THE 521 00:17:53,000 --> 00:17:53,520 1980s. 522 00:17:53,520 --> 00:17:55,560 AND I FOUND THIS YOUTUBE CLIP. 523 00:17:55,560 --> 00:17:56,960 THERE WAS A SHOW THAT WE USED TO 524 00:17:56,960 --> 00:17:58,800 WATCH ON TV EVERY SINGLE WEEK, 525 00:17:58,800 --> 00:18:01,120 AND I THOUGHT I WOULD JUST SHOW 526 00:18:01,120 --> 00:18:02,160 A COMMERCIAL. 527 00:18:02,160 --> 00:18:03,000 >> HELLO, EVERYONE IN TELEVISION 528 00:18:03,000 --> 00:18:03,760 LAND. 529 00:18:03,760 --> 00:18:06,200 IT'S ME, YOUR HOST OF MARTIAL 530 00:18:06,200 --> 00:18:10,200 ARTS THIGH TER RIGHT HERE Y 531 00:18:10,200 --> 00:18:16,360 NIGHT AT 11:00 P.M., JOIN ME FOR 532 00:18:16,360 --> 00:18:18,040 MARTIAL ARTS THEATER, WE SHOW 533 00:18:18,040 --> 00:18:20,640 CLASSIC MARTIAL ARTS FILMS AND 534 00:18:20,640 --> 00:18:22,120 ALSO PATTI HERE WILL HAVE A 535 00:18:22,120 --> 00:18:24,120 DEMONSTRATION OF ACTUAL MARTIAL 536 00:18:24,120 --> 00:18:25,320 ARTS FOR YOU EVERY WEEK. 537 00:18:25,320 --> 00:18:27,480 BUT DON'T WORRY, BECAUSE NONE OF 538 00:18:27,480 --> 00:18:29,920 THIS IS DANGEROUS. 539 00:18:29,920 --> 00:18:32,560 YOU'RE KILLING ME! 540 00:18:32,560 --> 00:18:35,080 >> SO THIS WAS THE OTHER SIDE OF 541 00:18:35,080 --> 00:18:36,120 MY HOMETOWN. 542 00:18:36,120 --> 00:18:37,720 AND AGAIN, WE WATCHED THIS SHOW 543 00:18:37,720 --> 00:18:39,880 EVERY SINGLE WEEK, AND YOU KNOW, 544 00:18:39,880 --> 00:18:42,440 REALLY SORT OF ENJOYED THE 545 00:18:42,440 --> 00:18:42,760 MOVIES. 546 00:18:42,760 --> 00:18:49,040 BUT IN THAT BACKGROUND, IN 1982, 547 00:18:49,040 --> 00:18:53,040 THERE WAS THE VINCENT CHIN CASE. 548 00:18:53,040 --> 00:18:54,160 FOR THOSE OF YOU WHO AREN'T 549 00:18:54,160 --> 00:18:55,920 AWARE OF THAT CASE, IT WAS A 550 00:18:55,920 --> 00:18:57,840 RACIALLY MOTIVATED MURDER THAT 551 00:18:57,840 --> 00:19:00,960 OCCURRED RIGHT OUTSIDE DETROIT. 552 00:19:00,960 --> 00:19:03,080 VINCENT CHIN WAS 27 YEARS OLD AT 553 00:19:03,080 --> 00:19:04,600 HIS BACHELOR'S PARTY AND GOT 554 00:19:04,600 --> 00:19:05,800 INTO AN ALTERCATION WITH TWO 555 00:19:05,800 --> 00:19:07,040 PEOPLE WHO WERE UNEMPLOYED AUTO 556 00:19:07,040 --> 00:19:10,400 WORKERS WHO BASICALLY BLAMED HIM 557 00:19:10,400 --> 00:19:12,080 FOR LOSING THEIR JOBS. 558 00:19:12,080 --> 00:19:13,640 HE WAS MURDERED WITH A BASEBALL 559 00:19:13,640 --> 00:19:14,200 BAT THAT EVENING. 560 00:19:14,200 --> 00:19:15,600 WHEN THE CASE WENT TO TRIAL, 561 00:19:15,600 --> 00:19:17,040 THERE WAS NEVER ANY DOUBT THAT 562 00:19:17,040 --> 00:19:18,880 THE TWO PEOPLE WERE GUILTY, BUT 563 00:19:18,880 --> 00:19:20,560 THE JUDGE BASICALLY SAID THAT 564 00:19:20,560 --> 00:19:21,880 THESE WEREN'T THE TIME OF MEN 565 00:19:21,880 --> 00:19:23,760 THAT YOU SEND TO JAIL, AND HE 566 00:19:23,760 --> 00:19:25,880 SENTENCED THEM TO THREE YEARS OF 567 00:19:25,880 --> 00:19:27,720 PROBATION AND A $3,000 FINE. 568 00:19:27,720 --> 00:19:29,080 SO THEY NEVER REALLY WENT TO 569 00:19:29,080 --> 00:19:31,360 JAIL FOR THAT. 570 00:19:31,360 --> 00:19:34,240 AND THAT SITUATION CAUSED A LOT 571 00:19:34,240 --> 00:19:35,240 OF DISAPPOINTMENT AND 572 00:19:35,240 --> 00:19:37,120 FRUSTRATION WITHIN THE ASIAN 573 00:19:37,120 --> 00:19:38,480 AMERICAN COMMUNITY, INITIALLY 574 00:19:38,480 --> 00:19:41,560 WITHIN DETROIT AND THEN 575 00:19:41,560 --> 00:19:42,400 NATIONALLY, AND IT WAS REALLY 576 00:19:42,400 --> 00:19:43,920 THE BEGINNING OF WHAT I WOULD 577 00:19:43,920 --> 00:19:44,880 CALL AN ASIAN AMERICAN ACTIVIST 578 00:19:44,880 --> 00:19:45,240 MOVEMENT. 579 00:19:45,240 --> 00:19:46,920 AND SEVERAL OF MY VERY CLOSE 580 00:19:46,920 --> 00:19:48,200 FAMILY FRIENDS WERE INVOLVED IN 581 00:19:48,200 --> 00:19:48,480 THIS. 582 00:19:48,480 --> 00:19:49,960 AND WHAT I LEARNED FROM THAT IS 583 00:19:49,960 --> 00:19:52,360 THAT YOU REALLY NEED TO SPEAK UP 584 00:19:52,360 --> 00:19:53,640 INDIVIDUALLY AND AS A COMMUNITY 585 00:19:53,640 --> 00:19:56,480 ABOUT THINGS LIKE THIS. 586 00:19:56,480 --> 00:19:57,800 THE OTHER THING I NOTICED EVEN 587 00:19:57,800 --> 00:19:59,320 AS A CHILD WAS THAT THERE WAS 588 00:19:59,320 --> 00:20:00,880 VERY STRONG SUPPORT FROM THE 589 00:20:00,880 --> 00:20:02,520 BLACK COMMUNITY, THE JEWISH 590 00:20:02,520 --> 00:20:03,880 COMMUNITY, AND OTHER GROUPS FOR 591 00:20:03,880 --> 00:20:06,360 THIS MOVEMENT BECAUSE THEY FELT 592 00:20:06,360 --> 00:20:07,720 SOLIDARITY FROM THEIR OWN 593 00:20:07,720 --> 00:20:08,760 HISTORICAL CHALLENGES. 594 00:20:08,760 --> 00:20:09,800 AND I'LL COME BACK AND TALK A 595 00:20:09,800 --> 00:20:13,600 LITTLE BIT MORE ABOUT THAT. 596 00:20:13,600 --> 00:20:14,800 SO WITHIN THAT SETTING, THIS WAS 597 00:20:14,800 --> 00:20:16,240 THE OTHER ELEMENT OF MY 598 00:20:16,240 --> 00:20:16,520 CHILDHOOD. 599 00:20:16,520 --> 00:20:19,840 AND I WOULD SAY THAT OVERALL, I 600 00:20:19,840 --> 00:20:21,520 HAD A VERY GOOD -- ACTUALLY A 601 00:20:21,520 --> 00:20:24,800 GREAT CHILDHOOD EXPERIENCE. 602 00:20:24,800 --> 00:20:27,080 BUT ONE OF THE -- I ALWAYS GREW 603 00:20:27,080 --> 00:20:29,000 UP PLAYING A LOT OF SPORTS, I 604 00:20:29,000 --> 00:20:31,120 WAS THE LEAD IN THE SCHOOL 605 00:20:31,120 --> 00:20:34,000 PRODUCTION OF MCBETH WHICH I 606 00:20:34,000 --> 00:20:35,840 MENTIONED TO A FEW PEOPLE AT NEI 607 00:20:35,840 --> 00:20:37,080 AND NOBODY BELIEVED ME, BUT IT 608 00:20:37,080 --> 00:20:37,680 WAS REAL. 609 00:20:37,680 --> 00:20:40,440 BUT AS A HIGH SCHOOL STUDENT, MY 610 00:20:40,440 --> 00:20:41,720 DAD, TOWARD THE END OF HIS 611 00:20:41,720 --> 00:20:45,800 CAREER, HAD A VERY HIGH ADD 612 00:20:45,800 --> 00:20:46,800 ADMINISTRATIVE POSITION AT FORD, 613 00:20:46,800 --> 00:20:48,600 AND HIS YOUNGER BROTHER, WHO'S 614 00:20:48,600 --> 00:20:52,160 MY UNCLE, HAD ALSO A VERY HIGH 615 00:20:52,160 --> 00:20:53,040 ADMINISTRATION POSITION AT ONE 616 00:20:53,040 --> 00:20:56,080 OF THE SEMICONDUCTOR -- MAJOR 617 00:20:56,080 --> 00:20:56,640 SEMICONDUCTOR COMPANIES. 618 00:20:56,640 --> 00:20:57,680 AND THEY WERE BOTH TRYING TO 619 00:20:57,680 --> 00:20:59,040 TELL ME, WHY DON'T YOU RUN FOR 620 00:20:59,040 --> 00:21:00,760 CLASS PRESIDENT IN HIGH SCHOOL, 621 00:21:00,760 --> 00:21:02,120 AND I WOULD ALWAYS RESPOND, 622 00:21:02,120 --> 00:21:03,560 WELL, I DON'T WANT TO BECOME 623 00:21:03,560 --> 00:21:03,960 CLASS PRESIDENT. 624 00:21:03,960 --> 00:21:05,600 I JUST WANT TO DO THOSE THINGS 625 00:21:05,600 --> 00:21:06,720 THAT INTEREST ME. 626 00:21:06,720 --> 00:21:08,440 AND PLUS, I DON'T THINK I WOULD 627 00:21:08,440 --> 00:21:11,160 EVER HAVE A CHANCE OF WINNING IN 628 00:21:11,160 --> 00:21:11,760 THIS SETTING. 629 00:21:11,760 --> 00:21:13,200 SO I NEVER REALLY SERIOUSLY 630 00:21:13,200 --> 00:21:14,240 CONSIDERED DOING THAT. 631 00:21:14,240 --> 00:21:15,400 BUT I'LL COME TALK A LITTLE BIT 632 00:21:15,400 --> 00:21:16,760 MORE ABOUT THAT. 633 00:21:16,760 --> 00:21:20,080 SO WITH THAT SAID, ABOUT MY 634 00:21:20,080 --> 00:21:23,400 BACKGROUND, I EVENTUALLY WENT 635 00:21:23,400 --> 00:21:25,000 TO -- EVENTUALLY BECAME AN 636 00:21:25,000 --> 00:21:26,120 ENGINEER, BECAME A DOCTOR, AND 637 00:21:26,120 --> 00:21:27,600 THEN WENT TO OPHTHALMOLOGY. 638 00:21:27,600 --> 00:21:30,040 SO I WAS DOING OPHTHALMOLOGY 639 00:21:30,040 --> 00:21:33,040 RESIDENCY AT JOHNS HOPKINS, AND 640 00:21:33,040 --> 00:21:34,240 THIS IS 1998. 641 00:21:34,240 --> 00:21:36,960 EVERY SINGLE DAY, WE WOULD HAVE 642 00:21:36,960 --> 00:21:37,520 LECTURES. 643 00:21:37,520 --> 00:21:38,640 SO ONE OF THE PEOPLE WOULD SPEAK 644 00:21:38,640 --> 00:21:41,360 AND THIS HAPPENS TO BE 645 00:21:41,360 --> 00:21:42,320 OCTOBER 1998. 646 00:21:42,320 --> 00:21:43,240 AND YOU KNOW, SOMEBODY WOULD 647 00:21:43,240 --> 00:21:45,440 COME IN AND SPEAK EITHER FROM 648 00:21:45,440 --> 00:21:46,400 JOHNS HOPKINS OR FROM THE 649 00:21:46,400 --> 00:21:47,800 COMMUNITY. 650 00:21:47,800 --> 00:21:50,240 AND IT TURNED OUT THAT IN 1999, 651 00:21:50,240 --> 00:21:52,400 WE HAD TWO GUEST SPEAKERS. 652 00:21:52,400 --> 00:21:56,000 ONE OF THEM WAS EMILY CHU, WHO 653 00:21:56,000 --> 00:21:57,640 WAS AT THE NEI AND I'M PROUD TO 654 00:21:57,640 --> 00:21:59,120 SAY IS ONE OF MY COLLEAGUES NOW 655 00:21:59,120 --> 00:22:00,920 AT THE NEI, AND THE SECOND GUEST 656 00:22:00,920 --> 00:22:02,440 SPEAKER A FEW MONTHS LATER WAS 657 00:22:02,440 --> 00:22:05,960 SOMEBODY NAMED PAUL LI, WHO WAS 658 00:22:05,960 --> 00:22:06,760 THEN AT DUKE UNIVERSITY AND IS 659 00:22:06,760 --> 00:22:08,880 NOW THE CHAIR OF OPHTHALMOLOGY 660 00:22:08,880 --> 00:22:10,400 AT THE UNIVERSITY OF MICHIGAN. 661 00:22:10,400 --> 00:22:12,120 BUT WHEN THOSE TWO SPOKE, IT 662 00:22:12,120 --> 00:22:14,080 OCCURRED TO ME THAT I HAD NEVER 663 00:22:14,080 --> 00:22:16,920 LISTENED TO A LECTURE FROM AN 664 00:22:16,920 --> 00:22:19,880 ASIAN SPEAKER BEFORE IN ALL 665 00:22:19,880 --> 00:22:21,920 THESE YEARS OF BEING AN 666 00:22:21,920 --> 00:22:23,080 OPHTHALMOLOGY RESIDENT. 667 00:22:23,080 --> 00:22:24,440 SO THAT REALLY STRUCK ME, THAT 668 00:22:24,440 --> 00:22:25,800 I'D NEVER SEEN ANYBODY WHO 669 00:22:25,800 --> 00:22:29,000 LOOKED LIKE ME AT THE PODIUM. 670 00:22:29,000 --> 00:22:30,800 AND THEN IT TURNED OUT THAT TWO 671 00:22:30,800 --> 00:22:33,080 YEARS LATER, WHEN I GOT MY FIRST 672 00:22:33,080 --> 00:22:37,000 JOB, MY FIRST CHAIR WAS A 673 00:22:37,000 --> 00:22:40,400 RETINAL SPECIALIST NAMED STANLEY 674 00:22:40,400 --> 00:22:41,000 CHANG AT COLUMBIA UNIVERSITY. 675 00:22:41,000 --> 00:22:42,880 AND IT REALLY GAVE ME AN 676 00:22:42,880 --> 00:22:43,480 INTERESTING PERSPECTIVE TO HAVE 677 00:22:43,480 --> 00:22:44,120 A CHAIRMAN WHO WAS ASIAN 678 00:22:44,120 --> 00:22:45,880 AMERICAN. 679 00:22:45,880 --> 00:22:47,720 AND I'LL BUILD ON SOME OF THAT 680 00:22:47,720 --> 00:22:48,080 AFTERWARDS. 681 00:22:48,080 --> 00:22:49,080 BUT IT WAS REALLY IN THIS 682 00:22:49,080 --> 00:22:51,960 SETTING THAT I STARTED MY 683 00:22:51,960 --> 00:22:52,160 CAREER. 684 00:22:52,160 --> 00:22:54,440 AND AS DR. GOTTESMAN SAID, YOU 685 00:22:54,440 --> 00:22:55,960 KNOW, I STARTED MY CAREER 686 00:22:55,960 --> 00:22:58,160 STUDYING A DISEASE CALLED 687 00:22:58,160 --> 00:22:59,120 RETINOPATHY OF PREMATURITY. 688 00:22:59,120 --> 00:23:00,400 I WANT TO TALK A LITTLE BIT 689 00:23:00,400 --> 00:23:02,480 ABOUT THAT SCIENCE AND HOW THIS 690 00:23:02,480 --> 00:23:03,800 LEADS INTO ARTIFICIAL 691 00:23:03,800 --> 00:23:04,440 INTELLIGENCE AND SOME OF THE 692 00:23:04,440 --> 00:23:08,200 THINGS THAT WE DID IN A.I. FOR 693 00:23:08,200 --> 00:23:08,880 OPHTHALMOLOGY, BUT WHAT I WANT 694 00:23:08,880 --> 00:23:12,280 TO TRY TO DO IS TO KEEP THE 695 00:23:12,280 --> 00:23:13,920 GENERAL PRINCIPLES EVIDENT, SO 696 00:23:13,920 --> 00:23:14,760 ASSUMING THAT YOU'RE NOT 697 00:23:14,760 --> 00:23:16,080 INTERESTED IN THIS PARTICULAR 698 00:23:16,080 --> 00:23:17,640 DISEASE, THAT I'M HOPING TO MAKE 699 00:23:17,640 --> 00:23:19,160 SOME PRINCIPLES COME OUT THAT 700 00:23:19,160 --> 00:23:20,520 ARE REALLY GOING TO APPLY TO 701 00:23:20,520 --> 00:23:24,120 ANYBODY WHO'S DOING SCIENCE OR 702 00:23:24,120 --> 00:23:28,560 CLINICAL CARE. 703 00:23:28,560 --> 00:23:29,440 THIS PARTICULAR DISEASE, WHY 704 00:23:29,440 --> 00:23:31,320 DOES IT MATTER, WELL, IT'S ONE 705 00:23:31,320 --> 00:23:34,240 OF THE LEADING CAUSES OF 706 00:23:34,240 --> 00:23:35,800 CHILDHOOD BLINDNESS AROUND THE 707 00:23:35,800 --> 00:23:36,960 WORLD. 708 00:23:36,960 --> 00:23:38,200 STEVIE WONDER IS SOMEBODY WHO 709 00:23:38,200 --> 00:23:40,720 WENT BLIND FROM RETINOPATHY OF 710 00:23:40,720 --> 00:23:41,640 PREMATURITY IN THE 1950s 711 00:23:41,640 --> 00:23:43,080 BEFORE THERE WAS A TREATMENT. 712 00:23:43,080 --> 00:23:44,520 NOW ALMOST ALL THE CASES OF 713 00:23:44,520 --> 00:23:45,520 BLINDNESS ARE PREVENTABLE IF 714 00:23:45,520 --> 00:23:46,920 THEY'RE DIAGNOSED EARLY AND 715 00:23:46,920 --> 00:23:47,800 TREATED APPROPRIATELY. 716 00:23:47,800 --> 00:23:49,600 THERE HAVE BEEN TENS OF MILLIONS 717 00:23:49,600 --> 00:23:51,640 OF DOLLARS OF NIH FUNDED 718 00:23:51,640 --> 00:23:52,520 CLINICAL TRIALS THAT HAVE GONE 719 00:23:52,520 --> 00:23:54,760 INTO DEMONSTRATING THIS. 720 00:23:54,760 --> 00:23:56,320 THE PROBLEM IS THAT WE'RE NOT 721 00:23:56,320 --> 00:23:57,640 VERY GOOD AT DIAGNOSING 722 00:23:57,640 --> 00:23:59,520 RETINOPATHY OF PREMATURITY. 723 00:23:59,520 --> 00:24:02,920 THE WAY THAT YOU DIAGNOSE SEVERE 724 00:24:02,920 --> 00:24:04,520 DISEASE IS THAT YOU LOOK AT THE 725 00:24:04,520 --> 00:24:06,600 BLOOD VESSELS AND IF THEY'RE 726 00:24:06,600 --> 00:24:11,880 DILATED, THICKER AND TORE TEU , 727 00:24:11,880 --> 00:24:12,720 WIGGLY, THAT'S SOMETHING CALLED 728 00:24:12,720 --> 00:24:13,960 PLUS DISEASE AND THAT'S BAD. 729 00:24:13,960 --> 00:24:15,160 ACCORDING TO THESE TENS OF 730 00:24:15,160 --> 00:24:17,440 MILLIONS OF DOLLARS OF NIH 731 00:24:17,440 --> 00:24:18,880 TRIALS, IN YOU SEE PLUS DISEASE 732 00:24:18,880 --> 00:24:20,400 IN BABIES AT HIGH RISK OF GOING 733 00:24:20,400 --> 00:24:21,480 BLIND, YOU NEED TO TREAT THAT 734 00:24:21,480 --> 00:24:22,840 BABY WITH LASER OR SOME OTHER 735 00:24:22,840 --> 00:24:25,760 TYPE OF DRUG INJECTION TO 736 00:24:25,760 --> 00:24:27,520 PREVENT THAT BLINDNESS AND YOU 737 00:24:27,520 --> 00:24:29,240 NEED TO DO IT REALLY QUICKLY. 738 00:24:29,240 --> 00:24:30,120 IN 2007 WE DID WHAT I THOUGHT 739 00:24:30,120 --> 00:24:32,680 WAS AN EMBARRASSINGLY BASIC 740 00:24:32,680 --> 00:24:34,000 STUDY WHERE WE PUT SOME PICTURES 741 00:24:34,000 --> 00:24:36,200 UP ON THE WEB AND WE ASKED 742 00:24:36,200 --> 00:24:37,720 21 WORLD EXPERTS TO GO IN, LOG 743 00:24:37,720 --> 00:24:40,200 IN AND YOU DIAGNOSE EACH IMAGE 744 00:24:40,200 --> 00:24:42,120 AS PLUS, PRE-PLUS OR NOT PLUS. 745 00:24:42,120 --> 00:24:44,120 IN OTHER WORDS, BAD, 746 00:24:44,120 --> 00:24:45,320 INTERMEDIATE OR NORMAL. 747 00:24:45,320 --> 00:24:47,880 AND HERE'S AN EXAMPLE. 748 00:24:47,880 --> 00:24:52,600 50% CALLED THIS ONE PLUS, 50% 749 00:24:52,600 --> 00:24:54,280 CALLED IT NOT PLUS. 750 00:24:54,280 --> 00:24:58,800 WE HAD IMAGES THAT SPLIT 50/50, 751 00:24:58,800 --> 00:24:59,520 60/40. 752 00:24:59,520 --> 00:25:02,760 OBVIOUSLY THIS IS A AN INTUITIVE 753 00:25:02,760 --> 00:25:04,760 PROBLEM IF THIS IS SUCH A BAD 754 00:25:04,760 --> 00:25:06,640 DISEASE AND CAUSES BABIES TO GO 755 00:25:06,640 --> 00:25:09,080 BLIND, WHY ARE WORLD EXPERTS 756 00:25:09,080 --> 00:25:10,280 SPLITTING 50/50 IN TERMS OF HOW 757 00:25:10,280 --> 00:25:11,400 THEY DIAGNOSE, SO DISAGREEMENT 758 00:25:11,400 --> 00:25:11,840 IN DIAGNOSIS. 759 00:25:11,840 --> 00:25:13,800 I WOULD ARGUE WE SEE THIS IN 760 00:25:13,800 --> 00:25:15,440 EVERY AREA OF MEDICINE. 761 00:25:15,440 --> 00:25:17,720 NOW, I WANT TO TALK ABOUT 762 00:25:17,720 --> 00:25:19,560 DISAGREEMENT IN PROCESS AS WELL, 763 00:25:19,560 --> 00:25:22,280 THE DIAGNOSTIC PROCESS. 764 00:25:22,280 --> 00:25:23,720 THIS IS GOING TO LEAD INTO WHAT 765 00:25:23,720 --> 00:25:25,480 I CALL THE SCIENCE AND THE ART 766 00:25:25,480 --> 00:25:28,120 OF MEDICAL CARE. 767 00:25:28,120 --> 00:25:28,880 SO THREE YEARS AFTER THE 768 00:25:28,880 --> 00:25:31,520 PREVIOUS SLIDE, I WAS ON A PANEL 769 00:25:31,520 --> 00:25:35,920 AT ONE OF THE NATIONAL MEETINGS 770 00:25:35,920 --> 00:25:37,040 AND WE WERE TALKING ABOUT PLUS 771 00:25:37,040 --> 00:25:38,160 DISEASE AND HOW DO YOU DIAGNOSE 772 00:25:38,160 --> 00:25:38,520 IT. 773 00:25:38,520 --> 00:25:40,240 ONE OF THE EXPERT PANELISTS MADE 774 00:25:40,240 --> 00:25:42,800 THE ANALOGY TO A SUPREME COURT 775 00:25:42,800 --> 00:25:44,520 CASE IN THE 1960s. 776 00:25:44,520 --> 00:25:47,280 THERE WAS A SUPREME COURT 777 00:25:47,280 --> 00:25:48,440 JUSTICE, POTTER STEWART, AND 778 00:25:48,440 --> 00:25:52,440 THEY WERE ARGUING A CASE ON 779 00:25:52,440 --> 00:25:54,000 PORNOGRAPHY, AND POTTER STEWART 780 00:25:54,000 --> 00:25:56,240 SAID, WELL, I CAN'T DEFINE 781 00:25:56,240 --> 00:25:56,880 PORNOGRAPHY, BUT I KNOW IT WHEN 782 00:25:56,880 --> 00:25:59,360 I SEE IT BECAUSE IT JUST LOOKS 783 00:25:59,360 --> 00:25:59,680 BAD. 784 00:25:59,680 --> 00:26:03,240 SO THAT EXPERT PANELIST'S 785 00:26:03,240 --> 00:26:04,320 ANALOGY WAS THAT PLUS DISEASE IS 786 00:26:04,320 --> 00:26:05,400 LIKE PORNOGRAPHY, THAT IT'S 787 00:26:05,400 --> 00:26:07,040 REALLY, REALLY HARD TO 788 00:26:07,040 --> 00:26:09,040 CHARACTERIZE WHAT IT MEANS TO 789 00:26:09,040 --> 00:26:09,960 HAVE PLUS DISEASE, BUT YOU KNOW 790 00:26:09,960 --> 00:26:11,160 IT WHEN YOU SEE IT BECAUSE IT 791 00:26:11,160 --> 00:26:16,200 JUST LOOKS BAD. 792 00:26:16,200 --> 00:26:17,200 THAT COMMENT BOTHERED ME FOR 793 00:26:17,200 --> 00:26:19,040 MONTHS BECAUSE I THOUGHT, WELL, 794 00:26:19,040 --> 00:26:20,280 WE'RE TRYING TO BE SCIENTIFIC 795 00:26:20,280 --> 00:26:21,480 ABOUT THINGS, HOW CAN WE JUST BE 796 00:26:21,480 --> 00:26:25,080 SAYING THAT THINGS JUST LOOK 797 00:26:25,080 --> 00:26:25,280 BAD? 798 00:26:25,280 --> 00:26:26,320 ON THE OTHER HAND, EVERY 799 00:26:26,320 --> 00:26:27,720 CLINICIAN WILL TELL YOU THAT 800 00:26:27,720 --> 00:26:30,000 SOMETIMES THINGS JUST LOOK BAD. 801 00:26:30,000 --> 00:26:33,160 SOMETIMES THE RETINA LOOKS 802 00:26:33,160 --> 00:26:36,200 AGGRESSIVE, IF AN ICU DOCTOR -- 803 00:26:36,200 --> 00:26:37,480 SOMETIMES THAT PATIENT JUST 804 00:26:37,480 --> 00:26:38,720 LOOKS SICK IN A WAY THAT EXPERTS 805 00:26:38,720 --> 00:26:40,000 CAN'T ALWAYS ARTICULATE. 806 00:26:40,000 --> 00:26:42,680 SO WE BECAME INTERESTED IN THIS 807 00:26:42,680 --> 00:26:43,680 QUESTION, COULD VARIABILITY IN 808 00:26:43,680 --> 00:26:46,440 THE DIAGNOSTIC PROCESS, IN OTHER 809 00:26:46,440 --> 00:26:47,440 WORDS, IF DIFFERENT PEOPLE HAVE 810 00:26:47,440 --> 00:26:48,560 DIFFERENT CONCEPTS OF WHAT IT 811 00:26:48,560 --> 00:26:50,560 MEANS TO JUST LOOK BAD, COULD 812 00:26:50,560 --> 00:26:51,600 THAT EXPLAIN SOME OF THE 813 00:26:51,600 --> 00:26:52,840 VARIABILITY THAT WE'RE SEEING ON 814 00:26:52,840 --> 00:26:54,120 THE PREVIOUS SLIDE. 815 00:26:54,120 --> 00:26:56,160 SO WE DESIGNED A SET OF STUDIES 816 00:26:56,160 --> 00:27:00,400 TO TRY TO CAPTURE AND ENCODE THE 817 00:27:00,400 --> 00:27:02,840 DETAILED QUALITATIVE THOUGHTS OF 818 00:27:02,840 --> 00:27:04,360 THE SEVEN MOST SEASONED ROP 819 00:27:04,360 --> 00:27:07,480 EXPERTS IN THE WORLD. 820 00:27:07,480 --> 00:27:09,280 THERE WERE PEOPLE WHO ORIGINALLY 821 00:27:09,280 --> 00:27:11,840 CHARACTERIZED THE DISEASE AND WE 822 00:27:11,840 --> 00:27:13,000 THOUGHT LET'S VIDEOTAPE THEM 823 00:27:13,000 --> 00:27:14,440 WHEN THEY'RE MAKING THESE 824 00:27:14,440 --> 00:27:15,080 DIAGNOSES. 825 00:27:15,080 --> 00:27:16,800 WE SPENT MONTHS PICKING IMAGES, 826 00:27:16,800 --> 00:27:18,480 SCRIPTING QUESTIONS, AND WE HAD 827 00:27:18,480 --> 00:27:19,880 THEM ANNOTATE WITH A PEN THE 828 00:27:19,880 --> 00:27:21,520 AREAS ON THE IMAGES THAT EITHER 829 00:27:21,520 --> 00:27:23,080 WORRIED THEM OR REASSURED THEM, 830 00:27:23,080 --> 00:27:24,960 AND WE THEN HAD A COGNITIVE 831 00:27:24,960 --> 00:27:27,200 PSYCHOLOGIST GO BACK AND ANALYZE 832 00:27:27,200 --> 00:27:29,080 HOURS OF VIDEOTAPE TRANSCRIPT TO 833 00:27:29,080 --> 00:27:32,200 TRY TO MODEL THEIR THOUGHT 834 00:27:32,200 --> 00:27:32,440 PROCESS. 835 00:27:32,440 --> 00:27:34,320 SO HERE'S AN EXAMPLE. 836 00:27:34,320 --> 00:27:37,040 THERE'S ONE IMAGE, AND THREE 837 00:27:37,040 --> 00:27:37,960 EXPERTS. 838 00:27:37,960 --> 00:27:39,040 EXPERT NUMBER ONE DIAGNOSED AS 839 00:27:39,040 --> 00:27:40,840 PLUS DISEASE, EXPERT NUMBER TWO 840 00:27:40,840 --> 00:27:42,960 DIAGNOSED IT AS PRE-PLUS 841 00:27:42,960 --> 00:27:44,280 DISEASE, EXPERT NUMBER THREE 842 00:27:44,280 --> 00:27:45,560 DIAGNOSED IT AS NORMAL, AND 843 00:27:45,560 --> 00:27:46,640 THEY'RE ALL LOOKING AT DIFFERENT 844 00:27:46,640 --> 00:27:47,920 PARTS OF THE RETINA WHEN THEY DO 845 00:27:47,920 --> 00:27:48,440 THAT. 846 00:27:48,440 --> 00:27:50,720 SO THE DIFFERENCE IN THE DYING 847 00:27:50,720 --> 00:27:51,360 DIAGNOSTIC PROCESS AS WELL AS 848 00:27:51,360 --> 00:27:55,120 THE OVERALL DIAGNOSIS. 849 00:27:55,120 --> 00:27:58,080 AND MINA HEWING WAS A VISITING 850 00:27:58,080 --> 00:27:59,000 RESEARCHER FROM GERMANY WHO 851 00:27:59,000 --> 00:28:00,280 WROTE A REALLY NEAT PAPER ON 852 00:28:00,280 --> 00:28:00,680 THIS. 853 00:28:00,680 --> 00:28:03,320 WHEN WE LOOKED BACK AT THE 854 00:28:03,320 --> 00:28:04,760 LITERATURE, THERE'S REALLY GOOD 855 00:28:04,760 --> 00:28:06,200 EVIDENCE THIS HAPPENS IN EVERY 856 00:28:06,200 --> 00:28:07,800 AREA OF MEDICINE. 857 00:28:07,800 --> 00:28:09,040 DIFFERENCE IN THE DIAGNOSTIC 858 00:28:09,040 --> 00:28:10,000 PROCESS AS WELL AS THE OVERALL 859 00:28:10,000 --> 00:28:10,640 DIAGNOSIS. 860 00:28:10,640 --> 00:28:13,240 NOW, IN FACT, WHEN YOU GET 861 00:28:13,240 --> 00:28:14,920 DIFFERENT EXPERTS, THIS IS EIGHT 862 00:28:14,920 --> 00:28:16,160 WORLD EXPERTS, TO LOOK AT 863 00:28:16,160 --> 00:28:18,920 DIFFERENT IMAGES, THIS IS 100 864 00:28:18,920 --> 00:28:21,120 DIFFERENT RETINAL IMAGES, THESE 865 00:28:21,120 --> 00:28:23,320 COLUMBUS, A LOT OF INTERESTING 866 00:28:23,320 --> 00:28:24,360 THINGS HAPPEN. 867 00:28:24,360 --> 00:28:26,520 PETE CAMPBELL WAS A JUNIOR 868 00:28:26,520 --> 00:28:29,000 COLLEAGUE IN OREGON, AND KRAMER 869 00:28:29,000 --> 00:28:30,400 IS A LONG TIME COLLABORATOR. 870 00:28:30,400 --> 00:28:31,520 WE WORKED TOGETHER FOR 10 YEARS 871 00:28:31,520 --> 00:28:32,720 AND SHE SPOKE AT NIH VIRTUALLY 872 00:28:32,720 --> 00:28:40,760 JUST A WEEK AGO ON ARTIFICIAL 873 00:28:40,760 --> 00:28:41,080 INTELLIGENCE. 874 00:28:41,080 --> 00:28:42,800 AND THEY WROTE A REALLY NICE 875 00:28:42,800 --> 00:28:44,160 SERIES OF PAPERS ON THIS. 876 00:28:44,160 --> 00:28:45,720 IF THAT EXPERT DIAGNOSED THAT 877 00:28:45,720 --> 00:28:46,880 IMAGE AS PLUS DISEASE, IN OTHER 878 00:28:46,880 --> 00:28:48,080 WORDS, REALLY SEVERE, THE BOX IS 879 00:28:48,080 --> 00:28:48,600 IN RED. 880 00:28:48,600 --> 00:28:51,080 IF THEY DIAGNOSE PRE-PLUS 881 00:28:51,080 --> 00:28:52,320 DISEASE, INTERMEDIATE, THE BOX 882 00:28:52,320 --> 00:28:52,840 IS YELLOW. 883 00:28:52,840 --> 00:28:54,520 IF THEY DIAGNOSE IT AS NORMAL, 884 00:28:54,520 --> 00:28:57,600 THE BOX IS IN GREEN. 885 00:28:57,600 --> 00:28:59,960 SO THE FIRST THING TO NOTICE 886 00:28:59,960 --> 00:29:02,400 HERE IS EXPERTS 7 AND 8, ALL 887 00:29:02,400 --> 00:29:04,560 WORLD EXPERTS, DIAGNOSE PLUS 888 00:29:04,560 --> 00:29:05,360 DISEASE SIX TIMES MORE OFTEN 889 00:29:05,360 --> 00:29:06,600 THAN EXPERT NUMBER ONE. 890 00:29:06,600 --> 00:29:11,040 IF YOU GO TO A DIRN DIFFERENT 891 00:29:11,040 --> 00:29:12,440 DATASET, 34 IMAGES, IT'S THAT 892 00:29:12,440 --> 00:29:14,640 EXACT SAME 6:1 RATIO. 893 00:29:14,640 --> 00:29:16,480 SO YOU'VE GOT OVER AND 894 00:29:16,480 --> 00:29:18,800 UNDERCALLERS. 895 00:29:18,800 --> 00:29:20,680 WE SEE THIS IN EVERY FIELD OF 896 00:29:20,680 --> 00:29:22,200 MEDICINE, THERE ARE SOME DOCTORS 897 00:29:22,200 --> 00:29:24,480 THAT ARE MORE AGGRESSIVE AND 898 00:29:24,480 --> 00:29:25,440 ALWAYS RECOMMEND SURGERY AND 899 00:29:25,440 --> 00:29:26,400 THERE ARE OTHERS THAT WILL 900 00:29:26,400 --> 00:29:27,240 OBSERVE YOU AND RECOMMEND 901 00:29:27,240 --> 00:29:27,640 PHYSICAL THERAPY. 902 00:29:27,640 --> 00:29:29,000 SO THIS IS WHAT IT LOOKS LIKE ON 903 00:29:29,000 --> 00:29:29,520 A GRID. 904 00:29:29,520 --> 00:29:31,240 ON THE OTHER HAND, IF YOU GO 905 00:29:31,240 --> 00:29:33,280 DOWN EACH COLUMN AND CALCULATE 906 00:29:33,280 --> 00:29:34,640 AN AVERAGE SCORE FOR EACH IMAGE, 907 00:29:34,640 --> 00:29:36,560 SO IF YOU GIVE IT ONE POINT FOR 908 00:29:36,560 --> 00:29:41,040 A NORMAL DIAGNOSIS, TWO POINTS 909 00:29:41,040 --> 00:29:42,640 FOR A PRE-PLUS DISEASE AND THREE 910 00:29:42,640 --> 00:29:44,200 POINTS FOR A PLUS DISEASE 911 00:29:44,200 --> 00:29:45,680 DIAGNOSIS AND CALCULATE TO AN 912 00:29:45,680 --> 00:29:47,640 AVERAGE SCORE AND CONVERT TO A 913 00:29:47,640 --> 00:29:48,440 COLOR, THAT'S THIS ROW, THE 914 00:29:48,440 --> 00:29:49,160 AVERAGE SCORE. 915 00:29:49,160 --> 00:29:53,040 SO ON THE RIGHT SIDE, THESE 916 00:29:53,040 --> 00:29:54,480 IMAGES ARE VERY GREEN BECAUSE 917 00:29:54,480 --> 00:29:55,720 EVERYBODY DIAGNOSED NORMAL, ON 918 00:29:55,720 --> 00:29:57,440 THE LEFT SIDE THEY'RE VERY RED 919 00:29:57,440 --> 00:29:58,320 BECAUSE EVERYONE DIAGNOSED PLUS. 920 00:29:58,320 --> 00:29:59,520 HERE IN THE MIDDLE, THERE'S 921 00:29:59,520 --> 00:30:01,960 EVERY SHADE IN BETWEEN. 922 00:30:01,960 --> 00:30:02,920 SO THE POINT I'M TRYING TO MAKE 923 00:30:02,920 --> 00:30:04,680 WITH THIS IS THAT THE REAL WORLD 924 00:30:04,680 --> 00:30:08,000 IS A CONTINUOUS SPEC SPECTRUF 925 00:30:08,000 --> 00:30:08,960 ABNORMALITY. 926 00:30:08,960 --> 00:30:11,320 EVIDENCE-BASED MEDICINE MAKES IT 927 00:30:11,320 --> 00:30:13,000 SOUND LIKE PEOPLE COME IN 928 00:30:13,000 --> 00:30:13,720 BUCKETS. 929 00:30:13,720 --> 00:30:15,560 NORMAL, MILD, MODERATE, SEVERE, 930 00:30:15,560 --> 00:30:17,120 AND PEOPLE DO NOT COME IN 931 00:30:17,120 --> 00:30:17,600 BUCKETS. 932 00:30:17,600 --> 00:30:19,280 THEY COME ON A CONTINUOUS 933 00:30:19,280 --> 00:30:20,280 SPECTRUM. 934 00:30:20,280 --> 00:30:21,920 AND IT'S A CLINICIAN'S JOB TO 935 00:30:21,920 --> 00:30:23,840 DRAW THE LINES AND PUT PEOPLE 936 00:30:23,840 --> 00:30:24,560 INTO BUCKETS. 937 00:30:24,560 --> 00:30:25,840 BUT THE PROBLEM IS THAT PEOPLE 938 00:30:25,840 --> 00:30:27,840 DON'T DO THAT CONSISTENTLY. 939 00:30:27,840 --> 00:30:29,760 SOME PEOPLE DRAW THOSE LINES 940 00:30:29,760 --> 00:30:30,960 HERE, SOME PEOPLE DRAW THEM 941 00:30:30,960 --> 00:30:32,120 HERE, AND THAT'S REALLY ONE OF 942 00:30:32,120 --> 00:30:34,400 THE AREAS WHERE ARTIFICIAL 943 00:30:34,400 --> 00:30:36,760 INTELLIGENCE CAN HELP US IMPROVE 944 00:30:36,760 --> 00:30:38,960 DIAGNOSTIC ACCURACY AND 945 00:30:38,960 --> 00:30:39,760 CONSISTENCY. 946 00:30:39,760 --> 00:30:41,080 THERE'S BEEN ENORMOUS AMOUNT OF 947 00:30:41,080 --> 00:30:44,160 WORK IN ARTIFICIAL INTELLIGENCE 948 00:30:44,160 --> 00:30:46,400 FOR OPHTHALMOLOGY, AND I 949 00:30:46,400 --> 00:30:48,400 CAPTURED THIS TWEET FROM ERIC 950 00:30:48,400 --> 00:30:50,240 TOPOL WITH TWO YEARS AGO, OF 951 00:30:50,240 --> 00:30:51,840 ALL THE MEDICAL SPECIALTIES, 952 00:30:51,840 --> 00:30:53,920 MOST PEOPLE THINK RADIOLOGY IS 953 00:30:53,920 --> 00:30:56,080 LEADING THE A.I. MOVEMENT BUT 954 00:30:56,080 --> 00:30:57,440 IT'S ACTUALLY OPHTHALMOLOGY. 955 00:30:57,440 --> 00:30:58,040 WHY IS THAT? 956 00:30:58,040 --> 00:31:00,000 BECAUSE THERE'S SO MUCH 957 00:31:00,000 --> 00:31:04,160 IMAGE-BASED TATA DATA AND SOH 958 00:31:04,160 --> 00:31:11,840 ACCOMPANIES CLINICAL DATA. 959 00:31:11,840 --> 00:31:14,600 FOR EXAMPLE, THE FIRST 960 00:31:14,600 --> 00:31:15,600 FDA-APPROVED AUTONOMOUS 961 00:31:15,600 --> 00:31:16,760 ARTIFICIAL INTELLIGENCE SYSTEM 962 00:31:16,760 --> 00:31:18,200 IN ANY FIELD OF MEDICINE WAS 963 00:31:18,200 --> 00:31:24,160 DEVELOPED BY MICHAEL ABRAMOFF AT 964 00:31:24,160 --> 00:31:25,760 UNIVERSITY OF IOWA AND IT 965 00:31:25,760 --> 00:31:27,280 DIAGNOSES DIABETIC RETINOPATHY. 966 00:31:27,280 --> 00:31:29,160 THE FOLKS AT GOOGLE HAVE DONE 967 00:31:29,160 --> 00:31:30,920 SOME BEAUTIFUL WORK INVOLVING 968 00:31:30,920 --> 00:31:32,320 KNOWLEDGE DISCOVERY REGARDING 969 00:31:32,320 --> 00:31:33,280 SYSTEMIC HEALTH, SO THEY WERE 970 00:31:33,280 --> 00:31:34,680 ABLE TO TRAIN DEEP LEARNING 971 00:31:34,680 --> 00:31:36,280 SYSTEMS THAT, FOR EXAMPLE, TOOK 972 00:31:36,280 --> 00:31:41,400 RETINAL IMAGES AND WERE ABLE TO 973 00:31:41,400 --> 00:31:43,240 TELL DID THAT IMAGE COME FROM A 974 00:31:43,240 --> 00:31:44,960 MALE PATIENT OR FEMALE PATIENT. 975 00:31:44,960 --> 00:31:47,120 THEY CAN DO THIS WITH ALMOST 976 00:31:47,120 --> 00:31:47,760 100% ACCURACY. 977 00:31:47,760 --> 00:31:49,960 WHEN THAT PAPER WAS PUBLISHED 978 00:31:49,960 --> 00:31:53,080 FOUR YEARS AGO, IT WAS STUNNING 979 00:31:53,080 --> 00:31:55,560 TO THE MEDICAL COMMUNITY BECAUSE 980 00:31:55,560 --> 00:31:57,640 NONE OF US CAN LOOK AT AN IMAGE 981 00:31:57,640 --> 00:31:59,680 LIKE THIS AND TELL, IS IT MALE 982 00:31:59,680 --> 00:32:01,360 OR FEMALE, BUT PRESUMABLY 983 00:32:01,360 --> 00:32:03,480 THERE'S SOME DATA IN THE PIXELS 984 00:32:03,480 --> 00:32:07,600 THAT ASSOCIATES WITH GENDER. 985 00:32:07,600 --> 00:32:08,760 SO MAYBE SOMEDAY WE CAN TAKE 986 00:32:08,760 --> 00:32:10,320 ADVANTAGE OF THAT FOREKNOWLEDGE 987 00:32:10,320 --> 00:32:10,720 DISCOVERY. 988 00:32:10,720 --> 00:32:12,200 AND IN MY PIECE OF THE WORLD, WE 989 00:32:12,200 --> 00:32:15,520 WORKED ON SYSTEMS -- ARTIFICIAL 990 00:32:15,520 --> 00:32:16,520 INTELLIGENCE SYSTEMS FOR PLUS 991 00:32:16,520 --> 00:32:17,840 DISEASE DIAGNOSIS IN RETINOPATHY 992 00:32:17,840 --> 00:32:18,600 OF PREMATURITY. 993 00:32:18,600 --> 00:32:29,000 AND THAT WAS WITH THREE 994 00:32:29,240 --> 00:32:29,520 INVESTIGATORS. 995 00:32:29,520 --> 00:32:31,200 IN A NUTSHELL, HOW WE TRIED TO 996 00:32:31,200 --> 00:32:35,240 APPROACH THIS IS A TWO-LEVEL 997 00:32:35,240 --> 00:32:36,840 CONVOLUTIONAL NEURAL NETWORK. 998 00:32:36,840 --> 00:32:38,320 THE FIRST LEVEL WAS SEGMENTING 999 00:32:38,320 --> 00:32:38,640 THE IMAGE. 1000 00:32:38,640 --> 00:32:41,680 IN OTHER WORDS, GOING FROM THAT 1001 00:32:41,680 --> 00:32:43,600 IMAGE TO IDENTIFY WHAT THE 1002 00:32:43,600 --> 00:32:44,640 VESSELS ARE. 1003 00:32:44,640 --> 00:32:45,440 A VESSEL MAP. 1004 00:32:45,440 --> 00:32:46,960 NOW WHY DO WE WANT TO DO THIS? 1005 00:32:46,960 --> 00:32:51,200 AT THE TIME, THIS WAS 2016, 1006 00:32:51,200 --> 00:32:52,400 2017, WE WERE CONCERNED THAT 1007 00:32:52,400 --> 00:32:54,200 THERE ARE DARKLY PIGMENTED 1008 00:32:54,200 --> 00:32:57,440 PATIENTS AND LIGHTLY PIGMENTED 1009 00:32:57,440 --> 00:32:59,400 PATIENTS, AND WE DIDN'T WANT THE 1010 00:32:59,400 --> 00:33:01,280 A.I. SYSTEMS TO LEARN RACE AS 1011 00:33:01,280 --> 00:33:02,800 PART OF THE DIAGNOSIS. 1012 00:33:02,800 --> 00:33:05,000 SO WE THOUGHT, LET'S JUST 1013 00:33:05,000 --> 00:33:06,840 CALCULATE THE VESSEL MAPS. 1014 00:33:06,840 --> 00:33:08,080 AND THEN WE'LL GO FROM THE 1015 00:33:08,080 --> 00:33:11,560 VESSEL MAPS TO A SECOND 1016 00:33:11,560 --> 00:33:12,480 CONVOLUTIONAL NEURAL NETWORK 1017 00:33:12,480 --> 00:33:15,760 THAT WILL THEN MAP TO A FUTURE 1018 00:33:15,760 --> 00:33:16,800 SPACE AND COME UP WITH WHAT'S 1019 00:33:16,800 --> 00:33:19,920 YOUR PROBABILITY OF PLUS VERSUS 1020 00:33:19,920 --> 00:33:21,360 PRIEST-PLUS VERSUS NORMAL? 1021 00:33:21,360 --> 00:33:23,120 IN FACT WHEN DO YOU THAT, OUR 1022 00:33:23,120 --> 00:33:24,920 SYSTEMS HAD VERY, VERY HIGH 1023 00:33:24,920 --> 00:33:28,280 ACCURACY FOR DIAGNOSING, IN FACT 1024 00:33:28,280 --> 00:33:29,160 IT COMPARED TO THE EXPERTS THAT 1025 00:33:29,160 --> 00:33:31,040 WE TESTED, IT BEAT ALL OF THE 1026 00:33:31,040 --> 00:33:32,240 EXPERTS EXCEPT FOR ONE. 1027 00:33:32,240 --> 00:33:35,320 SO WHERE ARE WE NOW? 1028 00:33:35,320 --> 00:33:37,400 WE HAVE FDA BREAKTHROUGH 1029 00:33:37,400 --> 00:33:38,600 DESIGNATION SO THERE'S NOT A 1030 00:33:38,600 --> 00:33:39,640 PRODUCT ABOUT THIS, BUT HAVING 1031 00:33:39,640 --> 00:33:40,440 THOSE DISCUSSIONS. 1032 00:33:40,440 --> 00:33:41,520 AND I DON'T HAVE ANY FINANCIAL 1033 00:33:41,520 --> 00:33:43,800 INTEREST IN THIS, AND IF I DID, 1034 00:33:43,800 --> 00:33:45,240 ROP WOULD NOT THE BE THE DISEASE 1035 00:33:45,240 --> 00:33:49,000 TO BE MAKING MONEY OFF. 1036 00:33:49,000 --> 00:33:51,600 SO ANYWAY, ONE OF THE QUESTIONS 1037 00:33:51,600 --> 00:33:54,480 THAT PEOPLE ALWAYS ASK IS, WHAT 1038 00:33:54,480 --> 00:33:55,120 ABOUT IMAGE QUALITY? 1039 00:33:55,120 --> 00:33:56,720 THE ENGINEERS WILL SAY, WELL, 1040 00:33:56,720 --> 00:33:58,360 GARBAGE IN, GARBAGE OUT, AND HOW 1041 00:33:58,360 --> 00:33:59,800 DO YOU KNOW YOU'VE GOT A GOOD 1042 00:33:59,800 --> 00:34:02,320 ENOUGH IMAGE TO MAKE A 1043 00:34:02,320 --> 00:34:03,040 DIAGNOSIS? 1044 00:34:03,040 --> 00:34:04,360 ONE OF THE PH.D. STUDENTS THAT I 1045 00:34:04,360 --> 00:34:09,000 WORKED WITH IN OREGON, AARON 1046 00:34:09,000 --> 00:34:10,240 COYNER, WORKED ON A SET OF 1047 00:34:10,240 --> 00:34:10,760 PROJECTS INVOLVING THAT 1048 00:34:10,760 --> 00:34:11,160 QUESTION. 1049 00:34:11,160 --> 00:34:12,600 THIS QUESTION COMES UP ALL THE 1050 00:34:12,600 --> 00:34:13,640 TIME IN MEDICAL DIAGNOSIS. 1051 00:34:13,640 --> 00:34:17,160 AND SO WE THOUGHT, LET'S DEVELOP 1052 00:34:17,160 --> 00:34:19,480 A CONVOLUTIONAL NETWORK TO TRAIN 1053 00:34:19,480 --> 00:34:21,280 BASED ON QUALITY OF IMAGES. 1054 00:34:21,280 --> 00:34:23,480 SO BASICALLY WE HAD THOUSANDS OF 1055 00:34:23,480 --> 00:34:26,480 IMAGES THAT WERE LABELED BY 1056 00:34:26,480 --> 00:34:27,920 EXPERTS, IS THAT IMAGE 1057 00:34:27,920 --> 00:34:28,800 ACCEPTABLE FOR DIAGNOSIS, 1058 00:34:28,800 --> 00:34:30,000 POSSIBLY ACCEPTABLE OR NOT 1059 00:34:30,000 --> 00:34:30,640 ACCEPTABLE? 1060 00:34:30,640 --> 00:34:32,160 WE USE THAT AS A BASIS FOR 1061 00:34:32,160 --> 00:34:33,960 TRAINING AND VALIDATING THESE 1062 00:34:33,960 --> 00:34:36,160 NEURAL NETWORKS. 1063 00:34:36,160 --> 00:34:37,160 SO IT TURNED OUT WHEN YOU DO 1064 00:34:37,160 --> 00:34:40,600 THAT, THEY HAD VERY HIGH 1065 00:34:40,600 --> 00:34:43,680 ACCURACY COMPARED TO EXPERTS FOR 1066 00:34:43,680 --> 00:34:44,720 IDENTIFYING IMAGE QUALITY. 1067 00:34:44,720 --> 00:34:48,880 AND IN FACT, WHEN WE HAD THAT 1068 00:34:48,880 --> 00:34:51,880 A.I. SYSTEM BASICALLY RANK 30 1069 00:34:51,880 --> 00:34:55,320 IMAGES FROM LOWEST QUALITY TO 1070 00:34:55,320 --> 00:34:56,920 HIGHEST QUALITY, IT ACTUALLY 1071 00:34:56,920 --> 00:34:57,680 CORRELATED FAIRLY WELL WITH WHAT 1072 00:34:57,680 --> 00:34:59,440 EXPERTS ORDERED IN TERMS OF 1073 00:34:59,440 --> 00:35:00,120 QUALITY, THE CORRELATION 1074 00:35:00,120 --> 00:35:01,720 COEFFICIENT WAS ABOUT .9. 1075 00:35:01,720 --> 00:35:03,520 SO RIGHT NOW WHERE ARE WE WITH 1076 00:35:03,520 --> 00:35:05,760 THAT, TRYING TO INCORPORATE INTO 1077 00:35:05,760 --> 00:35:07,800 OPERATIONAL A.I. SYSTEM. 1078 00:35:07,800 --> 00:35:11,400 SO A LOT OF WORK TO BE DONE. 1079 00:35:11,400 --> 00:35:12,720 SO LASTLY I JUST WANT TO SAY 1080 00:35:12,720 --> 00:35:16,440 THAT WE TALKED ABOUT CLINICAL 1081 00:35:16,440 --> 00:35:18,000 DIAGNOSIS BEING SUBJECTIVE AND 1082 00:35:18,000 --> 00:35:19,520 QUALITATIVE IN ROP AND REALLY SO 1083 00:35:19,520 --> 00:35:20,840 MANY OTHER AREAS IN 1084 00:35:20,840 --> 00:35:22,120 OPHTHALMOLOGY AND MEDICINE, AND 1085 00:35:22,120 --> 00:35:25,960 THE QUESTION IS, CAN WE MAKE IT 1086 00:35:25,960 --> 00:35:27,280 MORE QUANTITATIVE AND OBJECTIVE. 1087 00:35:27,280 --> 00:35:29,040 HERE'S ONE OF OUR ATTEMPTS TO DO 1088 00:35:29,040 --> 00:35:29,560 THAT. 1089 00:35:29,560 --> 00:35:31,800 WE TAKE THE OUTPUTS OF THOSE 1090 00:35:31,800 --> 00:35:35,640 A.I. SYSTEMS AND WE CONVERT TO A 1091 00:35:35,640 --> 00:35:36,600 1 TO 9 SCALE. 1092 00:35:36,600 --> 00:35:39,560 SO 1 IS MOST NORMAL, 9 IS MOST 1093 00:35:39,560 --> 00:35:42,840 SEVERE. 1094 00:35:42,840 --> 00:35:45,800 AND WHO DID THE WORK? 1095 00:35:45,800 --> 00:35:47,200 CAMPBELL, KIM, VISITING 1096 00:35:47,200 --> 00:35:51,200 RESEARCHER FROM KOREA, PAUL 1097 00:35:51,200 --> 00:35:52,840 CHAN, AND KRAMER DID THE BULK OF 1098 00:35:52,840 --> 00:35:53,280 THIS. 1099 00:35:53,280 --> 00:35:57,240 SO YB DO WE WHY DO WE WANT? 1100 00:35:57,240 --> 00:35:59,280 FROM A CLINICAL PERSPECTIVE WE 1101 00:35:59,280 --> 00:36:00,240 TALKED ABOUT THIS CONTINUOUS 1102 00:36:00,240 --> 00:36:01,080 SPECTRUM OF ABNORMALITY AND 1103 00:36:01,080 --> 00:36:02,680 THOUGHT IF WE GIVE NINE LEVELS, 1104 00:36:02,680 --> 00:36:05,000 THAT WILL MIMIC THAT SPECTRUM. 1105 00:36:05,000 --> 00:36:07,320 ON THE OTHER HAND, WHAT A REAL 1106 00:36:07,320 --> 00:36:10,320 DOCTOR WANTS TO KNOW AT THE 1107 00:36:10,320 --> 00:36:11,720 BEDSIDE IS, IS MY PATIENT 1108 00:36:11,720 --> 00:36:12,680 GETTING BETTER, WORSE OR ABOUT 1109 00:36:12,680 --> 00:36:15,320 THE SAME COMPARED TO THE LAST 1110 00:36:15,320 --> 00:36:15,800 EXAM? 1111 00:36:15,800 --> 00:36:18,160 YOU CAN'T DO THAT WITH THE 1112 00:36:18,160 --> 00:36:20,000 CURRENT THREE-LEVEL SCALES, 1113 00:36:20,000 --> 00:36:21,760 PLUS, PRE-PLUS, NORMAL. 1114 00:36:21,760 --> 00:36:23,760 IT'S JUST NOT GRANULAR ENOUGH. 1115 00:36:23,760 --> 00:36:24,560 SO NINE LEVELS WE THOUGHT WOULD 1116 00:36:24,560 --> 00:36:26,360 BE A REAL WAY TO BE ABLE TO 1117 00:36:26,360 --> 00:36:27,120 COMPARE. 1118 00:36:27,120 --> 00:36:29,080 NOW FROM A PRACTICAL 1119 00:36:29,080 --> 00:36:29,920 PERSPECTIVE, ONE THING THAT I'VE 1120 00:36:29,920 --> 00:36:32,800 HEARD A LOT IN GIVING TALKS IS, 1121 00:36:32,800 --> 00:36:35,680 A DOCTOR WILL SAY, WELL, AM I AT 1122 00:36:35,680 --> 00:36:37,000 MEDICAL-LEGAL LIABILITY IF A 1123 00:36:37,000 --> 00:36:38,720 COMPUTER TELLS ME THAT MY BABY 1124 00:36:38,720 --> 00:36:40,480 HAS PLUS DISEASE BUT I DECIDE 1125 00:36:40,480 --> 00:36:42,880 NOT TO TREAT THE BABY, WHAT IF 1126 00:36:42,880 --> 00:36:44,320 SOMETHING BAD HAPPENS AND AM I 1127 00:36:44,320 --> 00:36:45,640 GOING TO GET SUED BECAUSE OF 1128 00:36:45,640 --> 00:36:46,840 THAT? 1129 00:36:46,840 --> 00:36:50,280 AND I DON'T NECESSARILY BELIEVE 1130 00:36:50,280 --> 00:36:52,640 THAT. 1131 00:36:52,640 --> 00:36:54,480 BUT I'VE HEARD IT SO MANY TIMES. 1132 00:36:54,480 --> 00:36:56,960 THE RATIONALE HERE IS THAT FROM 1133 00:36:56,960 --> 00:36:59,160 A PRACTICAL PERSPECTIVE, THE 1134 00:36:59,160 --> 00:36:59,800 MACHINE WOULD THEN GIVE THE 1135 00:36:59,800 --> 00:37:02,040 DOCTOR A NUMBER, JUST LIKE ANY 1136 00:37:02,040 --> 00:37:02,680 OTHER NUMBER. 1137 00:37:02,680 --> 00:37:04,640 BLOOD PRESSURE, ANY OTHER LAB 1138 00:37:04,640 --> 00:37:06,040 TEST, AND THE CLINICIAN WOULD 1139 00:37:06,040 --> 00:37:08,040 THEN DECIDE BASED ON THAT NUMBER 1140 00:37:08,040 --> 00:37:08,760 WHAT TO DO WITH IT. 1141 00:37:08,760 --> 00:37:10,080 AND IN FACT, THERE MAY BE 1142 00:37:10,080 --> 00:37:11,600 NUANCES THAT IF YOU GO FROM A 1143 00:37:11,600 --> 00:37:14,120 5 TO A 6, IT MAY NOT BE AS BAD 1144 00:37:14,120 --> 00:37:16,240 AS IF YOU GO FROM A 2 TO A 1145 00:37:16,240 --> 00:37:17,880 6 DURING THE SAME TIME PERIOD. 1146 00:37:17,880 --> 00:37:19,040 SO THERE MAY BE JUST INFORMATION 1147 00:37:19,040 --> 00:37:21,840 THAT YOU COULD THEN GIVE TO THE 1148 00:37:21,840 --> 00:37:23,480 DOCTOR TO MAKE THAT DIAGNOSIS 1149 00:37:23,480 --> 00:37:26,000 AND MAKE THAT MANAGEMENT PLAN. 1150 00:37:26,000 --> 00:37:27,760 SO I THINK THAT WE'RE REALLY 1151 00:37:27,760 --> 00:37:31,560 MOVING FROM A PARADIGM IN ALL OF 1152 00:37:31,560 --> 00:37:36,240 MEDICAL CARE FROM QUALITATIVE TO 1153 00:37:36,240 --> 00:37:36,840 QUANTITATIVE, AND WE'LL COME 1154 00:37:36,840 --> 00:37:38,000 BACK AND TALK MORE ABOUT THAT. 1155 00:37:38,000 --> 00:37:40,040 SO I'M GOING TO STOP THIS FIRST 1156 00:37:40,040 --> 00:37:42,600 SECTION OF THE TALK BY SAYING 1157 00:37:42,600 --> 00:37:43,880 JUST WHAT HAVE WE LEARNED ABOUT 1158 00:37:43,880 --> 00:37:44,680 THIS PARTICULAR DISEASE. 1159 00:37:44,680 --> 00:37:46,480 WELL, I WOULD ARGUE THAT THERE'S 1160 00:37:46,480 --> 00:37:48,160 SCIENCE, THERE'S ART AND THERE'S 1161 00:37:48,160 --> 00:37:48,480 TECHNOLOGY. 1162 00:37:48,480 --> 00:37:50,800 IN TERMS OF THE SCIENCE, ALL OF 1163 00:37:50,800 --> 00:37:52,920 THESE NIH FUNDED CLINICAL 1164 00:37:52,920 --> 00:37:54,320 TRIALS, THEY'VE REVOLUTIONIZED 1165 00:37:54,320 --> 00:37:57,000 CLINICAL CARE IN ROP BUT THEY 1166 00:37:57,000 --> 00:37:57,640 ARE NOT PERFECT. 1167 00:37:57,640 --> 00:37:59,400 FOR EXAMPLE, PLUS DISEASE 1168 00:37:59,400 --> 00:38:01,000 DIAGNOSIS, WHICH IS THE BASIS OF 1169 00:38:01,000 --> 00:38:04,080 THOSE TRIALS, IS SUBJECTIVE AND 1170 00:38:04,080 --> 00:38:04,720 QUALITATIVE. 1171 00:38:04,720 --> 00:38:06,240 THAT REALLY GETS INTO THE ART OF 1172 00:38:06,240 --> 00:38:07,640 MEDICINE. 1173 00:38:07,640 --> 00:38:08,320 THAT EVIDENCE-BASED PRACTICE 1174 00:38:08,320 --> 00:38:10,800 CANNOT CAPTURE ALL THAT 1175 00:38:10,800 --> 00:38:12,080 QUALITATIVE NUANCE OF DISEASE. 1176 00:38:12,080 --> 00:38:13,360 IN OTHER WORDS, WHAT IT MEANS TO 1177 00:38:13,360 --> 00:38:14,880 JUST LOOK BAD. 1178 00:38:14,880 --> 00:38:16,200 AND I THINK THAT TAKES US INTO 1179 00:38:16,200 --> 00:38:20,040 THE TECHNOLOGY, AND IN THIS 1180 00:38:20,040 --> 00:38:21,240 CASE, ARTIFICIAL INTELLIGENCE 1181 00:38:21,240 --> 00:38:23,080 METHODS CAN HELPFULLY HOPE 1182 00:38:23,080 --> 00:38:25,000 CLINICIANS ONE DAY TO IMPROVE 1183 00:38:25,000 --> 00:38:26,200 OBJECTIVITY OF DIAGNOSIS AND 1184 00:38:26,200 --> 00:38:27,480 MOVE FROM QUALITATIVE TO 1185 00:38:27,480 --> 00:38:31,480 QUANTITY ITIVE. 1186 00:38:31,480 --> 00:38:33,000 I HOPE THAT'S A GOOD PARADIGM 1187 00:38:33,000 --> 00:38:35,360 FOR ALL OF US AT NIH, THAT WE'RE 1188 00:38:35,360 --> 00:38:35,960 TRYING TO DO THINGS THAT ARE 1189 00:38:35,960 --> 00:38:37,520 GOING TO IMPROVE HUMAN HEALTH, 1190 00:38:37,520 --> 00:38:38,600 AND GOOD MEDICAL CARE IS GOING 1191 00:38:38,600 --> 00:38:40,400 TO REQUIRE ALL THREE OF THOSE 1192 00:38:40,400 --> 00:38:41,880 THINGS: SCIENCE, ART AND 1193 00:38:41,880 --> 00:38:42,280 TECHNOLOGY. 1194 00:38:42,280 --> 00:38:43,440 AND I HOPE THAT'S SOMETHING THAT 1195 00:38:43,440 --> 00:38:44,200 WE CAN JUST KEEP IN MIND IN 1196 00:38:44,200 --> 00:38:46,560 EVERYTHING THAT HE WITH WE . 1197 00:38:46,560 --> 00:38:49,640 WITH SO WITH THAT SAID, I WANT 1198 00:38:49,640 --> 00:38:50,680 TO MOVE INTO TALKING A LITTLE 1199 00:38:50,680 --> 00:38:52,240 BIT ABOUT WHAT'S MISSING, IN 1200 00:38:52,240 --> 00:38:55,320 OTHER WORDS, WHY HASN'T A.I. 1201 00:38:55,320 --> 00:38:59,680 REVOLUTIONIZED MEDICAL CARE. 1202 00:38:59,680 --> 00:39:01,280 I WOULD JUST TRY TO SUMMARIZE 1203 00:39:01,280 --> 00:39:02,560 SOME OF THE FEW UNANSWERED 1204 00:39:02,560 --> 00:39:03,520 RESEARCH QUESTIONS. 1205 00:39:03,520 --> 00:39:04,960 THIS ISN'T ALL OF THE UNANSWERED 1206 00:39:04,960 --> 00:39:06,200 QUESTIONS BUT IT'S JUST A FEW 1207 00:39:06,200 --> 00:39:07,280 THAT DOVETAIL WITH WHAT WE'VE 1208 00:39:07,280 --> 00:39:09,400 TALKED ABOUT UP UNTIL NOW. 1209 00:39:09,400 --> 00:39:14,640 ONE OF THEM IS THAT I WOULD SAY 1210 00:39:14,640 --> 00:39:16,840 THE PERFORMANCE OF AI SYSTEMS IN 1211 00:39:16,840 --> 00:39:17,720 OPHTHALMOLOGY, RADIOLOGY, 1212 00:39:17,720 --> 00:39:19,360 EVERYTHING ELSE IMPROVES AS YOUR 1213 00:39:19,360 --> 00:39:20,280 TASK NARROWS. 1214 00:39:20,280 --> 00:39:22,000 FOR EXAMPLE, IF YOU ASK A 1215 00:39:22,000 --> 00:39:24,160 SYSTEM, RULE OUT PLUS DISEASE IN 1216 00:39:24,160 --> 00:39:25,640 RETINOPATHY OF PREMATURITY, 1217 00:39:25,640 --> 00:39:28,240 THEY'RE REALLY, REALLY GOOD. 1218 00:39:28,240 --> 00:39:29,720 BUT THE PROBLEM IS THAT CLINICAL 1219 00:39:29,720 --> 00:39:31,320 DIAGNOSIS ISN'T JUST INDIVIDUAL 1220 00:39:31,320 --> 00:39:33,920 DISCRETE QUESTIONS LIKE THIS. 1221 00:39:33,920 --> 00:39:35,720 CLINICAL DIAGNOSIS INVOLVES 1222 00:39:35,720 --> 00:39:38,080 NUMEROUS PARALLEL TASKS. 1223 00:39:38,080 --> 00:39:40,080 PATIENTS HAVE MULTIPLE DISEASES, 1224 00:39:40,080 --> 00:39:41,600 MULTIPLE FINDINGS, AND YOU 1225 00:39:41,600 --> 00:39:43,040 TYPICALLY NEED TO INTEGRATE 1226 00:39:43,040 --> 00:39:44,480 IMAGE DATA TOGETHER WITH 1227 00:39:44,480 --> 00:39:47,640 CLINICAL DATA, AND SO IT'S NOT 1228 00:39:47,640 --> 00:39:48,960 ENOUGH JUST TO BE ASKING THESE 1229 00:39:48,960 --> 00:39:50,120 NARROW QUESTIONS, WE'VE GOT TO 1230 00:39:50,120 --> 00:39:53,280 FIND WAYS TO INTEGRATE THESE 1231 00:39:53,280 --> 00:39:54,560 DIAGNOSTIC SYSTEMS TOGETHER, AND 1232 00:39:54,560 --> 00:39:56,480 ALSO COME UP WITH HOW ARE THESE 1233 00:39:56,480 --> 00:39:58,200 SYSTEMS REALLY GOING TO HELP 1234 00:39:58,200 --> 00:39:59,280 DOCTORS WITH THE INFORMATION 1235 00:39:59,280 --> 00:40:00,920 THEY NEED. 1236 00:40:00,920 --> 00:40:04,720 SO BIG CHALLENGES IN THIS AREA. 1237 00:40:04,720 --> 00:40:05,640 ANOTHER CHALLENGE IS WHAT I 1238 00:40:05,640 --> 00:40:07,520 WOULD CALL GENERALIZABILITY AND 1239 00:40:07,520 --> 00:40:10,160 BIAS. 1240 00:40:10,160 --> 00:40:12,360 MANY OF THESE STUDIES ARE DONE 1241 00:40:12,360 --> 00:40:17,760 IN ACADEMIC CENTERS. 1242 00:40:17,760 --> 00:40:20,520 THE POPULATIONS ARE OFTEN 1243 00:40:20,520 --> 00:40:21,040 RELATIVELY HOMOGENEOUS. 1244 00:40:21,040 --> 00:40:22,480 I USED TO WORK IN PORTLAND, 1245 00:40:22,480 --> 00:40:23,760 OREGON WHERE THE POPULATION IS 1246 00:40:23,760 --> 00:40:25,280 JUST NOT THAT DIVERSE. 1247 00:40:25,280 --> 00:40:26,240 THE POINT OBVIOUSLY IS THAT THE 1248 00:40:26,240 --> 00:40:28,320 REAL WORLD IS HETEROGENEOUS. 1249 00:40:28,320 --> 00:40:29,160 THERE'S DIFFERENT RACES, 1250 00:40:29,160 --> 00:40:32,320 DIFFERENT POPULATIONS, DIFFERENT 1251 00:40:32,320 --> 00:40:34,720 IMAGING DEVICES. 1252 00:40:34,720 --> 00:40:36,160 SO IN GENERAL, YOU CANNOT TAKE 1253 00:40:36,160 --> 00:40:37,360 RESULTS FROM ONE MACHINE AND 1254 00:40:37,360 --> 00:40:39,120 THEN JUST EXTRAPOLATE TO ANOTHER 1255 00:40:39,120 --> 00:40:40,760 MACHINE OR TAKE ONE POPULATION 1256 00:40:40,760 --> 00:40:41,880 AND EXTRAPOLATE THOSE TO ANOTHER 1257 00:40:41,880 --> 00:40:45,440 POPULATION. 1258 00:40:45,440 --> 00:40:47,000 ON TOP OF THAT, REAL WORLD DATA 1259 00:40:47,000 --> 00:40:48,600 IS OFTEN MUCH POORER QUALITY 1260 00:40:48,600 --> 00:40:52,520 THAN DATA THAT COMES IN CLINICAL 1261 00:40:52,520 --> 00:40:52,880 RESEARCH. 1262 00:40:52,880 --> 00:40:55,800 IT'S SUBJECT TO BIASES IN DATA 1263 00:40:55,800 --> 00:40:56,200 COLLECTION. 1264 00:40:56,200 --> 00:40:57,320 IN FACT, THERE'S A PAPER 1265 00:40:57,320 --> 00:40:59,360 PUBLISHED THAT ONLY 6% OF DEEP 1266 00:40:59,360 --> 00:41:00,840 LEARNING STUDIES HAD ANY 1267 00:41:00,840 --> 00:41:02,560 EXTERNAL VALIDATION. 1268 00:41:02,560 --> 00:41:05,120 AND WE ALL SAW THIS IN THE COVID 1269 00:41:05,120 --> 00:41:06,920 ERA, WHERE ARTIFICIAL 1270 00:41:06,920 --> 00:41:07,960 INTELLIGENCE WAS SUPPOSED TO 1271 00:41:07,960 --> 00:41:09,680 SOLVE ALL OF OUR PROBLEMS WITH 1272 00:41:09,680 --> 00:41:11,920 COVID DIAGNOSIS AND PROGNOSIS, 1273 00:41:11,920 --> 00:41:18,440 AND IT REALLY DIDN'T. 1274 00:41:18,440 --> 00:41:20,920 SO HOW DO WE APPROACH THIS? 1275 00:41:20,920 --> 00:41:25,960 LARGER DATASETS, MORE DATASETS 1276 00:41:25,960 --> 00:41:28,120 AND VALIDATION ACROSS 1277 00:41:28,120 --> 00:41:31,040 POPULATIONS AND DEVICES, DATA 1278 00:41:31,040 --> 00:41:32,280 SHARING. 1279 00:41:32,280 --> 00:41:33,680 LASTLY, UNCLEAR GROUND TRUTH. 1280 00:41:33,680 --> 00:41:34,920 WE TALKED ABOUT THIS A LOT WITH 1281 00:41:34,920 --> 00:41:36,480 PLUS DISEASE DIAGNOSIS AND IT'S 1282 00:41:36,480 --> 00:41:38,720 TRUE FOR EVERY AREA IN MEDICINE. 1283 00:41:38,720 --> 00:41:40,680 THERE'S A SCIENCE AND THERE'S AN 1284 00:41:40,680 --> 00:41:41,800 ART. 1285 00:41:41,800 --> 00:41:45,720 AND THERE'S DIAGNOSTIC 1286 00:41:45,720 --> 00:41:46,280 VARIABILITY. 1287 00:41:46,280 --> 00:41:47,560 SOMETIMES THERE'S JUST GREY 1288 00:41:47,560 --> 00:41:48,760 ZONES, LIKE THE LINE BETWEEN 1289 00:41:48,760 --> 00:41:50,120 MODERATE DISEASE AND MILD 1290 00:41:50,120 --> 00:41:51,520 DISEASE CAN JUST BE A LITTLE BIT 1291 00:41:51,520 --> 00:41:54,400 BLURRY SOMETIMES. 1292 00:41:54,400 --> 00:41:56,960 SO ON TOP OF THAT, THERE ARE AFN 1293 00:41:56,960 --> 00:41:59,280 VERY 1294 00:41:59,280 --> 00:42:01,040 OFTEN VERY UNCLEAR DEFINITIONS 1295 00:42:01,040 --> 00:42:03,440 OF WHAT I WOULD CALL WHAT DOES 1296 00:42:03,440 --> 00:42:05,080 IT MEAN TO HAVE MODERATE 1297 00:42:05,080 --> 00:42:08,080 DISEASE, SEVERE DISEASE, 1298 00:42:08,080 --> 00:42:09,080 REGARDLESS OF WHAT DISEASE IT 1299 00:42:09,080 --> 00:42:10,360 IS, AND THAT GETS INTO THAT 1300 00:42:10,360 --> 00:42:12,200 WHOLE ISSUE OF DYING DIAGNOSC 1301 00:42:12,200 --> 00:42:12,840 PROCESS. 1302 00:42:12,840 --> 00:42:14,080 DIFFERENT PEOPLE MAY HAVE 1303 00:42:14,080 --> 00:42:15,080 DIFFERENT DIAGNOSTIC PROCESSES. 1304 00:42:15,080 --> 00:42:16,480 SO I THINK THERE'S A LOT OF WORK 1305 00:42:16,480 --> 00:42:18,520 THAT COMMUNITY NEEDS TO DO IN 1306 00:42:18,520 --> 00:42:20,760 TERMS OF CONSENSUS DEFINITIONS 1307 00:42:20,760 --> 00:42:22,160 FOR DISEASE, AND FOR EXAMPLE, 1308 00:42:22,160 --> 00:42:24,360 HOW ARE WE GOING TO CROWD SOURCE 1309 00:42:24,360 --> 00:42:25,920 DIFFERENT GRADINGS IF YOU NEED 1310 00:42:25,920 --> 00:42:28,400 TO HAVE MULTIPLE PEOPLE MAKING 1311 00:42:28,400 --> 00:42:32,440 DIAGNOSES TO CREATE THAT GROUND 1312 00:42:32,440 --> 00:42:37,040 TRUTH. 1313 00:42:37,040 --> 00:42:37,920 SO A LOT OF DIFFERENT 1314 00:42:37,920 --> 00:42:38,440 CHALLENGES. 1315 00:42:38,440 --> 00:42:39,760 WITH THOSE CHALLENGES I WANT TO 1316 00:42:39,760 --> 00:42:40,760 SPEND THE NEXT FEW MINUTES 1317 00:42:40,760 --> 00:42:43,480 TALKING ABOUT WHAT ARE WE DOING 1318 00:42:43,480 --> 00:42:45,160 AT THE NATIONAL EYE INSTITUTE TO 1319 00:42:45,160 --> 00:42:45,880 ADDRESS THOSE CHALLENGES AND 1320 00:42:45,880 --> 00:42:47,320 WHAT ARE WE DOING NIH WIDE IN 1321 00:42:47,320 --> 00:42:48,160 SOME OF THOSE AREAS. 1322 00:42:48,160 --> 00:42:49,600 SO I THINK ONE OF THE THINGS 1323 00:42:49,600 --> 00:42:51,760 THAT WE NEED IS WHAT I WOULD 1324 00:42:51,760 --> 00:42:52,560 CALL DATA SHARING AND 1325 00:42:52,560 --> 00:42:54,760 HARMONIZATION. 1326 00:42:54,760 --> 00:42:58,480 AND THE PREMISE FOR THIS IS THAT 1327 00:42:58,480 --> 00:43:00,240 MODERN RESEARCH RELIES ON BIG 1328 00:43:00,240 --> 00:43:00,800 DATA. 1329 00:43:00,800 --> 00:43:01,960 IT DOESN'T MATTER IF YOU'RE 1330 00:43:01,960 --> 00:43:04,400 DOING ARTIFICIAL INTELLIGENCE 1331 00:43:04,400 --> 00:43:05,880 RESEARCH, MECHANISTIC RESEARCH 1332 00:43:05,880 --> 00:43:07,960 INVOLVING GENETICS AND OMICS AND 1333 00:43:07,960 --> 00:43:09,480 DISEASE MECHANISMS. 1334 00:43:09,480 --> 00:43:10,880 YOU KNOW, I THINK THAT YOU NEED 1335 00:43:10,880 --> 00:43:12,520 LARGE DATASETS AND THAT'S GOING 1336 00:43:12,520 --> 00:43:13,680 TO REQUIRE DATA SHARING. 1337 00:43:13,680 --> 00:43:15,680 WE'VE GOT A REALLY, REALLY 1338 00:43:15,680 --> 00:43:16,920 COMMON PARADIGM IN THIS COUNTRY 1339 00:43:16,920 --> 00:43:18,600 WHERE THERE ARE MULTIPLE CENTERS 1340 00:43:18,600 --> 00:43:20,400 AROUND THE COUNTRY WHERE PEOPLE 1341 00:43:20,400 --> 00:43:22,640 ARE WORKING ON THE EXACT SAME 1342 00:43:22,640 --> 00:43:23,880 PROBLEM BUT THOSE STUDIES ARE 1343 00:43:23,880 --> 00:43:26,120 ALL INDIVIDUALLY UNDERPOWERED TO 1344 00:43:26,120 --> 00:43:28,880 TRULY ANSWER THAT QUESTION AND 1345 00:43:28,880 --> 00:43:30,160 MOVE THE FIELD FORWARD. 1346 00:43:30,160 --> 00:43:33,800 IN FACT, NUMBER ONE, THOSE 1347 00:43:33,800 --> 00:43:35,680 INVESTIGATORS DON'T ALWAYS FEEL 1348 00:43:35,680 --> 00:43:37,000 INCENTIVIZED TO SHARE DATA, IN 1349 00:43:37,000 --> 00:43:39,240 FACT, THEY MAY FEEL 1350 00:43:39,240 --> 00:43:40,120 DISINCENTIVIZED TO SHARE DATA. 1351 00:43:40,120 --> 00:43:41,360 BUT EVEN IF THEY WANTED TO 1352 00:43:41,360 --> 00:43:43,120 SHARE, IN GENERAL, THEY CAN'T 1353 00:43:43,120 --> 00:43:45,160 JUST SHARE OUTSIDE THE BOX, 1354 00:43:45,160 --> 00:43:47,480 BECAUSE THOSE DATA WERE ALMOST 1355 00:43:47,480 --> 00:43:49,000 ALWAYS COLLECTED DIFFERENTLY AND 1356 00:43:49,000 --> 00:43:50,400 USING DIFFERENT DATA 1357 00:43:50,400 --> 00:43:51,280 REPRESENTATIONS. 1358 00:43:51,280 --> 00:43:55,280 SO I THINK THAT WE'RE LOSING A 1359 00:43:55,280 --> 00:43:56,520 LOT OF OPPORTUNITIES TO ADVANCE 1360 00:43:56,520 --> 00:43:57,640 THE FIELD BECAUSE OF LACK OF 1361 00:43:57,640 --> 00:43:59,040 DATA SHARING AND LACK OF 1362 00:43:59,040 --> 00:43:59,760 HARMONIZATION. 1363 00:43:59,760 --> 00:44:03,200 NOW, AT THE NIH, ALMOST ALL OF 1364 00:44:03,200 --> 00:44:04,880 YOU KNOW THAT WE'VE GOT A DATA 1365 00:44:04,880 --> 00:44:05,960 SHARING POLICY THAT'S GOING TO 1366 00:44:05,960 --> 00:44:08,000 GO INTO EFFECT NEXT YEAR AND 1367 00:44:08,000 --> 00:44:08,440 THAT'S A STICK. 1368 00:44:08,440 --> 00:44:09,800 WE'RE NOT GOING TO FUND YOUR 1369 00:44:09,800 --> 00:44:14,840 RESEARCH UNLESS YOU SHARE DATA. 1370 00:44:14,840 --> 00:44:16,600 I AND MANY PEOPLE AT NIH BELIEVE 1371 00:44:16,600 --> 00:44:18,640 WE REALLY NEED INCENTIVES FOR 1372 00:44:18,640 --> 00:44:20,200 DATA SHARING TO TRULY MOVE THE 1373 00:44:20,200 --> 00:44:24,080 FIELD FORWARD. 1374 00:44:24,080 --> 00:44:25,280 ONE OF THE THINGS WE DID IN MY 1375 00:44:25,280 --> 00:44:27,520 PART OF THE WORLD IS THAT I 1376 00:44:27,520 --> 00:44:31,720 WORKED WITH SOMEONE, MARCO 1377 00:44:31,720 --> 00:44:33,080 ZARBAN, EDITOR-IN-CHIEF OF ONE 1378 00:44:33,080 --> 00:44:34,120 OF THE MAJOR VISION JOURNALS 1379 00:44:34,120 --> 00:44:36,640 TOGETHER WITH TWO PEOPLE ON THE 1380 00:44:36,640 --> 00:44:40,320 EDITORIAL BOARD, TO CREATE A NEW 1381 00:44:40,320 --> 00:44:42,600 PUBLICATION TYPE CALLED A 1382 00:44:42,600 --> 00:44:43,800 DATASET/SOFTWARE LIBRARY. 1383 00:44:43,800 --> 00:44:45,480 AND THE POINT IS THAT YOU CAN 1384 00:44:45,480 --> 00:44:46,640 WRITE A PAPER THAT'S PURELY 1385 00:44:46,640 --> 00:44:49,040 BASED ON YOUR DATASET, AND 1386 00:44:49,040 --> 00:44:50,240 HOPEFULLY THAT WOULD BE ONE OF 1387 00:44:50,240 --> 00:44:51,640 THOSE THINGS THAT COULD 1388 00:44:51,640 --> 00:44:53,520 INCENTIVIZE DATA SHARING AND 1389 00:44:53,520 --> 00:44:55,520 TEAM SCIENCE WITHIN THE VISION 1390 00:44:55,520 --> 00:44:56,480 COMMUNITY. 1391 00:44:56,480 --> 00:44:58,120 SO IN OTHER WORDS, PEOPLE WOULD 1392 00:44:58,120 --> 00:44:59,560 GET ACADEMIC CREDIT, THEIR 1393 00:44:59,560 --> 00:45:00,920 DATASET WOULD BE CITED AND IT 1394 00:45:00,920 --> 00:45:03,400 WOULD BE FINDABLE USING PUBMED 1395 00:45:03,400 --> 00:45:03,640 SEARCH. 1396 00:45:03,640 --> 00:45:06,160 SO I THINK WE'RE GOING TO NEED 1397 00:45:06,160 --> 00:45:09,160 MORE OF THOSE CARROTS. 1398 00:45:09,160 --> 00:45:13,040 ALONG THE LINES OF DATA 1399 00:45:13,040 --> 00:45:14,000 HARMONIZATION, THINK WE NEED A 1400 00:45:14,000 --> 00:45:15,640 COMMON DATA MODEL FOR 1401 00:45:15,640 --> 00:45:16,680 OPHTHALMOLOGY, THAT'S PRETTY 1402 00:45:16,680 --> 00:45:16,880 CLEAR. 1403 00:45:16,880 --> 00:45:18,520 AND WE AT THE NEI HAVE STARTED 1404 00:45:18,520 --> 00:45:20,600 WORKING WITH THE AMERICAN 1405 00:45:20,600 --> 00:45:21,560 ACADEMY OF OPHTHALMOLOGY TO TRY 1406 00:45:21,560 --> 00:45:23,200 TO DEVELOP THAT DATA MODEL. 1407 00:45:23,200 --> 00:45:24,600 AND THERE'S SOME FOLKS HERE WHO 1408 00:45:24,600 --> 00:45:26,840 ARE DOING THE WORK, SALLY BAXTER 1409 00:45:26,840 --> 00:45:33,040 FROM UC SAN DIEGO, KERRY GETS GZ 1410 00:45:33,040 --> 00:45:34,520 AND A WHOLE OTHER GROUP OF 1411 00:45:34,520 --> 00:45:36,320 PEOPLE WHO'S WORKING ON THIS. 1412 00:45:36,320 --> 00:45:37,720 SO DATA SHARING AND HARM 1413 00:45:37,720 --> 00:45:38,840 ?IIZATION, I THINK REALLY, 1414 00:45:38,840 --> 00:45:40,440 REALLY IMPORTANT FOR THE FUTURE 1415 00:45:40,440 --> 00:45:43,400 OF CLINICAL RESEARCH AND JUST 1416 00:45:43,400 --> 00:45:44,360 RESEARCH IN GENERAL. 1417 00:45:44,360 --> 00:45:45,960 ANOTHER THING THAT I THINK WE'RE 1418 00:45:45,960 --> 00:45:50,720 REALLY GOING TO NEED IS IMAGING, 1419 00:45:50,720 --> 00:45:52,480 IMAGE STANDARDS. 1420 00:45:52,480 --> 00:45:54,720 NOW, IMAGING WE TALKED ABOUT AS 1421 00:45:54,720 --> 00:45:57,280 AVAILABILITY OF THESE IMAGES IS 1422 00:45:57,280 --> 00:46:00,800 ONE REASON THERE'S BEEN SUCH A 1423 00:46:00,800 --> 00:46:02,320 REVOLUTION IN A.I. AND VISION 1424 00:46:02,320 --> 00:46:02,880 SCIENCE. 1425 00:46:02,880 --> 00:46:04,280 THESE IMAGES COME FROM MACHINES 1426 00:46:04,280 --> 00:46:05,600 LIKE THIS. 1427 00:46:05,600 --> 00:46:08,200 AND THEY GENERATE PICTURES AND 1428 00:46:08,200 --> 00:46:08,440 NUMBERS. 1429 00:46:08,440 --> 00:46:10,640 THIS IS A GOLD MINE FOR 1430 00:46:10,640 --> 00:46:11,200 ARTIFICIAL INTELLIGENCE AND 1431 00:46:11,200 --> 00:46:14,240 OTHER SORT OF CLINICAL RESEARCH 1432 00:46:14,240 --> 00:46:14,800 IN GENERAL. 1433 00:46:14,800 --> 00:46:17,080 THE PROBLEM IS THAT IN GENERAL, 1434 00:46:17,080 --> 00:46:18,760 YOU CANNOT GET DATA OUT OF THESE 1435 00:46:18,760 --> 00:46:21,360 MACHINES EASILY AT ALL, OR AT 1436 00:46:21,360 --> 00:46:23,040 ALL BECAUSE THEY'RE LOCKED IN 1437 00:46:23,040 --> 00:46:24,760 PROPRIETARY STANDARDS THAT ARE 1438 00:46:24,760 --> 00:46:28,400 DEFINED AND REALLY OWNED BY THE 1439 00:46:28,400 --> 00:46:28,640 VENDORS. 1440 00:46:28,640 --> 00:46:30,000 SO THIS IS A HUGE PROBLEM. 1441 00:46:30,000 --> 00:46:31,720 IT'S A HUGE PROBLEM FOR RESEARCH 1442 00:46:31,720 --> 00:46:33,280 AND IT'S A HUGE PROBLEM FOR 1443 00:46:33,280 --> 00:46:35,480 CLINICAL CARE. 1444 00:46:35,480 --> 00:46:36,800 WHEN I WAS IN OREGON, I WORKED 1445 00:46:36,800 --> 00:46:39,600 WITH A PH.D. STUDENT WHO WANTED 1446 00:46:39,600 --> 00:46:41,160 TO DO MACHINE LEARNING FOR HIS 1447 00:46:41,160 --> 00:46:42,920 PH.D. THESIS, AND THE MACHINE 1448 00:46:42,920 --> 00:46:43,800 LEARNING DATA WAS GOING TO COME 1449 00:46:43,800 --> 00:46:44,920 OUT OF MACHINES LIKE THIS. 1450 00:46:44,920 --> 00:46:46,920 HE SPENT A YEAR, COULDN'T GET 1451 00:46:46,920 --> 00:46:48,360 ACCESS TO THE DATA, AND SO JUST 1452 00:46:48,360 --> 00:46:50,920 MOVED ON TO SOMETHING ELSE. 1453 00:46:50,920 --> 00:46:54,960 SO BIG, BIG PROBLEM. 1454 00:46:54,960 --> 00:46:55,720 NOW, I'VE REALLY BEEN TRYING TO 1455 00:46:55,720 --> 00:46:57,800 SOLVE THIS PROBLEM, AND YOU 1456 00:46:57,800 --> 00:47:00,640 KNOW, WE'VE WORKED OVER THE PAST 1457 00:47:00,640 --> 00:47:01,840 NINE MONTHS WITH A GROUP FROM 1458 00:47:01,840 --> 00:47:04,640 THE ONC, THE OFFICE OF NATIONAL 1459 00:47:04,640 --> 00:47:06,080 COORDINATOR FOR HEALTH I.T., AND 1460 00:47:06,080 --> 00:47:10,360 THAT'S LED BY MICKEY TROPOTHY. 1461 00:47:10,360 --> 00:47:12,160 WHY THE ONC? 1462 00:47:12,160 --> 00:47:15,800 THEY HAVE PURVIEW OVER 1463 00:47:15,800 --> 00:47:16,480 CERTIFYING DATA EXCHANGE. 1464 00:47:16,480 --> 00:47:18,680 WE'VE ALSO WORKED WITH THE FDA, 1465 00:47:18,680 --> 00:47:25,960 MEL VENA EDELMAN, AND WHY THE 1466 00:47:25,960 --> 00:47:27,440 FDA, BECAUSE THEY REGULATE 1467 00:47:27,440 --> 00:47:28,880 DEVICES, WHY US, BECAUSE WE 1468 00:47:28,880 --> 00:47:29,440 REGULATE GRANTS BASICALLY. 1469 00:47:29,440 --> 00:47:32,280 SO WHAT WE'VE BEEN TRYING TO GO 1470 00:47:32,280 --> 00:47:33,840 THROUGH IS WHAT CAN WE DO AND 1471 00:47:33,840 --> 00:47:35,600 WHAT ARE OUR LEVERS IN THE 1472 00:47:35,600 --> 00:47:38,080 FEDERAL GOVERNMENT ABOUT HOW WE 1473 00:47:38,080 --> 00:47:40,040 CAN MOVE THE FIELD FORWARD. 1474 00:47:40,040 --> 00:47:42,320 AND IT MAY BE A COMBINATION OF 1475 00:47:42,320 --> 00:47:44,480 STICKS AND CARROTS. 1476 00:47:44,480 --> 00:47:47,720 WE HAD A WORKSHOP ABOUT TWO 1477 00:47:47,720 --> 00:47:49,000 WEEKS AGO ABOUT THIS, AND IT WAS 1478 00:47:49,000 --> 00:47:51,440 REALLY, REALLY PRODUCTIVE, AND 1479 00:47:51,440 --> 00:47:52,160 HOPEFULLY WE'LL LEARN SOMETHING 1480 00:47:52,160 --> 00:47:54,800 THAT WILL HELP OUR COMMUNITY IN 1481 00:47:54,800 --> 00:47:56,000 VISION RESEARCH AS WELL AS OTHER 1482 00:47:56,000 --> 00:47:56,760 COMMUNITIES, BECAUSE THERE ARE 1483 00:47:56,760 --> 00:47:57,760 MANY, MANY COMMUNITIES AT NIH 1484 00:47:57,760 --> 00:47:59,760 AND AROUND THE COUNTRY THAT DEAL 1485 00:47:59,760 --> 00:48:02,000 WITH THIS EXACT PROBLEM WITH 1486 00:48:02,000 --> 00:48:03,760 IMAGING AND PROPRIETARY 1487 00:48:03,760 --> 00:48:07,760 STANDARDS. 1488 00:48:07,760 --> 00:48:09,720 SO IMAGING STANDARDS, REALLY 1489 00:48:09,720 --> 00:48:10,280 IMPORTANT. 1490 00:48:10,280 --> 00:48:12,120 ANOTHER THING I THINK WE NEED IS 1491 00:48:12,120 --> 00:48:14,520 JUST LARGER A.I.-READY DATASETS. 1492 00:48:14,520 --> 00:48:15,400 I JUST WANT TO MAKE SURE THAT 1493 00:48:15,400 --> 00:48:17,160 EVERYBODY HERE IS AWARE OF TWO 1494 00:48:17,160 --> 00:48:20,920 DIFFERENT PROGRAMS THAT ARE 1495 00:48:20,920 --> 00:48:21,160 TRANS-NIH. 1496 00:48:21,160 --> 00:48:24,800 ONE OF THEM IS A COMMON FUND 1497 00:48:24,800 --> 00:48:25,600 BRIDGE2AI. 1498 00:48:25,600 --> 00:48:27,560 THERE'S FIVE OF THESE I.C.s 1499 00:48:27,560 --> 00:48:28,960 THAT REALLY HELPED TO DEVELOP 1500 00:48:28,960 --> 00:48:30,080 THAT ORIGINAL PROJECT, BUT IT'S 1501 00:48:30,080 --> 00:48:32,240 GOING TO BE A $100 MILLION 1502 00:48:32,240 --> 00:48:35,400 PROJECT OVER FOUR YEARS, AND 1503 00:48:35,400 --> 00:48:36,880 ALMOST ALL THE WORK THAT WE'VE 1504 00:48:36,880 --> 00:48:38,800 DONE HISTORICALLY AT NIH HAS 1505 00:48:38,800 --> 00:48:40,840 BEEN WHAT I'D CALL 1506 00:48:40,840 --> 00:48:42,920 HYPOTHESIS-DRIVEN RESEARCH. 1507 00:48:42,920 --> 00:48:45,640 BUT BRIDGE 2AI IS DIFFERENT. 1508 00:48:45,640 --> 00:48:47,760 THE FUNDING HERE IS TO BUILD 1509 00:48:47,760 --> 00:48:49,840 DATASETS FROM MULTIPLE SOURCES. 1510 00:48:49,840 --> 00:48:52,440 SO IT'S IN A WAY 1511 00:48:52,440 --> 00:48:53,760 HYPOTHESIS-GENERATING. 1512 00:48:53,760 --> 00:48:55,120 A REALLY, REALLY DIFFERENT 1513 00:48:55,120 --> 00:48:55,720 PARADIGM FOR NIH. 1514 00:48:55,720 --> 00:48:57,800 SO HOPEFULLY ANNOUNCEMENTS ABOUT 1515 00:48:57,800 --> 00:49:01,960 THIS COMING UP RELATIVELY SOON. 1516 00:49:01,960 --> 00:49:03,000 ANOTHER PROJECT THAT I WANT TO 1517 00:49:03,000 --> 00:49:05,360 MAKE SURE PEOPLE ARE AWARE OF IS 1518 00:49:05,360 --> 00:49:06,440 AIM AHEAD. 1519 00:49:06,440 --> 00:49:10,080 AIM AHEAD IS RUN THROUGH THE 1520 00:49:10,080 --> 00:49:11,360 OFFICE OF DATA SCIENCE STRATEGY 1521 00:49:11,360 --> 00:49:17,040 THAT'S LED BY SUSAN GREGURICK, 1522 00:49:17,040 --> 00:49:18,480 AND JAMES, OUR PROGRAM OFFICER 1523 00:49:18,480 --> 00:49:20,320 INVOLVED WITH THIS PROJECT. 1524 00:49:20,320 --> 00:49:23,080 THERE ARE OBVIOUSLY MANY PROJECT 1525 00:49:23,080 --> 00:49:25,920 OFFICERS INVOLVED, BUT THE GOAL 1526 00:49:25,920 --> 00:49:27,560 OF BRIDGE2AI IS TO USE 1527 00:49:27,560 --> 00:49:28,320 ARTIFICIAL INTELLIGENCE TO 1528 00:49:28,320 --> 00:49:32,800 ADDRESS CHALLENGES OF HEALTH 1529 00:49:32,800 --> 00:49:34,720 DISPARITIES IN MINORITY HEALTH 1530 00:49:34,720 --> 00:49:35,480 IN THIS COUNTRY. 1531 00:49:35,480 --> 00:49:36,680 THERE'S MANY WAYS OF DOING THAT. 1532 00:49:36,680 --> 00:49:37,760 THE ONLY FEW EXAMPLES THAT I'LL 1533 00:49:37,760 --> 00:49:39,920 PUT ON THIS SLIDE ARE HOW MIGHT 1534 00:49:39,920 --> 00:49:40,960 YOU DO THAT? 1535 00:49:40,960 --> 00:49:42,760 WELL, NUMBER ONE, DEVELOPING 1536 00:49:42,760 --> 00:49:45,000 DIVERSE DATASETS THAT TRULY 1537 00:49:45,000 --> 00:49:46,600 REPRESENT THE AMERICAN 1538 00:49:46,600 --> 00:49:47,160 POPULATION. 1539 00:49:47,160 --> 00:49:50,560 NUMBER TWO, HELPING TO TRAIN A 1540 00:49:50,560 --> 00:49:50,960 DIVERSE WORKFORCE. 1541 00:49:50,960 --> 00:49:52,040 MANY OF THE COMMUNITY HEALTH 1542 00:49:52,040 --> 00:49:53,880 CLINICS THAT TAKE CARE OF 1543 00:49:53,880 --> 00:49:56,080 PATIENTS WHO ARE IN MEDICALLY 1544 00:49:56,080 --> 00:49:58,080 UNDERSERVED AREAS MAY NOT EVEN 1545 00:49:58,080 --> 00:49:59,080 HAVE ELECTRONIC HEALTH RECORD 1546 00:49:59,080 --> 00:50:01,320 SYSTEMS NOW, LET ALONE USING 1547 00:50:01,320 --> 00:50:03,840 ARTIFICIAL INTELLIGENCE. 1548 00:50:03,840 --> 00:50:06,560 SO ANOTHER WAY IS BY DETECTING 1549 00:50:06,560 --> 00:50:09,080 A.I. APPROACHES TO DETECT BIAS. 1550 00:50:09,080 --> 00:50:10,840 WE DEFINITELY DON'T WANT A WORLD 1551 00:50:10,840 --> 00:50:14,960 WHERE A.I. SYSTEMS REINFORCE 1552 00:50:14,960 --> 00:50:15,160 BIAS. 1553 00:50:15,160 --> 00:50:16,640 NOW JUST A WORD ABOUT BIAS. 1554 00:50:16,640 --> 00:50:17,680 BIAS CAN GET INTO THE MEDICAL 1555 00:50:17,680 --> 00:50:18,760 RECORD IN MANY DIFFERENT WAYS 1556 00:50:18,760 --> 00:50:20,920 AND ONE OF THEM IS DOCTORS AND 1557 00:50:20,920 --> 00:50:22,640 NURSES MAY JUST WRITE 1558 00:50:22,640 --> 00:50:23,840 OBSERVATIONS IN CHARTS THAT ARE 1559 00:50:23,840 --> 00:50:25,680 PURELY SUBJECTIVE. 1560 00:50:25,680 --> 00:50:29,880 YOU KNOW, "MY PATIENT WAS 1561 00:50:29,880 --> 00:50:30,200 NONCOMPLIANT." 1562 00:50:30,200 --> 00:50:31,240 WELL, WHAT DOES THAT EVEN MEAN? 1563 00:50:31,240 --> 00:50:36,160 SO I THINK IT'S VERY HARD TO 1564 00:50:36,160 --> 00:50:37,760 STUDY BIAS AND UNDERSTAND THE 1565 00:50:37,760 --> 00:50:39,000 SOURCES OF BIAS WHEN YOU'RE 1566 00:50:39,000 --> 00:50:41,200 LOOKING AT DATA THAT MAY BE 1567 00:50:41,200 --> 00:50:42,560 INHERENTLY BIASED AND HOW DO YOU 1568 00:50:42,560 --> 00:50:43,240 KNOW WHAT'S TRUE? 1569 00:50:43,240 --> 00:50:46,800 SO I HOPE THAT THINGS LIKE 1570 00:50:46,800 --> 00:50:48,160 IMAGES ARE GOING TO BE ONE WAY 1571 00:50:48,160 --> 00:50:49,480 TO PROVIDE OBJECTIVE DATA THAT 1572 00:50:49,480 --> 00:50:51,560 CAN REALLY ANCHOR THAT AND 1573 00:50:51,560 --> 00:50:53,000 PROVIDE MEANINGFUL WAYS TO STUDY 1574 00:50:53,000 --> 00:50:54,120 BIAS AND FIGURE OUT HOW CAN WE 1575 00:50:54,120 --> 00:50:55,040 DO A BETTER JOB. 1576 00:50:55,040 --> 00:50:57,440 SO OVERALL, WE NEED LARGER 1577 00:50:57,440 --> 00:50:59,800 DATASETS, A.I.-READY DATASETS. 1578 00:50:59,800 --> 00:51:00,760 THE LAST THING THAT I THINK THAT 1579 00:51:00,760 --> 00:51:04,520 WE NEED IS MEDICAL EDUCATION IN 1580 00:51:04,520 --> 00:51:07,440 ARTIFICIAL INTEL JESS INTA 1581 00:51:07,440 --> 00:51:07,680 SCIENCE. 1582 00:51:07,680 --> 00:51:09,040 I ASK THE QUESTION, WHAT'S THE 1583 00:51:09,040 --> 00:51:10,400 ONE TOOL USED BY EVERY SINGLE 1584 00:51:10,400 --> 00:51:12,040 DOCTOR AND SCIENTIST IN THIS 1585 00:51:12,040 --> 00:51:12,280 COUNTRY? 1586 00:51:12,280 --> 00:51:14,120 IT'S NOT A STETHOSCOPE, IT'S NOT 1587 00:51:14,120 --> 00:51:17,400 EVEN AN OPHTHALMOSCOPE IF YOU'RE 1588 00:51:17,400 --> 00:51:17,960 AN OPHTHALMOLOGIST. 1589 00:51:17,960 --> 00:51:19,440 I WOULD ARGUE THAT IT'S THE 1590 00:51:19,440 --> 00:51:20,920 COMPUTER, AND IT'S COMPUTING 1591 00:51:20,920 --> 00:51:21,640 TECHNOLOGY. 1592 00:51:21,640 --> 00:51:24,520 AND DESPITE THAT, THERE IS 1593 00:51:24,520 --> 00:51:26,240 BASICALLY NO SYSTEMATIC TRAINING 1594 00:51:26,240 --> 00:51:27,040 OR FRAMEWORK THAT PEOPLE GET 1595 00:51:27,040 --> 00:51:28,240 ABOUT WHAT THE IMPLICATIONS ARE 1596 00:51:28,240 --> 00:51:31,600 OF THIS TECHNOLOGY OR HOW TO USE 1597 00:51:31,600 --> 00:51:31,960 IT. 1598 00:51:31,960 --> 00:51:33,440 I WOULD COMPARE THAT TO MEDICAL 1599 00:51:33,440 --> 00:51:36,680 STUDENTS WILL SPEND MONTHS TO 1600 00:51:36,680 --> 00:51:38,240 YEARS OF FOCUSED DEDICATED TIME 1601 00:51:38,240 --> 00:51:39,880 LEARNING HOW TO TAKE A HISTORY, 1602 00:51:39,880 --> 00:51:41,080 HOW TO DO A PHYSICAL 1603 00:51:41,080 --> 00:51:42,640 EXAMINATION, HOW TO AUSCULTATE 1604 00:51:42,640 --> 00:51:43,040 THE HEART. 1605 00:51:43,040 --> 00:51:44,600 YET FOR THE ONE TOOL THAT EVERY 1606 00:51:44,600 --> 00:51:47,680 SINGLE PERSON USES, TYPICALLY 1607 00:51:47,680 --> 00:51:48,920 IT'S ZERO TIME THAT'S DEDICATED 1608 00:51:48,920 --> 00:51:49,920 TO LEARNING HOW TO DO IT AND 1609 00:51:49,920 --> 00:51:51,520 PEOPLE LEARN BY TRIAL AND ERROR 1610 00:51:51,520 --> 00:51:52,880 BASICALLY. 1611 00:51:52,880 --> 00:51:54,480 AND YOU KNOW, THIS IS 1612 00:51:54,480 --> 00:51:55,200 FUNDAMENTALLY CHANGING THE 1613 00:51:55,200 --> 00:51:58,160 PRACTICE AND THE SCIENCE OF 1614 00:51:58,160 --> 00:51:58,800 MEDICINE. 1615 00:51:58,800 --> 00:52:00,040 THE OTHER THING I WOULD POINT 1616 00:52:00,040 --> 00:52:01,360 OUT IS THAT THERE'S A PAIR TIME 1617 00:52:01,360 --> 00:52:02,440 IN EDUCATION WHERE THE PERSON 1618 00:52:02,440 --> 00:52:03,840 WITH MORE GREY HAIR TEACHES THE 1619 00:52:03,840 --> 00:52:05,680 PERSON WITH LESS GREY HAIR HOW 1620 00:52:05,680 --> 00:52:08,160 TO DO THINGS, YET IN THIS CASE, 1621 00:52:08,160 --> 00:52:09,440 IT'S THE PERSON WITH MORE GREY 1622 00:52:09,440 --> 00:52:11,080 HAIR WHO KNOWS LESS ABOUT THE 1623 00:52:11,080 --> 00:52:11,880 TECHNOLOGY. 1624 00:52:11,880 --> 00:52:13,520 AND SO WE'VE GOT TO FIGURE OUT 1625 00:52:13,520 --> 00:52:14,480 WAYS TO DO A BETTER JOB WITH 1626 00:52:14,480 --> 00:52:18,240 THIS. 1627 00:52:18,240 --> 00:52:20,120 SO I'M GOING TO CLOSE UP THIS 1628 00:52:20,120 --> 00:52:20,960 SECTION AND BASICALLY SAY WHAT 1629 00:52:20,960 --> 00:52:22,120 HAVE I LEARNED ABOUT THE FUTURE 1630 00:52:22,120 --> 00:52:22,920 OF MEDICINE? 1631 00:52:22,920 --> 00:52:24,040 WELL, I WOULD SAY THAT WE'VE 1632 00:52:24,040 --> 00:52:26,200 TALKED ABOUT THIS THEME OF GOING 1633 00:52:26,200 --> 00:52:27,600 FROM QUALITATIVE TO 1634 00:52:27,600 --> 00:52:28,200 QUANTITATIVE. 1635 00:52:28,200 --> 00:52:28,880 RIGHT? 1636 00:52:28,880 --> 00:52:31,000 AND THAT'S TRUE IN ARTIFICIAL 1637 00:52:31,000 --> 00:52:32,520 INTELLIGENCE, IT'S TRUE IN 1638 00:52:32,520 --> 00:52:34,440 MEDICAL IMAGING AND 1639 00:52:34,440 --> 00:52:35,000 EVIDENCE-BASED PRACTICE 1640 00:52:35,000 --> 00:52:36,440 GUIDELINES. 1641 00:52:36,440 --> 00:52:37,520 I ALSO WILL THINK ABOUT 1642 00:52:37,520 --> 00:52:39,640 QUESTIONS FROM THIS FRAMEWORK OF 1643 00:52:39,640 --> 00:52:40,920 WHAT IS YOUR ADDED VALUE? 1644 00:52:40,920 --> 00:52:42,920 WHEN I WAS LEARNING TO BECOME AN 1645 00:52:42,920 --> 00:52:44,400 OPHTHALMOLOGIST, I SPENT SEVERAL 1646 00:52:44,400 --> 00:52:47,640 YEARS USING THIS DEVICE, A 1647 00:52:47,640 --> 00:52:48,600 CONTACT LENS YOU PUT UP TO 1648 00:52:48,600 --> 00:52:50,560 SOMEBODY'S EYE, AND I LEARNED TO 1649 00:52:50,560 --> 00:52:52,200 EXAMINE THE NUANCE OF THE 1650 00:52:52,200 --> 00:52:53,640 PATIENT'S OPTIC NERVE AND THE 1651 00:52:53,640 --> 00:52:54,720 RETINA BY LOOKING AT THOUSANDS 1652 00:52:54,720 --> 00:52:57,480 OF PEOPLE. 1653 00:52:57,480 --> 00:52:58,360 AND THAT WAS MY ADDED VALUE, 1654 00:52:58,360 --> 00:52:59,800 THAT I COULD DETECT NUANCE IN 1655 00:52:59,800 --> 00:53:02,000 THE STRUCTURE OF THE RETINA. 1656 00:53:02,000 --> 00:53:04,200 NOBODY DOES THAT ANYMORE. 1657 00:53:04,200 --> 00:53:06,920 WE GET TESTS LIKE OCT IMAGES 1658 00:53:06,920 --> 00:53:09,040 FROM THE RETINA, AND THOSE TESTS 1659 00:53:09,040 --> 00:53:10,280 ARE QUANTITATIVE AND THEY SPIT 1660 00:53:10,280 --> 00:53:13,040 OUT NUMBERS. 1661 00:53:13,040 --> 00:53:14,520 SO THE QUESTION I WONDER 1662 00:53:14,520 --> 00:53:16,480 SOMETIMES IS THAT WHAT'S YOUR 1663 00:53:16,480 --> 00:53:17,400 ADDED VALUE? 1664 00:53:17,400 --> 00:53:19,160 IF YOU'RE A RETINAL 1665 00:53:19,160 --> 00:53:21,320 SUBSPECIALIST VERSUS A GENERAL 1666 00:53:21,320 --> 00:53:24,120 OPHTHALMOLOGIST VERSUS AN 1667 00:53:24,120 --> 00:53:25,680 OPTOMETRIST VERSUS A PATIENT 1668 00:53:25,680 --> 00:53:26,680 USING GOOGLE SEARCH AND YOU'RE 1669 00:53:26,680 --> 00:53:29,120 ALL SEEING THE EXACT SAME 1670 00:53:29,120 --> 00:53:30,800 NUMBERS, HOW DO WE GET ADDED 1671 00:53:30,800 --> 00:53:32,360 VALUE IN CLINICAL CARE? 1672 00:53:32,360 --> 00:53:35,080 SO DOCTORS DIAGNOSE PATIENTS. 1673 00:53:35,080 --> 00:53:37,000 WHERE A.I. SYSTEMS ARE REALLY 1674 00:53:37,000 --> 00:53:38,480 PROMISING, BUT YOU NEED TO HAVE 1675 00:53:38,480 --> 00:53:39,680 CLINICAL JUDGMENT TO COLLECT 1676 00:53:39,680 --> 00:53:40,080 THAT DATA. 1677 00:53:40,080 --> 00:53:42,440 THEY ALSO MANAGE DISEASE. 1678 00:53:42,440 --> 00:53:45,040 THEY MANAGE PEOPLE. 1679 00:53:45,040 --> 00:53:48,080 IF SOMEBODY HAS CANCER, ARE WE 1680 00:53:48,080 --> 00:53:49,440 GOING TO START CHEMOTHERAPY, 1681 00:53:49,440 --> 00:53:51,320 WHICH MAY GIVE YOU A LONGER 1682 00:53:51,320 --> 00:53:52,840 LIFESPAN ON AVERAGE, SLIGHTLY, 1683 00:53:52,840 --> 00:53:55,920 BUT MAY MAKE YOU MISERABLE FROM 1684 00:53:55,920 --> 00:53:56,640 SIDE EFFECTS, VERSUS ARE WE 1685 00:53:56,640 --> 00:53:58,880 GOING TO PUT YOU IN HOSPICE CARE 1686 00:53:58,880 --> 00:54:00,000 WHERE YOU CAN SPEND TIME WITH 1687 00:54:00,000 --> 00:54:01,360 YOUR FRIENDS AND FAMILY. 1688 00:54:01,360 --> 00:54:07,120 THOSE ARE THANK YOU MAN HUMAN DE 1689 00:54:07,120 --> 00:54:08,000 THERE'S NOT HE NECESSARILY A 1690 00:54:08,000 --> 00:54:10,000 RIGHT OR WRONG. 1691 00:54:10,000 --> 00:54:11,200 ADDING DOCTORS IN AN 1692 00:54:11,200 --> 00:54:12,160 INCREASINGLY QUANTITATIVE WORLD, 1693 00:54:12,160 --> 00:54:14,040 I THINK THERE MAY BE CONCERNS IF 1694 00:54:14,040 --> 00:54:14,680 THINGS LIKE ELECTRONIC HEALTH 1695 00:54:14,680 --> 00:54:15,640 RECORDS ARE SEPARATING PATIENTS 1696 00:54:15,640 --> 00:54:17,440 FROM DOCTORS, AND YOU KNOW, 1697 00:54:17,440 --> 00:54:18,480 DOCTORS ARE ONLY LEARNING TO 1698 00:54:18,480 --> 00:54:20,080 TALK TO SCREENS INSTEAD OF 1699 00:54:20,080 --> 00:54:22,400 TALKING TO PEOPLE. 1700 00:54:22,400 --> 00:54:23,760 SO I THINK THERE'S A LOT OF 1701 00:54:23,760 --> 00:54:25,000 IMPLICATIONS HERE OF WHAT THE 1702 00:54:25,000 --> 00:54:27,800 FUTURE IS OF MEDICINE. 1703 00:54:27,800 --> 00:54:29,000 SO I'M JUST GOING TO CLOSE UP 1704 00:54:29,000 --> 00:54:31,320 WITH A FEW ANECDOTES ABOUT GOING 1705 00:54:31,320 --> 00:54:32,640 BACK TO THE CONCEPT OF WHAT IT 1706 00:54:32,640 --> 00:54:34,760 HAS MEANT TO BE TO BE AN ASIAN 1707 00:54:34,760 --> 00:54:38,440 AMERICAN SCIENTIST. 1708 00:54:38,440 --> 00:54:38,840 SO ANECDOTES NOW. 1709 00:54:38,840 --> 00:54:39,520 IT WAS 2012. 1710 00:54:39,520 --> 00:54:42,880 AT THE TIME I WAS UK TA TALKY 1711 00:54:42,880 --> 00:54:43,800 FATHER. 1712 00:54:43,800 --> 00:54:45,240 BECAUSE MY FATHER IS AN ASIAN 1713 00:54:45,240 --> 00:54:47,640 PARENT, HE IN 2012 WAS STILL 1714 00:54:47,640 --> 00:54:49,640 TRYING TO TELL ME HOW TO DO MY 1715 00:54:49,640 --> 00:54:50,880 JOB, EVEN THOUGH HE'S NOT A 1716 00:54:50,880 --> 00:54:51,840 PHYSICIAN. 1717 00:54:51,840 --> 00:54:54,080 BUT SO HIS QUESTION TO ME WAS, 1718 00:54:54,080 --> 00:54:56,160 WHAT DO YOU DO DURING AN 1719 00:54:56,160 --> 00:54:57,000 ACADEMIC TALK? 1720 00:54:57,000 --> 00:54:58,920 YOU GO VISIT A PLACE FOR AN 1721 00:54:58,920 --> 00:55:00,320 HOUR, AND SO I TOLD HIM THAT 1722 00:55:00,320 --> 00:55:03,120 WHAT I DO IS I TALK FOR 45 1723 00:55:03,120 --> 00:55:04,440 MINUTES, AND THERE'S 15 MINUTES 1724 00:55:04,440 --> 00:55:07,760 OF QUESTION AND ANSWER AT THE 1725 00:55:07,760 --> 00:55:07,920 END. 1726 00:55:07,920 --> 00:55:09,720 SO HE PAUSES AND HE SAYS, WELL, 1727 00:55:09,720 --> 00:55:10,960 WHY DO YOU DO THAT? 1728 00:55:10,960 --> 00:55:12,320 THAT MAKES NO SENSE. 1729 00:55:12,320 --> 00:55:13,640 SHOULDN'T IT BE BACKWARDS? 1730 00:55:13,640 --> 00:55:15,400 LIKE SHOULDN'T YOU TALK FOR 15 1731 00:55:15,400 --> 00:55:17,880 MINUTES AND THEN DISCUSS FOR 45 1732 00:55:17,880 --> 00:55:19,560 MINUTES, BECAUSE THE DISCUSSION 1733 00:55:19,560 --> 00:55:21,080 AND THE Q & A IS WHERE YOU 1734 00:55:21,080 --> 00:55:23,640 REALLY LEARN AND EXCHANGE IDEAS, 1735 00:55:23,640 --> 00:55:26,240 AND WHERE YOU CAN REALLY TELL, 1736 00:55:26,240 --> 00:55:27,720 DOES A SPEAKER REALLY KNOW WHAT 1737 00:55:27,720 --> 00:55:28,280 THEY'RE TALKING ABOUT. 1738 00:55:28,280 --> 00:55:29,520 I THOUGHT ABOUT THAT AND REALLY 1739 00:55:29,520 --> 00:55:30,680 THOUGHT THAT, YOU KNOW, IT MADE 1740 00:55:30,680 --> 00:55:32,280 SOME SENSE, BUT WE JUST DON'T DO 1741 00:55:32,280 --> 00:55:34,080 THINGS THAT WAY. 1742 00:55:34,080 --> 00:55:35,160 IT TURNS OUT THAT ONE OF THE 1743 00:55:35,160 --> 00:55:36,640 THINGS I THOUGHT BACK TO WAS 1744 00:55:36,640 --> 00:55:39,320 THAT THE PREVIOUS YEAR, I HAD 1745 00:55:39,320 --> 00:55:41,840 GONE TO TWO VERY SIMILAR TRIPS 1746 00:55:41,840 --> 00:55:42,560 INTERNATIONALLY. 1747 00:55:42,560 --> 00:55:44,960 IN AUGUST 2011, I WAS IN BRAZIL, 1748 00:55:44,960 --> 00:55:47,440 AND IN NOVEMBER 2011, I WAS IN 1749 00:55:47,440 --> 00:55:47,680 TAIWAN. 1750 00:55:47,680 --> 00:55:50,800 AND IN BOTH OF THOSE PLACES, I 1751 00:55:50,800 --> 00:55:55,000 WAS INVITED BY PEOPLE WHO I 1752 00:55:55,000 --> 00:55:57,920 KNEW, THEY'RE BOTH RETINA 1753 00:55:57,920 --> 00:55:59,680 SPECIALISTS, MET PEOPLE, MADE 1754 00:55:59,680 --> 00:56:01,560 SOME FRIENDS AND COLLEAGUES, AND 1755 00:56:01,560 --> 00:56:03,680 I GAVE A LOT OF TALKS. 1756 00:56:03,680 --> 00:56:06,280 AND AFTER THIS TALK IN BRAZIL, 1757 00:56:06,280 --> 00:56:08,320 AFTER 45 MINUTES, ASKED THE 1758 00:56:08,320 --> 00:56:09,640 QUESTION, OKAY, WELL, I'M DONE, 1759 00:56:09,640 --> 00:56:11,640 WHAT QUESTIONS DO PEOPLE HAVE? 1760 00:56:11,640 --> 00:56:13,120 AND SO MANY PEOPLE RAISED THEIR 1761 00:56:13,120 --> 00:56:14,200 HANDS AND WE HAD A DISCUSSION 1762 00:56:14,200 --> 00:56:16,280 THAT ACTUALLY WENT MUCH LONGER 1763 00:56:16,280 --> 00:56:18,520 THAN 15 MINUTES. 1764 00:56:18,520 --> 00:56:19,920 THREE MONTHS LATER, I GAVE THE 1765 00:56:19,920 --> 00:56:21,400 EXACT SAME TALK IN TAIWAN, AND 1766 00:56:21,400 --> 00:56:23,400 AT THE END, WHAT QUESTIONS DO 1767 00:56:23,400 --> 00:56:25,600 PEOPLE HAVE, NOT A SINGLE HAND 1768 00:56:25,600 --> 00:56:27,200 WENT UP. 1769 00:56:27,200 --> 00:56:28,600 AND I HAD NEVER HAD THAT HAPPEN 1770 00:56:28,600 --> 00:56:30,200 IN MY CAREER WHERE NOBODY ASKED 1771 00:56:30,200 --> 00:56:31,720 A QUESTION. 1772 00:56:31,720 --> 00:56:34,040 AND THOSE OF YOU IN THE AUDIENCE 1773 00:56:34,040 --> 00:56:37,720 WHO COME FROM BACKGROUNDS IN 1774 00:56:37,720 --> 00:56:41,520 ASIA WILL RECOGNIZE THAT THAT'S 1775 00:56:41,520 --> 00:56:43,000 NOT ALWAYS ENCOURAGED, THAT 1776 00:56:43,000 --> 00:56:43,920 SOMETIMES PEOPLE DON'T WANT TO 1777 00:56:43,920 --> 00:56:45,320 LOOK BAD IF THEY ASK A QUESTION. 1778 00:56:45,320 --> 00:56:46,720 THE PARADIGM MAY BE THAT THE 1779 00:56:46,720 --> 00:56:48,480 PERSON WITH MORE GREY HAIR IS 1780 00:56:48,480 --> 00:56:49,120 CONSIDERED AN EXPERT AND YOU 1781 00:56:49,120 --> 00:56:50,600 DON'T WANT TO QUESTION THEIR 1782 00:56:50,600 --> 00:56:51,080 EXPERTISE. 1783 00:56:51,080 --> 00:56:52,560 EVEN WHEN THE EXPERT WAS 41 1784 00:56:52,560 --> 00:56:54,600 YEARS OLD IN 2011, YOU KNOW, 1785 00:56:54,600 --> 00:56:56,960 THEY MAY NOT HAVE WANTED TO ASK 1786 00:56:56,960 --> 00:56:57,360 ANYTHING. 1787 00:56:57,360 --> 00:56:59,440 SO I THOUGHT, WELL, ISN'T THAT 1788 00:56:59,440 --> 00:57:00,800 INTERESTING, THAT MY PARENTS 1789 00:57:00,800 --> 00:57:03,280 AFTER HAVING BEEN IN THIS 1790 00:57:03,280 --> 00:57:04,200 COUNTRY MUCH LONGER THAN THEY 1791 00:57:04,200 --> 00:57:05,960 WERE IN TAIWAN, ASKED THIS 1792 00:57:05,960 --> 00:57:07,640 QUESTION, WELL, MAYBE THE WAY 1793 00:57:07,640 --> 00:57:12,160 THAT WE DO THINGS IN THE U.S. 1794 00:57:12,160 --> 00:57:13,400 SHOULD BE MORE FOCUSED ON 1795 00:57:13,400 --> 00:57:14,720 EXCHANGE OF IDEAS, LET ALONE THE 1796 00:57:14,720 --> 00:57:16,120 WAY THEY WERE DOING THINGS IN 1797 00:57:16,120 --> 00:57:18,320 TAIWAN, IN NOVEMBER 2011. 1798 00:57:18,320 --> 00:57:20,480 SO IT TURNED OUT THAT THREE 1799 00:57:20,480 --> 00:57:24,280 YEARS LATER, BACK IN TAIWAN, 1800 00:57:24,280 --> 00:57:25,760 GIVING A SERIES OF TALKS JUST 1801 00:57:25,760 --> 00:57:27,240 LIKE BEFORE, SAW A LITTLE BIT OF 1802 00:57:27,240 --> 00:57:28,800 CULTURAL SITES, BUT IT TURNED 1803 00:57:28,800 --> 00:57:32,840 OUT THAT WE DID ONE THING THAT 1804 00:57:32,840 --> 00:57:35,600 WAS DIFFERENT. 1805 00:57:35,600 --> 00:57:37,520 WE BASICALLY SPENT MONTHS 1806 00:57:37,520 --> 00:57:38,320 SCRIPTING AN EVENT WHERE WE HAD 1807 00:57:38,320 --> 00:57:39,520 PEOPLE THERE COME UP WITH 1808 00:57:39,520 --> 00:57:41,680 RESEARCH PROPOSALS. 1809 00:57:41,680 --> 00:57:43,080 SO IT DOESN'T MATTER IF YOU'RE 1810 00:57:43,080 --> 00:57:44,080 THE DEPARTMENT CHAIR OR IF 1811 00:57:44,080 --> 00:57:45,640 YOU'RE A RESEARCH FELLOW, 1812 00:57:45,640 --> 00:57:48,520 EVERYBODY WOULD SIT AROUND THE 1813 00:57:48,520 --> 00:57:49,840 CIRCLE HERE AND WE PUT THESE UP 1814 00:57:49,840 --> 00:57:51,360 ON THE SCREEN AND DISCUSS THEM, 1815 00:57:51,360 --> 00:57:52,600 AND WE'D CRITICIZE THEM. 1816 00:57:52,600 --> 00:57:53,880 WHY ARE YOU DOING THINGS THIS 1817 00:57:53,880 --> 00:57:55,240 WAY, WHY ARE YOU -- AND MY JOB 1818 00:57:55,240 --> 00:57:56,640 WAS TO MODERATE THAT DISCUSSION. 1819 00:57:56,640 --> 00:58:00,160 AND IT WAS FASCINATING BECAUSE I 1820 00:58:00,160 --> 00:58:01,200 WOULD BASICALLY TAKE WHAT I 1821 00:58:01,200 --> 00:58:03,000 THOUGHT THAT I'VE LEARNED, AND 1822 00:58:03,000 --> 00:58:04,360 MAYBE WE NEED 45 MINUTES OF 1823 00:58:04,360 --> 00:58:07,240 DISCUSSION AND 15 MINUTES OF 1824 00:58:07,240 --> 00:58:08,120 PRESENTATION, AND TO TRY TO COME 1825 00:58:08,120 --> 00:58:10,880 UP WITH A BETTER SYSTEM. 1826 00:58:10,880 --> 00:58:12,080 I HOPE THAT'S SOMETHING WE CAN 1827 00:58:12,080 --> 00:58:15,080 REALLY DO AT NIH IN TERMS OF WE 1828 00:58:15,080 --> 00:58:16,360 INTEGRATE THE GOOD THINGS FROM 1829 00:58:16,360 --> 00:58:17,760 EVERY CULTURE AND COME UP WITH 1830 00:58:17,760 --> 00:58:19,120 WHAT'S THE BEST PROCESS FOR 1831 00:58:19,120 --> 00:58:26,320 DOING SCIENCE. 1832 00:58:26,320 --> 00:58:29,160 SO WITH THAT, I'M GOING TO GO UP 1833 00:58:29,160 --> 00:58:29,480 TO 2021. 1834 00:58:29,480 --> 00:58:30,920 IN SOME WAYS, WHAT WE SEE IN THE 1835 00:58:30,920 --> 00:58:34,680 NEWS NOW IS VIOLENCE DIRECTED 1836 00:58:34,680 --> 00:58:35,480 TOWARD ASIAN AMERICAN, STOP 1837 00:58:35,480 --> 00:58:36,400 ASIAN HATE MOVEMENT. 1838 00:58:36,400 --> 00:58:38,600 WE'VE GOT THINGS LIKE SHOOTINGS 1839 00:58:38,600 --> 00:58:39,440 IN ATLANTA. 1840 00:58:39,440 --> 00:58:40,640 AND SOMETIMES I FEEL LIKE WE'RE 1841 00:58:40,640 --> 00:58:42,400 GOING BACK TO THE 1980s, BACK 1842 00:58:42,400 --> 00:58:45,840 IN THE ERA OF VINCENT CHIN AND 1843 00:58:45,840 --> 00:58:46,880 MARTIAL ARTS THEATER. 1844 00:58:46,880 --> 00:58:48,880 AND ONE OF THE THINGS, AFTER THE 1845 00:58:48,880 --> 00:58:53,200 ATLANTA SPA SHOOTINGS, I SAW 1846 00:58:53,200 --> 00:58:54,440 MANY UNIVERSITIES HAD THEIR 1847 00:58:54,440 --> 00:58:55,440 STATEMENTS ABOUT WHAT DOES THIS 1848 00:58:55,440 --> 00:58:56,560 MEAN, WE'RE GOING TO SPEAK OUT 1849 00:58:56,560 --> 00:58:58,160 AGAINST THIS, AND THE ONE THAT 1850 00:58:58,160 --> 00:58:59,960 REALLY STUCK OUT AT ME THE MOST 1851 00:58:59,960 --> 00:59:02,080 ACTUALLY CAME FROM HOWARD 1852 00:59:02,080 --> 00:59:05,000 UNIVERSITY. 1853 00:59:05,000 --> 00:59:06,680 THIS IS A COPY OF THAT LETTER 1854 00:59:06,680 --> 00:59:08,200 FROM THE PRESIDENT OF HOWARD. 1855 00:59:08,200 --> 00:59:09,760 IT SAID, HERE, IT IS HOWARD'S 1856 00:59:09,760 --> 00:59:11,400 MOST SACRED DUTY TO STAND UP TO 1857 00:59:11,400 --> 00:59:14,400 HATE IN ALL FORMS AGAINST ALL 1858 00:59:14,400 --> 00:59:14,760 PEOPLE. 1859 00:59:14,760 --> 00:59:17,000 IT'S NOT ENOUGH TO REPUDIATE THE 1860 00:59:17,000 --> 00:59:19,080 SHOOTINGS, WE MUST ALSO CONDEMN 1861 00:59:19,080 --> 00:59:21,320 THE PHYSICAL ATTACKS, VERBAL 1862 00:59:21,320 --> 00:59:22,880 ASSAULTS, DEMONIZING BEHAVIOR 1863 00:59:22,880 --> 00:59:24,760 AND THE OTHERRING OF THE ASIAN 1864 00:59:24,760 --> 00:59:25,560 AMERICAN COMMUNITY. 1865 00:59:25,560 --> 00:59:26,720 OF COURSE AS A HISTORICALLY 1866 00:59:26,720 --> 00:59:27,480 BLACK COLLEGE AND UNIVERSITY, 1867 00:59:27,480 --> 00:59:32,360 THIS SORT OF TRAGEDY RESONATES 1868 00:59:32,360 --> 00:59:34,320 STRONGLY AMONG OUR CAMPUS. 1869 00:59:34,320 --> 00:59:38,960 A FEAR HAS BEEN EMBEDDED INTO 1870 00:59:38,960 --> 00:59:42,000 THEIR DNA. 1871 00:59:42,000 --> 00:59:48,200 I READ THIS AND REALLY THOUGHT, 1872 00:59:48,200 --> 00:59:51,080 WOW, WHAT MY HOPE IS FOR NIH, 1873 00:59:51,080 --> 00:59:58,520 THE DIFFERENT COMMUNITIES, THE 1874 00:59:58,520 --> 01:00:01,920 AA AND AANHPI COMMUNITY, ALSO TO 1875 01:00:01,920 --> 01:00:03,320 UNDERSTAND AND EMPATHIZE WITH 1876 01:00:03,320 --> 01:00:04,320 OTHER COMMUNITIES JUST AS I SAW 1877 01:00:04,320 --> 01:00:05,240 FROM THIS LETTER. 1878 01:00:05,240 --> 01:00:06,560 MY HOPE IS THAT IF WE CAN DO 1879 01:00:06,560 --> 01:00:09,360 THAT AT NIH, I DON'T KNOW IF I 1880 01:00:09,360 --> 01:00:10,760 COULD SAY IT BETTER THAN WAYNE 1881 01:00:10,760 --> 01:00:11,760 FREDERICK, WHO'S THE PRESIDENT, 1882 01:00:11,760 --> 01:00:13,400 A BETTER WORLD IS WITHIN OUR 1883 01:00:13,400 --> 01:00:15,360 GRASP IF WE UNITED, CAN 1884 01:00:15,360 --> 01:00:17,000 ERADICATE THAT WHICH DIVIDES US, 1885 01:00:17,000 --> 01:00:18,400 AND WEAVE MORE TIGHTLY THE 1886 01:00:18,400 --> 01:00:23,120 FABRICS THAT BIND US. 1887 01:00:23,120 --> 01:00:24,360 SO WITH THAT, ALMOST DONE, AND 1888 01:00:24,360 --> 01:00:27,640 WHAT HAVE I LEARNED AS AN ASIAN 1889 01:00:27,640 --> 01:00:29,000 AMERICAN CLINICIAN-SCIENTIST? 1890 01:00:29,000 --> 01:00:30,160 WELL, NUMBER ONE, IT'S GOOD TO 1891 01:00:30,160 --> 01:00:33,680 DRAW FROM DIFFERENT CULTURES. 1892 01:00:33,680 --> 01:00:35,200 IN ASIAN CULTURE, MY PARENTS 1893 01:00:35,200 --> 01:00:37,320 GREW UP IN, THERE'S THINGS LIKE 1894 01:00:37,320 --> 01:00:37,600 DISCIPLINE. 1895 01:00:37,600 --> 01:00:38,720 HERE WHAT I LEARNED IS THE 1896 01:00:38,720 --> 01:00:40,960 EXCHANGE OF IDEAS. 1897 01:00:40,960 --> 01:00:42,640 THE 15 MINUTES VERSUS 45 1898 01:00:42,640 --> 01:00:42,920 MINUTES. 1899 01:00:42,920 --> 01:00:44,480 AND WHAT I LEARNED AS A KID IS 1900 01:00:44,480 --> 01:00:45,640 MAYBE IT'S NOT THE WORST THING 1901 01:00:45,640 --> 01:00:47,480 TO BE DIFFERENT. 1902 01:00:47,480 --> 01:00:48,640 I FELT LIKE, WELL, I'M NOT GOING 1903 01:00:48,640 --> 01:00:52,640 TO BE CLASS PRESIDENT, BUT MAYBE 1904 01:00:52,640 --> 01:00:53,800 I NEVER FELT THAT NEED TO 1905 01:00:53,800 --> 01:00:54,640 CONFORM WITH OTHER PEOPLE AND 1906 01:00:54,640 --> 01:00:55,840 MAYBE THAT HELPED ME THINK 1907 01:00:55,840 --> 01:00:56,800 OUTSIDE THE BOX. 1908 01:00:56,800 --> 01:00:58,600 I WAS AN ENGINEER WHO BECAME A 1909 01:00:58,600 --> 01:01:01,520 DOCTOR, AND A DOCTOR WHO BECAME 1910 01:01:01,520 --> 01:01:04,160 A BIOMEDICAL INFORMATICS PERSON. 1911 01:01:04,160 --> 01:01:07,640 SO MAYBE THAT SORT OF HELPED ME. 1912 01:01:07,640 --> 01:01:10,120 I LEARNED THINGS LIKE THE 1913 01:01:10,120 --> 01:01:11,800 IMPORTANCE OF EFFECTIVE 1914 01:01:11,800 --> 01:01:13,600 COMMUNICATION, AND OF SPEAKING 1915 01:01:13,600 --> 01:01:14,520 UP, AND THAT WAS THE REASON I 1916 01:01:14,520 --> 01:01:16,000 TALKED ABOUT ANECDOTES LIKE 1917 01:01:16,000 --> 01:01:19,480 WALTER CON CITE CRONKITE AT 1918 01:01:19,480 --> 01:01:19,680 CHIN. 1919 01:01:19,680 --> 01:01:20,880 I WOULD SAY I'VE LEARNED THE 1920 01:01:20,880 --> 01:01:23,000 NEED FOR MENTORSHIP AND ROLE 1921 01:01:23,000 --> 01:01:26,120 MODELS AND PEOPLE WHO LOOK LIKE 1922 01:01:26,120 --> 01:01:26,600 YOU. 1923 01:01:26,600 --> 01:01:33,760 THIS IS REASONS I SHOWED SLIDES 1924 01:01:33,760 --> 01:01:35,720 OF CAREER DEVELOPMENT. 1925 01:01:35,720 --> 01:01:39,000 THESE ARE PICTURES FROM THE 1926 01:01:39,000 --> 01:01:39,480 CHINESE AMERICAN 1927 01:01:39,480 --> 01:01:40,000 OPHTHALMOLOGICAL SOCIETY. 1928 01:01:40,000 --> 01:01:41,920 HERE'S A PICTURE OF ME, HERE'S 1929 01:01:41,920 --> 01:01:45,600 EMILY CHU AT ONE OF THESE 1930 01:01:45,600 --> 01:01:46,920 DINNERS. 1931 01:01:46,920 --> 01:01:48,120 SIDE SAY THAT, YES, DEFINITELY 1932 01:01:48,120 --> 01:01:49,240 THE ROLE OF COLLABORATION, TEAM 1933 01:01:49,240 --> 01:01:51,040 WORK AND SCIENCE, AND THAT'S WHY 1934 01:01:51,040 --> 01:01:58,840 I SHOWED PICTURES OF PEOPLE FROM 1935 01:01:58,840 --> 01:02:00,400 OTHER BACKGROUNDS, REGARDLESS OF 1936 01:02:00,400 --> 01:02:01,760 WHETHER IT'S PEOPLE FROM OTHER 1937 01:02:01,760 --> 01:02:02,840 SCIENTIFIC BACKGROUNDS OR PEOPLE 1938 01:02:02,840 --> 01:02:04,040 LIKE THE PRESIDENT OF HOWARD 1939 01:02:04,040 --> 01:02:04,320 UNIVERSITY. 1940 01:02:04,320 --> 01:02:05,400 THOSE ARE THE REASONS THAT I'VE 1941 01:02:05,400 --> 01:02:07,080 PUT THOSE SLIDES UP. 1942 01:02:07,080 --> 01:02:09,920 AND SO MY VERY LAST SLIDE IS 1943 01:02:09,920 --> 01:02:11,680 WHAT IS OUR PAST, PRESENT AND 1944 01:02:11,680 --> 01:02:11,920 FUTURE? 1945 01:02:11,920 --> 01:02:14,240 SO WE STARTED THIS TALK WITH 1946 01:02:14,240 --> 01:02:17,200 K.T. JEANG, I WANT TO END IT 1947 01:02:17,200 --> 01:02:18,840 WITH K.T. JEANG. 1948 01:02:18,840 --> 01:02:23,960 ONE OF HIS BIG POINTS IN 2008, 1949 01:02:23,960 --> 01:02:27,440 2010 AND 2013 WAS ADVOCACY FOR 1950 01:02:27,440 --> 01:02:28,160 LEADERSHIP AMONG ASIAN AMERICAN, 1951 01:02:28,160 --> 01:02:32,080 AND THAT'S WHERE WE WERE IN THE 1952 01:02:32,080 --> 01:02:32,640 PAST. 1953 01:02:32,640 --> 01:02:38,280 NOW IN 2022, WE HAVE THREE I.C. 1954 01:02:38,280 --> 01:02:40,160 DIRECTORS OF ASIAN AMERICAN 1955 01:02:40,160 --> 01:02:43,640 DESCENT. 1956 01:02:43,640 --> 01:02:45,280 I THINK K.T. JEANG WOULD HAVE 1957 01:02:45,280 --> 01:02:46,920 BEEN, I HOPE, IMPRESSED BY THAT 1958 01:02:46,920 --> 01:02:49,320 PROGRESS. 1959 01:02:49,320 --> 01:02:50,520 AND THE LAST THING I WANT TO SAY 1960 01:02:50,520 --> 01:02:53,960 IN TERMS OF THE FUTURE IS THAT I 1961 01:02:53,960 --> 01:02:56,640 GAVE THAT ANECDOTE OF NEVER 1962 01:02:56,640 --> 01:02:58,480 WANTING TO BE CLASS PRESIDENT. 1963 01:02:58,480 --> 01:03:01,360 AND FOUR YEARS AGO, OUR DAUGHTER 1964 01:03:01,360 --> 01:03:04,600 CAME HOME FROM SCHOOL, SHE WAS A 1965 01:03:04,600 --> 01:03:06,720 JUNIOR IN HIGH SCHOOL BACK THEN, 1966 01:03:06,720 --> 01:03:09,240 SHE CAME BACK HOME, IT WAS 1967 01:03:09,240 --> 01:03:11,240 8:00 P.M., SHE RAN UPSTAIRS, WAS 1968 01:03:11,240 --> 01:03:14,720 DOING SOME WORK, DENT DIDN'TR 1969 01:03:14,720 --> 01:03:16,360 EATER DINNER. 1970 01:03:16,360 --> 01:03:17,480 WE SAID WHAT ARE YOU DOING? 1971 01:03:17,480 --> 01:03:18,880 SHE SAID, I WANT TO RUN FOR 1972 01:03:18,880 --> 01:03:19,560 CLASS PRESIDENT. 1973 01:03:19,560 --> 01:03:21,000 IT TURNED OUT SHE DECIDED TO RUN 1974 01:03:21,000 --> 01:03:22,720 FOR STUDENT BODY, VICE PRESIDENT 1975 01:03:22,720 --> 01:03:24,120 INSTEAD, AND SHE WON. 1976 01:03:24,120 --> 01:03:25,520 AND SO THIS WAS ERICA WHEN SHE 1977 01:03:25,520 --> 01:03:26,880 WAS IN HIGH SCHOOL WITH HER 1978 01:03:26,880 --> 01:03:28,120 FRIENDS. 1979 01:03:28,120 --> 01:03:30,800 AND TO ME, I THINK THAT WHEN I 1980 01:03:30,800 --> 01:03:32,000 LOOK AT THE CHANGE THAT OCCURRED 1981 01:03:32,000 --> 01:03:34,400 BETWEEN MY GENERATION AND MY 1982 01:03:34,400 --> 01:03:35,600 DAUGHTER'S GENERATION, THAT 1983 01:03:35,600 --> 01:03:36,920 REALLY GIVES ME A LOT OF 1984 01:03:36,920 --> 01:03:38,320 OPTIMISM FOR THE FUTURE. 1985 01:03:38,320 --> 01:03:40,680 SO I HOPE THAT THIS HAS REALLY 1986 01:03:40,680 --> 01:03:42,600 BEEN ABLE TO DO JUSTICE TO 1987 01:03:42,600 --> 01:03:44,160 DR. JEANG'S LEGACY, AND THANK 1988 01:03:44,160 --> 01:03:45,680 YOU VERY MUCH FOR THE PRIVILEGE 1989 01:03:45,680 --> 01:03:53,000 OF ALLOWING ME TO SPEAK HERE. 1990 01:03:53,000 --> 01:03:55,160 >> THANK YOU VERY MUCH FOR A 1991 01:03:55,160 --> 01:03:56,640 VERY INFORMATIVE TALK. 1992 01:03:56,640 --> 01:03:59,000 WE HAVE TIME FOR A LITTLE BIT OF 1993 01:03:59,000 --> 01:04:04,760 QUESTION AND ANSWER. 1994 01:04:04,760 --> 01:04:06,000 PEOPLE IN THE AUDIENCE CAN USE 1995 01:04:06,000 --> 01:04:07,640 THE SEND LIVE FEEDBACK BUTTON TO 1996 01:04:07,640 --> 01:04:08,000 ASK QUESTIONS. 1997 01:04:08,000 --> 01:04:11,360 WE HAVE ONE QUESTION ALREADY 1998 01:04:11,360 --> 01:04:12,160 FROM TOM SNYDER. 1999 01:04:12,160 --> 01:04:16,280 HAVE YOU CONNECTED WITH NIST AND 2000 01:04:16,280 --> 01:04:18,040 IEEE ABOUT THE DATA 2001 01:04:18,040 --> 01:04:22,200 STANDARDIZATION? 2002 01:04:22,200 --> 01:04:23,080 >> TOM, THANKS FOR ASKING ABOUT 2003 01:04:23,080 --> 01:04:23,400 THAT. 2004 01:04:23,400 --> 01:04:26,400 WE ACTUALLY HAVEN'T CONNECTED TO 2005 01:04:26,400 --> 01:04:28,160 NIST YET, AND HAVE WONDERED 2006 01:04:28,160 --> 01:04:30,080 ABOUT DOING THAT. 2007 01:04:30,080 --> 01:04:31,840 OF AND THANKS FOR THAT. 2008 01:04:31,840 --> 01:04:35,760 SO FAR WE'VE JUST WORKED WITH 2009 01:04:35,760 --> 01:04:37,960 ONCFDA ABOUT IT. 2010 01:04:37,960 --> 01:04:40,160 THERE ARE STANDARDS FOR 2011 01:04:40,160 --> 01:04:41,880 RADIOLOGY THAT HAVE BEEN ADOPTED 2012 01:04:41,880 --> 01:04:46,800 FOR DECADES, DICOM OBVIOUSLY, 2013 01:04:46,800 --> 01:04:48,120 AND SEVERAL OF THE SPEAKERS FROM 2014 01:04:48,120 --> 01:04:52,240 OUR WORKSHOP WERE FOLKS FROM THE 2015 01:04:52,240 --> 01:04:57,520 RADIOLOGY COMMUNITY, AND WE'LL 2016 01:04:57,520 --> 01:05:00,440 DEFINITELY THINK MORE ABOUT 2017 01:05:00,440 --> 01:05:00,640 THAT. 2018 01:05:00,640 --> 01:05:02,920 >> ONE QUESTION I'VE ALWAYS HAD, 2019 01:05:02,920 --> 01:05:05,520 YOU MENTION THE ABILITY OF THE 2020 01:05:05,520 --> 01:05:08,080 A.I. TO IDENTIFY SEX FROM THE 2021 01:05:08,080 --> 01:05:09,280 RETINAL PATTERNS. 2022 01:05:09,280 --> 01:05:12,320 WHY IS IT SO HARD TO FIGURE OUT 2023 01:05:12,320 --> 01:05:18,040 WHAT THE A.I. IS SEEING? 2024 01:05:18,040 --> 01:05:20,640 >> ROLAND, THERE'S NEVER BEEN A 2025 01:05:20,640 --> 01:05:24,440 GOOD WAY TO LOOK UNDERNEATH WHAT 2026 01:05:24,440 --> 01:05:26,440 THE EXPLANATIONS OF THESE DEEP 2027 01:05:26,440 --> 01:05:27,920 NEURAL NETWORKS ARE DOING, IN 2028 01:05:27,920 --> 01:05:30,440 OTHER WORDS, WHAT LAYER OF THESE 2029 01:05:30,440 --> 01:05:31,840 CONVOLUTIONAL NETWORKS IS REALLY 2030 01:05:31,840 --> 01:05:34,520 CORRESPONDING TO WHAT. 2031 01:05:34,520 --> 01:05:37,480 FOR A WHILE, EXPLAINABLE A.I. 2032 01:05:37,480 --> 01:05:41,600 WAS A VERY HOT TOPIC. 2033 01:05:41,600 --> 01:05:42,960 WHY DID THESE SYSTEMS COME UP 2034 01:05:42,960 --> 01:05:44,360 WITH THE CONCLUSION THEY COME UP 2035 01:05:44,360 --> 01:05:44,560 WITH. 2036 01:05:44,560 --> 01:05:50,200 IN FACT, MY VERY LAST NIH GRANT 2037 01:05:50,200 --> 01:05:52,400 HAD A SPECIFIC AIM THAT DEALT 2038 01:05:52,400 --> 01:05:55,520 WITH EXPLAINABLE A.I. IN ROP. 2039 01:05:55,520 --> 01:06:00,360 I'M NOT SO SURE HOW IMPORTANT 2040 01:06:00,360 --> 01:06:03,520 THAT'S GOING TO BE LOOKING 2041 01:06:03,520 --> 01:06:03,920 FORWARD. 2042 01:06:03,920 --> 01:06:06,920 MY GUESS IS I'M NOT SURE IT'S 2043 01:06:06,920 --> 01:06:10,120 BETTER THAN ANYBODY ELSE'S 2044 01:06:10,120 --> 01:06:12,000 GUESS. 2045 01:06:12,000 --> 01:06:13,480 BUT I THINK WE'RE AT A 2046 01:06:13,480 --> 01:06:14,320 TRANSITIONAL POINT IN MEDICINE 2047 01:06:14,320 --> 01:06:15,280 WHERE PEOPLE DON'T QUITE KNOW 2048 01:06:15,280 --> 01:06:16,640 WHAT TO DO WITH THESE 2049 01:06:16,640 --> 01:06:18,560 TECHNOLOGIES OR HOW TO INTERPRET 2050 01:06:18,560 --> 01:06:20,360 THEM. 2051 01:06:20,360 --> 01:06:22,400 AND THEY MAY NOT TRUST THEM AS A 2052 01:06:22,400 --> 01:06:23,400 RESULT. 2053 01:06:23,400 --> 01:06:25,120 ON THE OTHER HAND, HOW MANY 2054 01:06:25,120 --> 01:06:27,800 DOCTORS REALLY UNDERSTAND HOW AN 2055 01:06:27,800 --> 01:06:29,960 MRI MACHINE WORKS, OR HOW THAT 2056 01:06:29,960 --> 01:06:36,320 MACHINE WORKS THAT GIVES YOU 2057 01:06:36,320 --> 01:06:37,840 YOUR BLOOD COUNT, OR FOR THAT 2058 01:06:37,840 --> 01:06:38,960 MATTER, HOW MANY DOCTORS TRULY 2059 01:06:38,960 --> 01:06:40,280 UNDERSTAND THE INTRICACY OF WHAT 2060 01:06:40,280 --> 01:06:41,920 WENT INTO THAT STUDY DESIGN, 2061 01:06:41,920 --> 01:06:43,640 THAT CLINICAL TRIAL THAT PROVED 2062 01:06:43,640 --> 01:06:46,480 WHATEVER IT PROVED, AND IT 2063 01:06:46,480 --> 01:06:47,680 WOULDN'T SURPRISE ME THAT IF 2064 01:06:47,680 --> 01:06:49,200 THESE SYSTEMS REALLY HAVE VALUE, 2065 01:06:49,200 --> 01:06:51,920 THAT PEOPLE WILL JUST TRUST WHAT 2066 01:06:51,920 --> 01:06:53,000 THEY SAY, BUT IT'S GOING TO BE 2067 01:06:53,000 --> 01:06:55,320 UP TO THE RESEARCH COMMUNITY TO 2068 01:06:55,320 --> 01:06:58,080 PROVE THAT THEY REALLY WORK, AND 2069 01:06:58,080 --> 01:06:59,720 THAT THEY WORK ACROSS DIFFERENT 2070 01:06:59,720 --> 01:07:00,000 POPULATIONS. 2071 01:07:00,000 --> 01:07:02,000 SO IT'S A LONG ANSWER, AND I 2072 01:07:02,000 --> 01:07:04,120 KNOW IT'S A LITTLE BIT NUANCED 2073 01:07:04,120 --> 01:07:05,640 TO THAT QUESTION. 2074 01:07:05,640 --> 01:07:07,360 >> THAT ACTUALLY FEEDS INTO MY 2075 01:07:07,360 --> 01:07:08,800 NEXT QUESTION. 2076 01:07:08,800 --> 01:07:10,280 I GREW UP WATCHING A LOT OF 2077 01:07:10,280 --> 01:07:11,200 SCIENCE FICTION. 2078 01:07:11,200 --> 01:07:13,600 I DON'T KNOW IF YOU EVER SAW THE 2079 01:07:13,600 --> 01:07:16,600 BUCK RODGERS WITH DR. THEOPOLIS, 2080 01:07:16,600 --> 01:07:18,960 THE A.I. WHO WAS PART OF THE 2081 01:07:18,960 --> 01:07:21,480 LEADERSHIP TEAM, AND I THINK 2082 01:07:21,480 --> 01:07:23,680 ABOUT THE ANALOGIES BETWEEN 2083 01:07:23,680 --> 01:07:25,760 MANAGING HUMAN DIVERSITY WITHIN 2084 01:07:25,760 --> 01:07:28,160 A TEAM WITH DIVERSE PERSPECTIVES 2085 01:07:28,160 --> 01:07:31,000 AND JUST INCORPORATING THE A.I. 2086 01:07:31,000 --> 01:07:32,760 INTO THAT IS SORT OF JUST 2087 01:07:32,760 --> 01:07:34,680 EXTENDING THAT SPECTRUM OF 2088 01:07:34,680 --> 01:07:38,440 MANAGING DIVERSITY. 2089 01:07:38,440 --> 01:07:39,840 HOW DO YOU FEEL ABOUT THAT? 2090 01:07:39,840 --> 01:07:41,480 DO YOU THINK OF A.I. AS A 2091 01:07:41,480 --> 01:07:46,000 PARTNER? 2092 01:07:46,000 --> 01:07:51,520 >> SORROW LAND, THERE'S E 2093 01:07:51,520 --> 01:07:52,160 PARTS TO THAT QUESTION. 2094 01:07:52,160 --> 01:07:55,520 ONE OF THEM IS, IS THE A.I. A 2095 01:07:55,520 --> 01:07:58,760 PARTNER, AND I THINK -- I HOPE 2096 01:07:58,760 --> 01:08:01,000 THE ANSWER IS DEFINITELY YES. 2097 01:08:01,000 --> 01:08:03,600 THERE ARE A LOT OF PEOPLE IN THE 2098 01:08:03,600 --> 01:08:04,600 DOCTOR COMMUNITY WHO ARE WORRIED 2099 01:08:04,600 --> 01:08:06,520 THAT THE A.I. IS GOING TO TAKE 2100 01:08:06,520 --> 01:08:08,040 THEIR JOB, AND I DON'T SEE IT 2101 01:08:08,040 --> 01:08:08,800 THAT WAY AT ALL. 2102 01:08:08,800 --> 01:08:11,520 I SEE IT AS BEING A PARTNERSHIP. 2103 01:08:11,520 --> 01:08:13,960 I SEE IT AS BASICALLY GIVING 2104 01:08:13,960 --> 01:08:15,680 MORE DATA TO THE DOCTORS TO HELP 2105 01:08:15,680 --> 01:08:17,200 MAKE WHAT THEIR MANAGEMENT PLAN 2106 01:08:17,200 --> 01:08:18,200 IS. 2107 01:08:18,200 --> 01:08:21,280 THE SECOND ISSUE, THOUGH, IS A 2108 01:08:21,280 --> 01:08:25,280 QUESTION ABOUT WHERE DOES THE 2109 01:08:25,280 --> 01:08:29,080 A.I. STOP AND WHERE DOES THE 2110 01:08:29,080 --> 01:08:29,560 DOCTORING BEGIN. 2111 01:08:29,560 --> 01:08:30,920 I THINK THAT'S A LITTLE MORE OF 2112 01:08:30,920 --> 01:08:31,760 A NUANCED QUESTION. 2113 01:08:31,760 --> 01:08:33,280 SOME OF THOSE EXAMPLES THAT I 2114 01:08:33,280 --> 01:08:36,360 THINK YOU'RE GETTING AT, YOU 2115 01:08:36,360 --> 01:08:39,400 KNOW, DIVERSITY AMONG PEOPLE OR 2116 01:08:39,400 --> 01:08:40,360 GETTING PEOPLE TO WORK TOGETHER, 2117 01:08:40,360 --> 01:08:41,640 I'M NOT SURE I'M IMMEDIATELY 2118 01:08:41,640 --> 01:08:43,720 SEEING THAT AS AN A.I. TASK AS 2119 01:08:43,720 --> 01:08:48,200 MUCH AS IT IS A HUMAN TASK. 2120 01:08:48,200 --> 01:08:49,960 TO READ EMOTION OR NUANCE OF 2121 01:08:49,960 --> 01:08:50,680 BEHAVIOR. 2122 01:08:50,680 --> 01:08:52,640 BUT I THINK IN BROADER STROKES, 2123 01:08:52,640 --> 01:08:54,440 I WOULD SAY THAT THE REAL 2124 01:08:54,440 --> 01:08:57,200 QUESTION THAT WE HAVE TO FIGURE 2125 01:08:57,200 --> 01:08:59,400 OUT IS WHERE DOES COMPUTING STOP 2126 01:08:59,400 --> 01:09:02,480 AND WHERE DOES HUMAN DOCTORING 2127 01:09:02,480 --> 01:09:02,880 PICK UP. 2128 01:09:02,880 --> 01:09:07,840 AND HOW DO WE, ALONG THOSE SAME 2129 01:09:07,840 --> 01:09:09,800 LINES, TRAIN THE HUMAN DOCTORS 2130 01:09:09,800 --> 01:09:11,680 FOR THE SKILLSET THAT ARE GOING 2131 01:09:11,680 --> 01:09:13,680 TO BEST NEED TO DO WHAT 2132 01:09:13,680 --> 01:09:15,640 DOCTORING REQUIRES IN THE 21ST 2133 01:09:15,640 --> 01:09:15,960 CENTURY. 2134 01:09:15,960 --> 01:09:17,160 I THINK WE'RE A LONG WAY FROM 2135 01:09:17,160 --> 01:09:22,440 HAVING FIGURED THOSE THINGS OUT. 2136 01:09:22,440 --> 01:09:26,840 >> SO YOU MENTIONED THE 2137 01:09:26,840 --> 01:09:30,160 RETINOPATHY OF PREMATURITY AND 2138 01:09:30,160 --> 01:09:31,520 DIABETIC RETINOPATHY. 2139 01:09:31,520 --> 01:09:34,480 ARE THERE ANY OTHER EYE DISEASES 2140 01:09:34,480 --> 01:09:40,600 WHERE A.I. COULD BE PROMISING? 2141 01:09:40,600 --> 01:09:42,200 >> THERE ARE SO MANY DIFFERENT 2142 01:09:42,200 --> 01:09:45,480 EYE DISEASES THAT ARE BASED ON A 2143 01:09:45,480 --> 01:09:46,760 SIMILAR PARADIGM WHERE YOU LOOK 2144 01:09:46,760 --> 01:09:51,800 AT AN IMAGE AND YOU MAKE A 2145 01:09:51,800 --> 01:09:53,960 DIAGNOSIS, OR EVEN SOME WHERE 2146 01:09:53,960 --> 01:09:55,560 YOU DON'T LOOK AT AN IMAGE AND 2147 01:09:55,560 --> 01:09:56,040 YOU MAKE A DIAGNOSIS. 2148 01:09:56,040 --> 01:10:01,160 SO THERE ARE MANY, MANY 2149 01:10:01,160 --> 01:10:02,960 APPLICATIONS. 2150 01:10:02,960 --> 01:10:04,640 AS YOU CAN IMAGINE, I WOULD 2151 01:10:04,640 --> 01:10:06,400 ALMOST SAY COUNTLESS PAPERS THAT 2152 01:10:06,400 --> 01:10:07,760 HAVE BEEN PUBLISHED WITHIN THE 2153 01:10:07,760 --> 01:10:13,280 PAST FIVE YEARS. 2154 01:10:13,280 --> 01:10:14,200 BUT I'D SAY THE MOST COMPLOB 2155 01:10:14,200 --> 01:10:16,920 BLINDING COMMON ARE 2156 01:10:16,920 --> 01:10:20,840 GOING TO BE THINGS LIKE DIABETIC 2157 01:10:20,840 --> 01:10:22,360 RETINOPATHY, AGE-RELATED MACULAR 2158 01:10:22,360 --> 01:10:23,640 DEGENERATION AND GLAUCOMA. 2159 01:10:23,640 --> 01:10:30,120 WITHIN THOSE AREAS, THERE'S BEEN 2160 01:10:30,120 --> 01:10:31,640 A.I. WORK IN ALL OF THEM AT 2161 01:10:31,640 --> 01:10:33,080 VARYING LEVELS OF MATURITY. 2162 01:10:33,080 --> 01:10:36,200 I WOULD HIGHLIGHT WITHIN THE 2163 01:10:36,200 --> 01:10:40,560 NATIONAL EYE INSTITUTE, EMILY 2164 01:10:40,560 --> 01:10:41,560 CHU -- THEY'VE DONE SOME NICE 2165 01:10:41,560 --> 01:10:42,840 WORK FOR MACULAR DEGENERATION 2166 01:10:42,840 --> 01:10:46,400 WORK. 2167 01:10:46,400 --> 01:10:47,480 IT'S GOING TO BE HOW ARE WE 2168 01:10:47,480 --> 01:10:48,480 GOING TO TAKE THOSE AND APPLY 2169 01:10:48,480 --> 01:10:53,800 THEM TO REAL WORLD CARE. 2170 01:10:53,800 --> 01:10:55,160 IT'S GOING TO BE FINDING WHAT 2171 01:10:55,160 --> 01:10:56,480 THE FDA WOULD CALL INDICATIONS 2172 01:10:56,480 --> 01:10:57,880 FOR USE. 2173 01:10:57,880 --> 01:10:59,120 >> I THINK WE'LL WRAP IT UP WITH 2174 01:10:59,120 --> 01:11:04,480 A QUESTION FROM PETER KILMARKS. 2175 01:11:04,480 --> 01:11:05,760 HE SAYS GREAT TALK, MIKE. 2176 01:11:05,760 --> 01:11:10,240 CAN YOU COMMENT ON INTERNATIONAL 2177 01:11:10,240 --> 01:11:10,640 COLLABORATIONS? 2178 01:11:10,640 --> 01:11:12,880 SHOULD GETTING DIVERSE MATERIAL 2179 01:11:12,880 --> 01:11:15,160 INCLUDE POPULATIONS IN OTHER 2180 01:11:15,160 --> 01:11:16,760 COUNTRIES? 2181 01:11:16,760 --> 01:11:19,200 ALSO ARE THERE APPLICATIONS IN 2182 01:11:19,200 --> 01:11:22,600 RESOURCE-POOR SETTINGS WITH 2183 01:11:22,600 --> 01:11:23,520 LIMITED MEDICAL CARE IN THE U.S. 2184 01:11:23,520 --> 01:11:25,800 AND ABROAD? 2185 01:11:25,800 --> 01:11:27,360 >> YEAH, PETER, THANKS FOR 2186 01:11:27,360 --> 01:11:29,240 ASKING THAT. 2187 01:11:29,240 --> 01:11:31,040 I THINK THERE ARE TWO PARTS TO 2188 01:11:31,040 --> 01:11:31,680 YOUR QUESTION. 2189 01:11:31,680 --> 01:11:33,560 THERE'S A VALIDATION PART AND 2190 01:11:33,560 --> 01:11:35,280 THERE'S A APPLICATION PART. 2191 01:11:35,280 --> 01:11:39,920 IN THE VALIDATION PART, I THINK 2192 01:11:39,920 --> 01:11:40,920 DEFINITELY YES. 2193 01:11:40,920 --> 01:11:43,480 ONE OF THE -- ONE WAY TO LOOK AT 2194 01:11:43,480 --> 01:11:48,120 THAT WOULD BE IF WE WANT TO 2195 01:11:48,120 --> 01:11:50,600 VALIDATE THESE SYSTEMS ACROSS 2196 01:11:50,600 --> 01:11:53,760 DIFFERENT RACES, FOR EXAMPLE, 2197 01:11:53,760 --> 01:11:56,160 THERE ARE PEOPLE OF DIFFERENT 2198 01:11:56,160 --> 01:11:57,320 RACES IN DIFFERENT COUNTRIES 2199 01:11:57,320 --> 01:11:58,480 THAT WE COULD VALIDATE THESE 2200 01:11:58,480 --> 01:11:59,600 SYSTEMS ON. 2201 01:11:59,600 --> 01:12:01,280 ANOTHER WAY TO LOOK AT 2202 01:12:01,280 --> 01:12:06,520 VALIDATION IS THAT DISEASE ISN'T 2203 01:12:06,520 --> 01:12:08,520 JUST ABOUT WHO HAS THE DISEASE 2204 01:12:08,520 --> 01:12:09,960 AND HOW OLD THE PATIENT IS, WHAT 2205 01:12:09,960 --> 01:12:11,960 TYPE OF PERSON IT IS. 2206 01:12:11,960 --> 01:12:13,280 THERE MAY BE ENVIRONMENTAL 2207 01:12:13,280 --> 01:12:14,760 FACTORS THAT ARE DIFFERENT IN 2208 01:12:14,760 --> 01:12:16,680 ONE SETTING VERSUS ANOTHER 2209 01:12:16,680 --> 01:12:17,120 COUNTRY. 2210 01:12:17,120 --> 01:12:18,280 FOR EXAMPLE, IN THE DISEASE THAT 2211 01:12:18,280 --> 01:12:21,360 I STUDY, RETINOPATHY OF 2212 01:12:21,360 --> 01:12:22,480 PREMATURITY, OXYGEN MANAGEMENT 2213 01:12:22,480 --> 01:12:24,560 IS A BIG PART OF THAT DISEASE, 2214 01:12:24,560 --> 01:12:25,080 AND IT'S MANAGED DIFFERENTLY IN 2215 01:12:25,080 --> 01:12:26,560 THE HIGHLY RESOURCED WORLD 2216 01:12:26,560 --> 01:12:28,400 VERSUS THE LESS HIGHLY RESOURCED 2217 01:12:28,400 --> 01:12:28,880 WORLD. 2218 01:12:28,880 --> 01:12:29,960 THE DISEASE DOESN'T ALWAYS LOOK 2219 01:12:29,960 --> 01:12:30,840 THE SAME. 2220 01:12:30,840 --> 01:12:32,400 SO THERE'S A DIFFERENT COMPONENT 2221 01:12:32,400 --> 01:12:35,320 OF VALIDATION. 2222 01:12:35,320 --> 01:12:36,400 THE SECOND PART OF YOUR 2223 01:12:36,400 --> 01:12:38,120 QUESTION, PETER, WAS DELIVERY OF 2224 01:12:38,120 --> 01:12:38,560 CARE. 2225 01:12:38,560 --> 01:12:41,680 I REALLY HOPE THAT A.I. CAN BE 2226 01:12:41,680 --> 01:12:44,360 ONE OF THOSE AREAS THAT, LOOK, 2227 01:12:44,360 --> 01:12:47,160 IN THIS COUNTRY, I THINK THIS 2228 01:12:47,160 --> 01:12:48,040 PANDEMIC SHOWED US THAT WE'VE 2229 01:12:48,040 --> 01:12:50,680 GOT A LOT OF INEQUITIES IN HOW 2230 01:12:50,680 --> 01:12:52,200 TO DELIVER CARE AND MAYBE A.I. 2231 01:12:52,200 --> 01:12:54,120 CAN BE ONE OF THOSE AREAS THAT 2232 01:12:54,120 --> 01:12:55,560 HELPS US DELIVER BETTER CARE TO 2233 01:12:55,560 --> 01:12:59,080 A WIDER RANGE OF PEOPLE, AND 2234 01:12:59,080 --> 01:13:00,160 THAT SAME THING COULD BE SAID 2235 01:13:00,160 --> 01:13:01,280 FOR OTHER COUNTRIES. 2236 01:13:01,280 --> 01:13:03,080 SO I REALLY HOPE THERE'S 2237 01:13:03,080 --> 01:13:04,120 POTENTIAL FOR A LOT OF WORK THAT 2238 01:13:04,120 --> 01:13:06,520 COULD BE DONE. 2239 01:13:06,520 --> 01:13:11,880 >> ONE MORE QUESTION FROM ASAN 2240 01:13:11,880 --> 01:13:12,480 OOLA. 2241 01:13:12,480 --> 01:13:14,680 ARE THERE ANY INITIATIVES FOR 2242 01:13:14,680 --> 01:13:20,400 USE OF A.I. AND GENOMIC DATA AT 2243 01:13:20,400 --> 01:13:21,920 NEI OR AT OTHER PARTS OF NIH? 2244 01:13:21,920 --> 01:13:26,600 >> YEAH, ASAN, DEFINITELY YES. 2245 01:13:26,600 --> 01:13:30,760 WITHIN THE NEI INTRAMURALLY AN 2246 01:13:30,760 --> 01:13:34,480 ALSO WITHIN OUR BROADER WORLD. 2247 01:13:34,480 --> 01:13:39,200 I THINK YOU CAN LOOK AT A LOT OF 2248 01:13:39,200 --> 01:13:40,840 POSSIBLE APPLICATIONS, BUT MAYBE 2249 01:13:40,840 --> 01:13:47,600 ONE OF THEM IS JUST IN TERMS OF 2250 01:13:47,600 --> 01:13:48,760 ONE OF THE THINGS THAT WE DO IN 2251 01:13:48,760 --> 01:13:51,040 MY PIECE OF THE WORLD, I'M 2252 01:13:51,040 --> 01:13:53,880 SPEAKING PURELY AS A RESEARCHER 2253 01:13:53,880 --> 01:13:55,160 NOW, IS THAT WE TRY TO DEVELOP 2254 01:13:55,160 --> 01:13:57,480 RISK MODELS FOR DISEASE. 2255 01:13:57,480 --> 01:14:01,320 YOU KNOW, BASED ON YOUR 2256 01:14:01,320 --> 01:14:02,600 APPEARANCE TODAY, WHAT ARE YOU 2257 01:14:02,600 --> 01:14:05,000 LIKELY TO LOOK LIKE NEXT WEEK. 2258 01:14:05,000 --> 01:14:06,400 AND WE CAN DEVELOP PREDICTIVE 2259 01:14:06,400 --> 01:14:10,360 MODELS THAT ARE BASED ON THAT. 2260 01:14:10,360 --> 01:14:11,920 JUST PURELY FROM IMAGING DATA. 2261 01:14:11,920 --> 01:14:14,960 WE CAN DEVELOP MODELS BASED ON 2262 01:14:14,960 --> 01:14:16,760 IMAGING DATA PLUS CLINICAL DATA, 2263 01:14:16,760 --> 01:14:18,000 AND I SEE NO REASON WE SHOULDN'T 2264 01:14:18,000 --> 01:14:20,360 BE DEVELOPING RISK MODELS BASED 2265 01:14:20,360 --> 01:14:23,280 ON, FOR EXAMPLE, IMAGE PLUS 2266 01:14:23,280 --> 01:14:24,840 CLINICAL PLUS GENETIC DATA. 2267 01:14:24,840 --> 01:14:27,080 SO I THINK THAT THERE'S A LOT OF 2268 01:14:27,080 --> 01:14:28,880 POTENTIAL APPLICATIONS OF THAT 2269 01:14:28,880 --> 01:14:30,680 THAT COULD BE REALLY EXCITING 2270 01:14:30,680 --> 01:14:33,680 BOTH FROM A A PREDICTIVE VIEW D 2271 01:14:33,680 --> 01:14:34,960 ALSO MAYBE FROM A MECHANISTIC 2272 01:14:34,960 --> 01:14:35,760 VIEW. 2273 01:14:35,760 --> 01:14:37,120 >> OKAY. 2274 01:14:37,120 --> 01:14:38,920 SO THE HITS JUST KEEP ON COMING. 2275 01:14:38,920 --> 01:14:42,880 WE HAVE ONE FROM LUKE NELSON, 2276 01:14:42,880 --> 01:14:45,400 WHO'S A TRAINEE AT NIH. 2277 01:14:45,400 --> 01:14:50,040 , WHO ASKS, WHAT ADVICE DO YOU 2278 01:14:50,040 --> 01:14:52,600 HAVE FOR A MEDICAL STUDENT IN 2279 01:14:52,600 --> 01:14:54,080 OPHTHALMOLOGY WHO'S DOING SOME 2280 01:14:54,080 --> 01:14:57,520 CODING AND MIGHT BE INTERESTED, 2281 01:14:57,520 --> 01:15:01,440 WHAT'S A GOOD FUTURE DIRECTION 2282 01:15:01,440 --> 01:15:02,880 SKILLSET TO START TO BUILD THAT 2283 01:15:02,880 --> 01:15:03,880 WILL HELP SOMEBODY IN THE 2284 01:15:03,880 --> 01:15:07,320 FUTURE? 2285 01:15:07,320 --> 01:15:09,640 >> LUKE, I'M GLAD YOU ASKED THAT 2286 01:15:09,640 --> 01:15:11,840 QUESTION BECAUSE I THINK WE NEED 2287 01:15:11,840 --> 01:15:15,800 MORE PEOPLE IN MEDICINE WHO CAN 2288 01:15:15,800 --> 01:15:18,600 UNDERSTAND THE INTERSECTION 2289 01:15:18,600 --> 01:15:20,800 BETWEEN DATA SCIENCE AND 2290 01:15:20,800 --> 01:15:23,200 CLINICAL MEDICINE. 2291 01:15:23,200 --> 01:15:24,800 I THINK THE BEST WAY TO ANSWER 2292 01:15:24,800 --> 01:15:27,680 THAT CONCISELY WOULD BE THAT I 2293 01:15:27,680 --> 01:15:30,680 THINK THAT THE PATHWAY TO GET 2294 01:15:30,680 --> 01:15:32,280 THERE IS GOING TO DEPEND ON 2295 01:15:32,280 --> 01:15:34,280 WHERE YOU'D LIKE TO BE 2296 01:15:34,280 --> 01:15:35,040 EVENTUALLY. 2297 01:15:35,040 --> 01:15:39,640 I THINK THAT ON ONE IF I WERE TO 2298 01:15:39,640 --> 01:15:41,480 GO ON A SPECTRUM, ONE END WOULD 2299 01:15:41,480 --> 01:15:43,760 BE THAT YOU COULD BE THE 2300 01:15:43,760 --> 01:15:44,840 ELECTRONIC HEALTH RECORD PERSON 2301 01:15:44,840 --> 01:15:47,480 WITHIN YOUR CLINICAL PRACTICE AS 2302 01:15:47,480 --> 01:15:48,360 A DOCTOR, SOMEWHERE ELSE IN THE 2303 01:15:48,360 --> 01:15:49,560 MIDDLE WOULD BE THAT YOU COULD 2304 01:15:49,560 --> 01:15:53,840 BE THE PERSON WHO COLLABORATES 2305 01:15:53,840 --> 01:15:56,120 WITH ENGINEERS OR SCIENTISTS TO 2306 01:15:56,120 --> 01:16:00,920 APPLY MECHANISMS TO HEALTHCARE. 2307 01:16:00,920 --> 01:16:02,200 ON THE OTHER END OF THE 2308 01:16:02,200 --> 01:16:04,360 SPECTRUM, YOU COULD BE THAT 2309 01:16:04,360 --> 01:16:06,280 SCIENTIST WHO DEVELOPS THOSE 2310 01:16:06,280 --> 01:16:07,280 METHODOLOGIES AND THEN APPLIES 2311 01:16:07,280 --> 01:16:08,520 THEM TO HEALTHCARE, AND I THINK 2312 01:16:08,520 --> 01:16:09,920 THAT DEPENDING ON WHERE YOU ARE 2313 01:16:09,920 --> 01:16:12,880 IN THAT SPECTRUM WOULD REQUIRE 2314 01:16:12,880 --> 01:16:15,640 MORE OR LESS METHODOLOGICAL 2315 01:16:15,640 --> 01:16:17,120 TRAINING IN THINGS LIKE COMPUTER 2316 01:16:17,120 --> 01:16:21,360 SCIENCE AND DATA SCIENCE. 2317 01:16:21,360 --> 01:16:22,720 >> I THINK WE'RE GOING TO WRAP 2318 01:16:22,720 --> 01:16:23,240 IT UP. 2319 01:16:23,240 --> 01:16:27,840 THANK YOU FOR A -- HOLD ON. 2320 01:16:27,840 --> 01:16:28,040 YES. 2321 01:16:28,040 --> 01:16:29,480 THANK YOU FOR A WONDERFUL 2322 01:16:29,480 --> 01:16:29,880 SESSION. 2323 01:16:29,880 --> 01:16:33,080 WE HAD 200-PLUS VIEWERS OUT 2324 01:16:33,080 --> 01:16:36,960 THERE IN VIDEOCAST-LAND, AND ON 2325 01:16:36,960 --> 01:16:40,000 BEHALF OF DR. GOTTESMAN, ALL OF 2326 01:16:40,000 --> 01:16:42,320 NIH, OUR PARTNERS IN THE OFFICE 2327 01:16:42,320 --> 01:16:46,440 OF EQUITY, DIVERSITY AND 2328 01:16:46,440 --> 01:16:48,000 INCLUSION AND THE JEANG FAMILY, 2329 01:16:48,000 --> 01:16:50,960 WE THANK YOU VERY MUCH FOR A 2330 01:16:50,960 --> 01:16:51,440 WONDERFUL TALK. 2331 01:16:51,440 --> 00:00:00,000 >>THANK YOU VERY MUCH, ROLAND.