I'M ADAM THOMAS. ON BEHALF OF THE DATA SCIENCE TEAM I WELCOME JASON PRIEM, FROM UNIVERSITY OF NORTH CAROLINA CHAPEL HILL, HE MET HEATHER AT A HACKATHON, THEY CREATED IMPACTSTORY, WHICH DELIVERED PRODUCTS JASON IS GOING TO TALK TO YOU INCLUDING IMPACTSTORY, UNPAYWALL AND OADOI. WITH THAT, JASON. [APPLAUSE] >> THANKS FOR THE INTRODUCTIONR AND BRIEF, WHICH GIVES US MORE TIME TO TALK ABOUT SCIENCE STUFF WHICH I LOVE. I SHOULD HAVE SOME SLIDES UP HERE HOPEFULLY SOON. ALL RIGHT. FANTASTIC. THIS IS MY TITLE. I ADDED A LITTLE. I THOUGHT A LOT OF Y'ALL ARE PRETTY ON BOARD WITH OPEN SCIENCE, THE OPEN SCIENCE IDEA ALREADY. I WANTED TO TALK A LITTLE BIT, THOSE OF US EXCITED ABOUT OPEN SCIENCE TALK ABOUT ALL THE POTENTIAL WE HAVE, WHICH IS GREAT. SOMETIMES IT CAN BE NICE TO LIGHT A FIRE UNDER PEOPLE TO SAY NOT ONLY DO WE HAVE ALL THESE GOOD THINGS THAT WE COULD BE GETTING, WE ALSO HAVE GOT A CATASTROPHE THAT WE CAN AVERT. SOMETIMES THAT CAN HELP LIGHT A FIRE. WE'LL TALK ABOUT THAT, REALLY EXCITED ABOUT THAT. SPECIFICALLY, WE'RE GOING TO GO OVER A BRIEF HISTORY OF SCIENCE, OBVIOUSLY WE'LL HAVE TO KEEP IT BRIEF BECAUSE WE'VE ONLY GOT 50 MINUTES. WE'LL TALK ABOUT SCIENCE DOOMSDAY, NOT MY TERM. OPEN SCIENCE, THE SECOND SCIENCE REVOLUTION, FINALLY TALK ABOUT OPEN SCIENCE TODAY WITH A FOCUS ON HOW YOU GUYS WITH HELP, HOW WE CAN HELP AND EVERYBODY LISTENING CAN HELP. LET'S TALKED ABOUT FACTS AND MODELS. LET'S ALSO TALK ABOUT THE MOST IMPORTANT SCIENTIFIC INSTRUMENT, METCALFE'S LAW AND SUCCESS. THE FIRES SCIENTIFIC REVOLUTION HAPPENED 16th CENTURY, A BIG DIFFERENCE IN THE WAY WE THOUGHT ABOUT SCIENCE, JOHN DON HAD A NICE POEM WILL THE OLD AND NEW PHILOSOPHY. AT THE TIME OF THE SCIENCE REVOLUTION, SCIENTISTS WERE CONSIDERED PHILOSOPHER. THE NEW WAY OF THINKING, WHAT MADE IT SPECIAL? ARISTOTLE SAID I SEE ROCKS LYING AROUND, THE NATURAL STATE IS LAYING AROUND. SO I'M NOT GOING TO DO A LOT OF OTHER OBSERVATIONS. I'VE SEEN WHAT I NEED TO SEE. NOW I'M GOING TO THINK WHY WOULD THAT BE, WHAT DOES THAT MEAN FOR THE WORLD, CREATING CREATE A COSMOLOGIC SYSTEM. IT WANTS TO RETURN TO ITS NATURAL STATE. SCIENTIFIC APPROACH IS STILL INTERESTED IN CREATING MODELS, INTERESTED IN UNDERSTANDING THE FACTS AND BUILDING THEORETICAL LENSES WE CAN VIEW THE WORLD THROUGH. WE'VE GOT A FAMOUS MODEL, UNIVERSITY GRAVITATIONAL LAW FROM NEWTON. THE DIFFERENCE BETWEEN THAT AND OLD PHILOSOPHY, WE'RE ALSO INTERESTED IN GATHERING LOTS OF FACTS. GALILEO, DROPPING CANNONBALLS FROM THE TOWER OF PISA MAY OR MAY NOT HAVE HAPPENED. HE WAS DROPPING CANNONBALLS DOWN THE RAMPS, THOUSANDS OF OBSERVATIONS TO REALIZE THE SIZE DOESN'T MATTER VERY MUCH, LIKE ARISTOTLE TOLD US. WE'VE GOT THE UNION OF THE IDEA, GATHER OODLES OF FACTS AND BUILD MODELS AROUND THOSE FACTS. WHAT'S COOL IS THAT WITH TECHNOLOGY, WE CAN REALLY IMPROVE OUR ABILITY TO GATHER FACTS. YOU CAN SEE SO MUCH LOOKING AT THE STARS. INVENT THIS TELESCOPE, NOW THAT TECHNOLOGY BRINGS US A BUNCH OF EXTRA FACTS. WE CAN SEE BLURS AROUND JUPITER. IT'S SUPPOSED TO BE A PERFECT SPHERE ACCORDING TO ARISTOTLE EVERYTHING IS SPHERICAL. WHAT ARE THESE BUMPS? THAT HELPS US BUILD A NEW MODEL. WE CAN JUST KEEP BUILDING BETTER AND BETTER TOOLS. WHEN WE RUN OUT OF FACTS, BUILD A BIGGER TELESCOPE. WE HAVE THE HUBBELL SPACE TELESCOPE, WE CANNOT ONLY LOOK AT JUPITER BUT DEEP INTO THE BEGINNINGS OF THE UNIVERSE. KEEP BUILDING MORE EXPENSIVE AND FANCIER AND MORE IMPRESSIVE TOOLS. THERE'S NOT ANY CONCEPTUAL END TO OUR ABILITY TO GATHER FACTS ABOUT THE WORLD. RIGHT? THERE'S NO PARTICULAR REASON THAT HAS TO STOP. EVENTUALLY WE'LL RUN OUT OF ATOMS IN THE UNIVERSE OR SOMETHING, BUT REALISTICALLY WE CAN KEEP BUILDING BETTER AND MORE IMPRESSIVE TOOLS. THERE'S PLENTY OF TELESCOPES PLANNED BIGGER THAN HUBBELL, THEY WILL SEE MORE THINGS. NO REASON THAT HAS TO STOP. HOW DO WE GET MORE MODELS? THERE'S TWO PARTS. IN SCIENCE WE THINK ABOUT GATHERING OF FACTS BECAUSE THAT WAS SUCH AN IMPORTANT PART OF THE SCIENTIFIC REVOLUTION BUT THE BUILDING OF MODELS IS IMPORTANT TOO. WE'VE GOT A MESS OF FACTS, A BAG OF FACTS, IT'S NOT USEFUL OR SCIENCE. HOW DO WE BUILD MORE MODELS, BUILD THAT? FACT MACHINES ARE EASY TO BUILD. CONTINUING TO BUILD MORE AND MORE, BETTER AND BETTER FACT MACHINES, BUT MODEL MACHINES ARE HARD TO BUILD. THE ONLY THING FOR THE HISTORY OF SCIENCE THAT WE'VE REALLY KNOWN THAT COULD BUILD MODELS RELIABLY IS IN HERE, BETWEEN YOUR EARS. WE DON'T KNOW HOW TO MAKE THAT BETTER VERY WELL. CAN YOU TRAIN A PERSON FOR LONGER BUT YOU CAN'T GET QUALITATIVE IMPROVEMENTS THE SAME WAY YOU CAN BUILD A MUCH BIGGER TELESCOPE. E NEED MORE FACTS. WE BUILD BETTER FACT MACHINES. IF WE NEED MORE MODELS WE'VE GOT TO ADD HUMAN BRAINS, BRING MORE HUMAN BRAINS IN TO GIVE US PROCESSING POWER. WHAT'S THE MOST IMPORTANT SCIENTIFIC INTERESTING OF INSTRUMENT, TAKE A GUESS, THE MOST IMPORTANT SCIENTIFIC INSTRUMENT EVER INVENTED? >> BRAINS. >> NOT INVENTED BUT VERY IMPORTANT, AGREED. BUT INVENTED. YEAH? >> WRITING. >> WRITING, YEAH. >> (INAUDIBLE). >> YEAH, OKAY. I'M GOING WITH THE PRINTING PRESS. OVERALL THAT'S PROBABLY THE MOST IMPORTANT INSTRUMENT EVER INVENTED, THAT'S WHAT WE USE TO COORDINATE THE BRAINS. IF WE'RE IN THE ROOM, YOU CAN TALK. MILLIONS, EVEN THOUSANDS OF PEOPLE WORKING ON A PROBLEM YOU NEED TECHNOLOGY TO SPLIT THE JOB AND BRING MORE BRAINS TO THE PARTY. FACTS IS THE EASY PART. BRINGING THE BRAINS IS THE HARD PART. WE'VE RELIED ON THAT MACHINE, ON THIS TYPE OF USE OF THE MACHINE, THIS IS THE FIRST PURELY SCIENTIFIC JOURNAL PUBLISHED BY OLDENBERG, 1665, PHILOSOPHIC TRANSACTION, THEY DIDN'T BELIEVE IN BREVITY. THEY HAVE A COOL JOURNAL, PRINTING PRESS, THAT'S HOW WE'RE LINKING BRAINS TOGETHER. IT'S REALLY POWERFUL. I GUESS IT WOULD HAVE BEEN LIKE THE LATE 19th CENTURY, METCALFE CAME UP WITH THIS OBSERVATION LOOKING AT TELEPHONE NETWORKS. ONE TELEPHONE IS NOT USEFUL. WHAT CAN YOU DO WITH ONE? TWO IS COOL, I CAN CALL MY FRIEND. AS I CONTINUE ADDING TELEPHONES TO THE NETWORK, I GET THIS COOL EFFECT THAT I GET ONE CONNECTION FROM TWO. FROM FIVE I GET WHAT IS THAT, SOMETHING LIKE 11 OR WHATEVER. THE POINT IS THAT THIS, THESE LINES, THESE CONNECTIONS, SCALE DIFFERENTLY THAN THE NUMBER OF NOTES, EDGES, SCALES FASTER THAN THE NUMBER OF NODES. IT SQUARES WITH APPROXIMATELY HALF THE SQUARE. AS YOU ADD NODES, IT GOES UP. WHAT THAT MEANS IS AS YOU ADD CONNECTIONS TO NETWORK, THE VALUE OF THE NETWORK INCREASES POLY NUMERALLY, IN AN ACCELERATING WAY MORE OR LESS EXPONENTIALLY. AS WE ADD SCIENTISTS TO THE NETWORK, THE NUMBER OF CONNECTIONS BETWEEN SCIENTISTS GROWS AND WE SEE MORE GROWTH, THE JOURNAL NETWORK IS THE CENTERPIECE OF THAT. THE JOURNAL NETWORK THAT WE HAVE, BASED ON THE PRINTING PRESS, IS ONE OF THE MOST VALUABLE THINGS WE'VE EVER CREATED AS A SPECIES BECAUSE THE MORE SCIENTISTS WE ADD, THE MORE CONNECTIONS WE ADD, THE JOURNAL CREATES THE CONNECTIONS. THE JOURNAL SYSTEM IS COOL BECAUSE IT ALLOWS US TO THROW MORE SCIENTIFIC BRAINS INTO THE FRAY, WHICH ALLOWS US TO INCREASE OUR MODELING POWER. UNSURPRISINGLY, BECAUSE OF THAT, WE'RE ABLE TO CONTINUE, THIS IS THE GROSS WAY OF MEASURING SCIENTISTS, ALWAYS TOUGH TO SAY WHAT COUNTS AS A SCIENTIST OR NOT . GLOBAL Ph.D.s, WE CAN SEE BECAUSE OF THE PRINTING SYSTEM, THE JOURNAL SYSTEM, WE'RE ABLE TO CONTINUE GROWING SCIENTISTS INTO THE PROBLEM VERY, VERY QUICKLY. WE'RE CONTINUING TO BE ABLE TO ADD MORE AND MORE SCIENTISTS. IN FACT, THIS ACTUALLY IS AN EXPONENTIAL GROWTH. THE FATHER OF SCIENCE METRICS FOUND ALL THESE COOL THINGS ABOUT THE GROWTH OF SCIENCE. ONE THING HE FOUND, THE NUMBER OF SCIENTISTS DOUBLES EVERY 18 YEARS. THAT'S TRUE NOW, IT'S TRUE AT THE BEGINNING OF SCIENCE. IT'S BEEN TRUE FOR THE HISTORY OF SCIENCE. ONE OF THE COOL PROPERTIES OF THAT, BETWEEN 80 AND 90% OF SCIENTISTS IN HISTORY ARE CURRENTLY ALIVE. THE VAST MAJORITY OF ALL SCIENTISTS IN HISTORY ARE CURRENTLY WORKING TODAY. THAT'S ALWAYS BEEN TRUE, INTERESTINGLY. THAT WAS TRUE IN 18th CENTURY AS MUCH AS TODAY. WE CAN SEE ENORMOUS GROWTH SUPPORTED BY THIS COMMUNICATION NETWORK. WE CAN ALSO SEE A LARGE GROWTH IN THE NUMBER OF JOURNALS. WE CAN LOOK IN THREE DIFFERENT PHASES. THIS IS EXPONENTIAL GROWTH, A LOG SCALE. WE'RE SEEING GROWTH IN THE RATE OF GROWTH. THIS IS EARLY DATE. I GOT THIS EARLY DATE, RIGHT, KIND OF BEFORE PRINTING WAS WIDESPREAD. WE GOT THIS INDUSTRIALIZED ERA HERE WHEN PRINTING WAS VERY WIDESPREAD. AND MASSIVE ACCELERATION BETWEEN THE WARS AND BIG SCIENCE, HEAVILY GOVERNMENT-SUPPORTED SCIENCE. AGAIN, THIS IS THE NUMBER OF JOURNAL ARTICLES, NOT SURPRISINGLY WE SEE THAT GROWING. WHAT'S INTERESTING, THE NUMBER OF SCIENTISTS DOUBLES EVERY 18 YEARS, NUMBER OF JOURNAL ARTICLES DOUBLES EVERY 9 YEARS. MORE IN THE LAST NINE YEARS PUBLISHED THAN THE HISTORY OF HUMANKIND UP UNTIL THAT TIME. IF YOU THINK YOU HAVE A LOT OF PAPERS TO READ, THAT MAY OF MAY BE WITH. IT'S WORKING. WE CAN BUILT FACTS, USE THE JOURNAL SYSTEM TO BRING THE BRAINS TO THE PARTY, IT'S ACTUALLY SUCCEEDED. THIS IS A GRAPH OF CROP YIELDS PER ACRE, OF INTEREST TO PEOPLE WHO DON'T WANT TO DIE, HOW CAN WE MAKE MORE FOOD, WE'VE BEEN WORKING ON THIS A LONG TIME. YIELD PER ACRE GROWS AS PEOPLE DO SCIENCE ON HOW WE CAN GROW MORE YIELD PER ACRE. WE'RE LOOKING AT THE NUMBER OF YEARS OF LIFE SAVED PER THOUSAND PEOPLE, RIGHT? IF YOU CAN MAKE ONE SINGLE PERSON LIVE ONE EXTRA YEAR, THAT'S ONE YEAR OF LIFE SAVED FOR A THOUSAND PEOPLE, ADD IT UP. WE CAN SEE THE DEATH RATE GOES DOWN THE NUMBER OF YEARS PER THOUSAND PEOPLE HAS GONE UP. AND THEN FINALLY A FAMOUS ONE, MOORE'S LAW, WHERE THE NUMBER OF SEMICONDUCTORS WE CAN PUT ON A COMPUTER CHIP DOUBLES EVERY 18 MONTHS. AGAIN, THAT'S EXPONENTIAL CHANGE, RIGHT? THAT'S VERY, VERY POWERFUL. THIS LOOKS LINEAR BUT AS YOU CAN SEE ON THE LEFT-HAND SIDE THAT'S A LOG SCALE. EVERY TWO MONTHS IT DOUBLES. THAT CREATES A CHANGE VERY, VERY, VERY FAST. THAT NUMBER OF TRANSISTORS ON A CHIP TRANSLATES TO HOW FAST YOUR COMPUTER GOES. YOU SEE CONSTANT IMPROVEMENTS, EVERYONE POINTS OUT THAT THING IN MY POCKET HAS MORE COMPUTER POWER THAN WE WOULD HAVE NEEDED TO SEND PEOPLE TO THE MOON. IT GROWS VERY, VERY, VERY, FAST. SO THAT'S ALL THE GOOD NEWS. IT DOESN'T LIKE THERE'S PROBLEMS, LIKE WE NEED OPEN SCIENCE OR ANYTHING TO BE CHANGED. THAT'S NOT THE CASE. SCIENCE CANNOT CONTINUE TO GROW AT THE RATE IT'S GROWN. DEREK PRICE POINTED OUT, COINED THE IDEA OF SCIENCE DOOMSDAY, AT SOME POINT WE'LL HAVE SCIENCE DOOMSDAY, FOR AS SIMPLE A REASON THE NUMBER OF SCIENTISTS INCREASING OF FASTER THAN THE POPULATION. SO JUST CONCEPTUALLY, AT SOME POINT WE CAN'T MAKE EVERY HUMAN BEING A SCIENTIST. SO THAT LEVEL OF GROWTH CAN'T CONTINUE. BUT THERE'S SOME OTHER MARKERS WE CAN SEE SCIENCE DOOMSDAY IS APPROACHING. BEFORE WE GET INTO THAT I WANT TO PUT SOME CONTEXT AROUND WHY THIS IS HAPPENING. I WANT TO TALK ABOUT THIS BOOK, INFLUENTIAL BOOK IN COMPUTER SCIENCE BY FRED BROOKS, "THE MYTHICAL MAN-MONTH." HE WAS WITH IBM, IN THE HEY-DAY OF BIG BLUE, HUNDREDS OF PEOPLE, GRAY FLANNEL SUITS, TYPING ON IBM, MAKING BIG, BIG COMPLICATED SYSTEMS FOR THE GOVERNMENT. HE WAS MANAGING A PARTICULAR JOB, A PARTICULAR CONTRACT, IBM WAS TRYING TO MEET, SOFTWARE. HE NOTICED IT WAS A BIT BEHIND. HAD HE LET'S SAY 100 PEOPLE WORKING ON IT. A BIT BEHIND. WITH 100 PEOPLE I CAN GET HALF THE WORK DONE IN SIX MONTHS. OR WHATEVER. I'M JUST GOING TO ADD ANOTHER 100 PEOPLE AND GET THE WHOLE WORK DONE IN SIX MONTHS, STRAIGHTFORWARD. WE SAY PERSON-MONTH TODAY. THAT'S HOW MUCH WORK THERE IS. I ADD THESE, SHOVEL IN BRAINS, RIGHT? SAME AS WITH SCIENCE, ADD BRAINS TO THE PROBLEM. EVENTUALLY WE'LL GET IT DONE. HE DISCOVERED SOMETHING REALLY UNNERVING, WHICH WAS AS HE ADDED PEOPLE, HE SAW DIMINISHING RETURNS ON EVERY ADDITIONAL PERSON ADDED TO THE POINT EVENTUALLY ADDING PEOPLE MADE THE THING GO SLOWER. WELL, THAT IS REALLY CONCERNING. HOW DO WE MAKE IT GO FASTER THEN? SO THE CONCEPTUAL MATH, HE WAS TALKING ABOUT COMBINATORIAL EXPLOSIONS, AS YOU ADD, NON-LINEAR EFFECTS. THIS IS THE DARK SIDE, THE FLIP SIDE OF METCALFE'S LAW, THE VALUE OF THE NETWORK GOES UP AS WE ADD NODES, BUT THE COMPLEXITY AND DIFFICULTY OF MAINTAINING ALL OF THOSE EDGES GOES UP IN A NON-LINEAR WAY. THIS IS THE ACTUAL SCALING FACTOR. SO THIS IS A CURVE THAT DEMONSTRATES THAT. AS WE ADD EDGES, COMPLEXITY SCALES IN A NON-LINEAR WAY DIFFICULTY OF MAINTAININGED SCALES WITH BIG IMPLICATIONS FOR SCIENCE, BAD IMPLICATIONS FOR SCIENCE. WE CANNOT INDEFINITELY CONTINUE TO SHOVEL BRAINS INTO SCIENCE, AND EXPECT TO SEE CONTINUED IMPROVEMENTS IN THE PROCESSING POWER THAT WE'RE BRINGING TO SCIENCE. I'LL POINT OUT AGAIN THE NUMBER OF PUBLICATIONS DOUBLES EVERY 18 YEARS. SORRY, NUMBER OF SCIENTISTS DOUBLES EVERY 18 YEARS. NUMBER OF PUBLICATIONS DOUBLES EVERY NINE YEARS. CONCEPTUALLY HOW IS THAT GOING TO WORK? IF THERE'S TWICE AS MANY PUBLICATIONS IN NINE YEARS, BUT IT TAKES 18 YEARS TO GET TWICE AS MANY SCIENTISTS THERE'S MORE AND MORE AND MORE PUBLICATIONS PER SCIENTIST, HOW ARE WE GOING WE'LL SAY LET'S HYPERSPECIALIZE, I'LL JUST FOCUS ON THE STUFF I KNOW, RIGHT? SO I'M GOING TO DROP OUT OF THE MY SLIDES, WHICH IS PERHAPS ILL ADVISED. THERE'S AN INTERNET FAMOUS -- ANYBODY SEEN THIS, ILLUSTRATED GUY WITH THE Ph.D.? A COUPLE NODS, SUPER COOL. I'LL WALK YOU THROUGH THIS. IMAGINE THE CIRCLE CONTAINS ALL KNOWLEDGE. ELEMENTARY, HIGH, BACHELOR DEGREE A SPECIALTY. MASTER'S DEGREE DEEPENS THE SPECIALTY. RESEARCH PAPERS TAKES YOU TO EDGE OF HUMAN KNOWLEDGE. THIS ONE PARTICULAR THING YOU'RE ALL THE WAY TO THE EDGE. ONCE YOU'RE ON THE BOUNDARY YOU FOCUS, PUSH AT THE BOUNDARY FOR A FEW YEARS, WE'RE ZOOMED IN. ONE DAY THE BOUNDARY GIVES AWAY. AND THAT DENT YOU'VE MADE IS CALLED A Ph.D. [LAUGHTER] THAT'S THAT TINY DENT YOU'VE MADE. NOW THE BEST PART. SO THIS IS WHAT YOU SEE. THAT'S YOUR WORLD, RIGHT? [LAUGHTER] MOST OF US IN ACADEMIA HAVE BEEN THERE. DON'T FORGET THE BIGGER PICTURE. THAT'S YOUR Ph.D. THERE'S SOMETHING I THINK GLORIOUS ABOUT THIS. I LOVE THE IDEA OF -- SORRY -- OF SO MANY PEOPLE WORKING ON SO MANY TINY LITTLE THINGS BUT INDIVIDUALLY EACH MAKING THEIR OWN CONTRIBUTION. AND THERE'S SOMETHING REALLY NICE ABOUT THAT. BUT IT COMES AT HIGH COST THAT WE DON'T THINK ABOUT MUCH. JOHAN BOLAN DID THIS, WE'RE NOT CONNECTED TO ALL KNOWLEDGE. WE LOSE TRACK OF ALL KNOWLEDGE. WE LOSE TRACK OF KNOWLEDGE NEXT TO US. MOST OF US HERE HAVE DONE RESEARCH, HAVE EXPERIENCED THE PHENOMENON OF ENCOUNTERING SOME BODY OF RESEARCH IN SOME PERHAPS DISTANT RELATED OR CLOSELY RELATED FIELD THAT'S PRETTY MUCH JUST WHAT WE'RE DOING, BUT THAT NO ONE CITES AND WE'VE NEVER HEARD OF. HAS ANYBODY EVER HEARD THAT? YEAH, RIGHT? MOST OF US, RIGHT? WHOA, THESE PEOPLE ARE DOING ALL THIS STUFF. THEY ARE OUT ON THE EDGE OF THE CIRCLE. YOU'RE OUT ON THE EDGE OF THE CIRCLE. YOU DON'T COMMUNICATE, THE NETWORK IS SIMILAR. PEOPLE DOING BRAIN STUDIES HERE, THESE PEOPLE IN CHEMICAL ENGINEERING OVER HERE, THERE'S NOT MUCH COMMUNICATION THERE. SO THE HYPERSPECIALIZATION FRED BROOKS ISSUE OF COMBINATORIAL EXPLOSION IS NOT VERY SUCCESSFUL. AS A RESULT OF THIS COMBINATORIAL EXPLOSION AND MASSIVE INCREASE THE NON-LINEAR GROWING INCREASE OF AMOUNT OF INFORMATION GENERATED, THE FACT THIS NETWORK GROWS AT A NON-LINEAR WAY, WE'RE STARTING TO SEE THE SYSTEM THAT WE CAME UP WITH, AS SCIENTISTS, GROANING AND GETTING READY TO BREAK. THIS IS AN EXAMPLE FROM THE '70s, 2005, HOW MANY PAPERS A RESEARCHER READS PER YEAR, INCREASED SUBSTANTIALLY. BUT THE AMOUNT OF TIME SPENT HAS INCREASED ONLY A LITTLE BIT. IF YOU DIVIDE THAT OUT, IT'S LIKE HALF AN HOUR, 45 MINUTES, SOMETHING LIKE THAT, HEATHER? ORIGINALLY SOMETHING LIKE 45 MINUTES, NOW 30 MINUTES A PAPER. IF IT CONTINUES, IT'S 15 MINUTES, WHAT'S THE POINT? YOU'RE NOT GETTING ANYTHING OF VALUE. YOU'RE SAYING LET'S DO THE SAME THING BUT FASTER, LIKE LUCY AND ETHEL IN THE CHOCOLATE FACTORY. THAT'S GOING TO BREAK. WE CAN SEE THE RESULTS IN THE ACTUAL OUTPUT OF SCIENCE. THIS IS THEORETICAL. WE CAN IMAGINE WE GOT THESE CURVES, DIFFERENT SHAPES, LITERATURE DOUBLES FASTER. WE CAN SEE RESULTS. WHAT'S COOL ABOUT THIS, THIS IS ANOTHER PAPER I GOT FROM RICHARD GOLD WHO DID A PRESENTATION ON SOMETHING RELATED, FANTASTIC. THIS IS THE AMOUNT OF LIFE SAVED PER 100,000 PEOPLE PER PUBLICATION . THE BLUE LINE, YEARS OF PEOPLE'S LIVES BEING SAVED PER PAPER. ALL OF THOSE ARE GOING DOWN. IT MEANS PER-PAPER PRODUCTIVITY IS GOING DOWN. EVEN THOUGH WE'RE GENERATING MORE AND MORE PAPERS, EACH OF THEM DOES LESS AND LESS. AND SO IN ORDER TO SEE THESE GAINS, IN ORDER TO SEE IMPROVEMENTS IN QUALITY OF LIFE SCIENCE IS GENERATING WE'RE HAVING TO SPEND ENORMOUS AMOUNTS OF MONEY AND MANPOWER BECAUSE OF DIMINISHING RETURNS FROM THE EXPLOSION, COMBINATORIAL EXPLOSION OF NUMBER OF EDGES IN THE GRAPH, WE'RE SEEING DIMINISHING RETURNS. THESE ARE THE SAME CROP GRAPHS. BLUE LINE LOOKS FLAT, IT'S RISING BY PERCENT. LOOK HOW MUCH WE HAVE TO SPEND TO GET THAT, GETTING DIMINISHING RETURNS PER RESEARCHER, ADDING THE GREEN LINE TO CONTINUE SEEING THE SAME IMPROVEMENTS. FINALLY THE MOORE'S LAW IS MOST COMPELLING, STRAIGHT LINE, 35%, THAT'S MOORE'S LAW, EXPONENTIAL GROWTH. THIS IS HOW MUCH WE HAVE TO ADD TO RESEARCH TO CONTINUE SEEING THAT GROWTH. IT'S DIMINISHING RETURNS PER RESEARCHER, DIMINISHING RETURNS PER PAPER, THE SIGN WE'RE STARTING TO RUN IN TO THIS SOUND BARRIER, THAT COMES FROM DOING THINGS THE WAY WE'VE BEEN DOING IT, FROM THE LEGACY OF 300 YEARS OR MORE OF SCIENTIFIC REVOLUTION, WE'RE STARTING TO SEE THAT STOP WORKING. AND THAT SHOULD BE A SCARY THOUGHT FOR THOSE INVOLVED IN RESEARCH ENTERPRISE. A LOT OF WHAT WE DO IS PREDICATED ON THE IDEA AS LONG AS WE KEEP KIND OF FOLLOWING IN THE FOOTSTEPS OF GALILEO, NEWTON, WHO STARTED THE SCIENTIFIC BALL ROLLING SEVERAL HUNDRED YEARS AGO, AS LONG AS WE CONTINUE THAT'S GOING TO KEEP WORKING. IT'S NOT GOING TO KEEP WORKING. WE'RE APPROACHING SCIENTIFIC DOOMSDAY. I DON'T WANT TO END ON SAD NEWS. THAT IS, WE HAVE SOME STRATEGIES THAT WE CAN USE TO DO A SECOND -- CREATE A SECOND SCIENTIFIC REVOLUTION, TO DO THINGS DIFFERENT THAN GALILEO, NEWTON AND THE PEOPLE AFTER THAT DID, AND INSTEAD WE CAN ADD MACHINE POWER, NOT JUST TO THE FACT PART OF THE SCIENTIFIC ENTERPRISE, BUT TO THE MODEL-MAKING PART OF THE SCIENTIFIC ENTERPRISE. THAT'S UNPRECEDENTED. WE DON'T HAVE TO ADD MORE BRAINS. WE CAN GET MORE OUT OF EACH BRAIN. I WANT TO START BY SAYING THIS TALKS ABOUT OPEN SCIENCE, I'M A PASSIONATE BELIEVER IN OPEN SCIENCE, MY WHOLE ACADEMIC CAREER HAS BEEN AROUND ACADEMIC SCIENCE. BUT IT'S NOT ABOUT OPEN. THE POINT OF OPEN SCIENCE ISN'T MAKING EVERYTHING OPEN. THAT'S GREAT. I MEAN, THAT'S WHAT WE WANT TO DO. AND IT HAS LOTS OF GREAT EFFECTS LIKE REGULAR PEOPLE CAN READ IT AND CITIZEN SCIENTISTS CAN JOIN IN, ALL REALLY WONDERFUL. BUT THE DEEPER POINT, THE DEEPER IMPORT OF OPEN SCIENCE IS BUILDING AN ENVIRONMENT, ACTUALLY YOU KNOW WHAT? I'VE GOT A COOL SLIDE FIRST. SCIENCE ISN'T ALWAYS OPEN. THAT'S WHY THE POINT OF OPEN SCIENCE ISN'T OPEN. SCIENCE WAS OPEN BEFORE. WHAT WAS DONE IN EUROPE BEFORE THE SCIENTIFIC REVOLUTION, ALCHEMY, PEOPLE IN UNDERGROUND ROOMS TRYING TO TURN LEAD INTO GOLD, SECRET BOOKS THEY WEREN'T SHARING, ONLY ONCE YOU STARTED SHARING WE WERE ABLE TO GET SCIENCE. MERTON TALKS ABOUT HIS FOUR NORMS OF SCIENCE. THE CENTERPIECE NORM, YOU SHARE YOUR STUFF. AS A SCIENTIST YOU TELL OTHER PEOPLE. THIS WHOLE POINT WE NEED TO RECRUIT OTHER BRAINS TO GET MORE PROCESSING POWER, YOU CAN'T DO THAT IF YOU DON'T TELL THE OTHER BRAINS WHAT YOU'RE DOING. OPENNESS IS COSH CORE TO SCIENCE, ALWAYS HAS BEEN. OPEN SCIENCE IN 2018 IS NOT ABOUT JUST SAYING, OH, WE WANT SCIENCE TO BE OPEN, JUST BECAUSE. BUT RATHER WE WANT SCIENCE TO BE OPEN SO THAT WE CAN BUILD MACHINES WE NEED TO SOLVE THIS SCIENTIFIC DOOMSDAY. WE NEED TO BUILD MACHINES TO MAKE OUR BRAINS ACTUALLY MORE PRODUCTIVE, NOT JUST ADD BRAINS BUT INCREASE THE PRODUCTIVITY OF THE BRAINS AND TO DO THAT WE NEED THE RAW MATERIAL. WE NEED THE MACHINES TO HAVE ACCESS TO THE RAW MATERIAL. AND IN MY VIEW THAT'S THE CORE OF WHAT OPEN SCIENCE IS ABOUT. SO WHAT DOES THIS OPEN SCIENCE WORLD LOOK LIKE? IN BROAD TERMS WE START WITH OPEN EVERYTHING, RIGHT? A LOT OF WORK IN OPEN SCIENCE ABOUT MAKING BIG DATA AND CODE AND SCIENTIFIC ARTICLES OPEN, A LOT OF THAT IS BUILT AROUND TRYING TO START THIS BOTTOM PART GOING, RIGHT? TRYING TO GET THAT SUBSTRATE, THAT RAW MATERIAL, AVAILABLE. AND THEN WE'VE GOT MACHINES THAT THINK. WE'VE GOT MACHINES THAT CAN USE ALL THESE, CREATE INFERENCE, DRAW COOL PICTURES, TEST HYPOTHESES THAT WE'VE HEARD ABOUT. IN FACT, I WAS DOING NOTES ON THIS IN THE UBER ON THE WAY HERE, AND I GOT SOME GOOD ADVICE FROM UBER DRIVER, TOTALLY SILENT THE WHOLE DRIVE. HE'S LIKE, YOU SHOULD PUT SOME A.I. IN THERE. I WAS LIKE, THAT'S GOOD ADVICE, SIR, I THINK I WILL. THIS IS IN HONOR OF MY UBER DRIVER. THIS IS WHERE A.I. GOES, IN THERE. THERE'S A LOT OF MAGIC THAT CAN HAPPEN, COOL THINGS CAN HAPPEN. SELF-DRIVING CARS, IMAGINE SELF-DRIVING SCIENCE, A FUNDAMENTAL CHANGE IN OUR BRAIN'S ABILITY TO WORK. NOT JUST ADDING MORE BRAINS. AND FINALLY GOT A CREATIVE PERSON AT THE TOP. WE'RE NOT AT THE POINT AND NEVER MAY BE WHERE WE WANT TO TAKE HUMANS COMPLETELY OUT OF THE LOOP. WE WANT A MULTIPLIER EFFECT ON WHAT HUMANS ARE ABLE TO DO. THE EXAMPLE IS THIS F-16 WHEN IT WAS CREATED IT WAS MEANT TO BE THE PREMIER FIGHTER AIRCRAFT OR WHATEVER. AND ONE OF THE INTERESTING THINGS ABOUT IT, ONE OF THE FIRST AIRPLANES INHERENTLY UNSTABLE. WHAT THAT MEANS IF YOU WERE TO TRY AND FLY THE AIRPLANE, YOU COULD NOT. IT'S SO UNSTABLE, THAT A HUMAN BEING DOESN'T HAVE FAST ENOUGH REFLEXES TO MAKE IT GO. IT NEEDED A COMPUTER THAT WAS THINKING MUCH, MUCH FASTER THAN A HUMAN BEING TO KEEP IT STABLE. WHEN IT STARTS TO FALL THIS WAY, ADJUST. CONSTANTLY ADJUSTING. COMPUTER CAN DO IT FAST ENOUGH, HUMAN BEING COULDN'T. YOU'RE ABLE TO BUILD AN AIRPLANE TO PERFORM MUCH BETTER THAN OPERATED BY A HUMAN BEING BECAUSE WE'RE RELYING ON THE COMPUTER TO DO WHAT IT'S GOOD AT, MAKING MILLIONS OF DECISIONS REALLY, REALLY QUICKLY, AND LET THE HUMAN DO WHAT IT'S GOOD AT, MAKING DECISIONS, I YOU WANT TO GO TO THIS PLACE, WHATEVER. WE NEED TO DO THE SAME THING FOR SCIENCE. THAT'S WHAT OPEN SCIENCE IS DESIGNED TO EVENTUALLY DO. IN SOME WAYS IT'S BEGINNING TO DO ALREADY. BUILD A SCIENCE SO FAST AND PERFORMING SO WELL, THAT NO HUMAN BEING COULD EVER KEEP UP WITH IT. WE WANT INHERENTLY UNSTABLE. CONSTANTLY IN A STATE OF FLUX, AND IF I AS A HUMAN BEING WERE TRYING TO DO IT, COULDN'T DO IT. THAT REQUIRES A DEGREE OF HUMILITY. TO BE HONEST THAT'S PART OF WHY OPEN SCIENCE IS SOMETIMES CHALLENGING FOR PEOPLE, BECAUSE WE AS HUMANS HAVE TO START TAKING A BACK SEAT, WE'VE REACHED DIMINISHED RETURNS BY THROWING BRAINS AT THE PROBLEM, WE'RE NOT SEEING CONTINUED GAINS ON THAT PHILOSOPHY OR THAT APPROACH, AND WE NEED TO STEP BACK AND SAY HEY, OUR BRAINS ARE NOT STRONG ENOUGH. WE NEED ADAPTIVE TECHNOLOGY FOR OUR BRAIN BECAUSE WE CAN'T DO IT. WE NEED GOOGLE, NOT YAHOO. WHEN THE WEB CAME OUT, LOTS OF WEB PAGES, HARD TO NAVIGATE, HARD TO FIND WHAT YOU WANT. SO I WAS A TEACHER, A MIDDLE SCHOOL TEACHER AT THE TIME. ONE THING WE LEARNED, THEY SAID YOU SHOULD USE THIS SEARCH ENGINE CALLED YAHOO, BACK IN THE DAY. YOU SHOULD USE THIS SEARCH ENGINE YAHOO, BECAUSE THEY HAVE EXPERT HUMAN BEINGS WHO READ THE INTERNET AND FIND THE GOOD PAGES AND THEY SHOW YOU. THAT MADE SENSE. OH, GOOD, I'LL HIRE SOME PERSON TO READ THE WHOLE INTERNET AND FIND THE GOOD PARTS AND THEY WILL SERVE TO US. LIKE PEER REVIEW, EVERY PIECE OF LITERATURE WILL BE READ BEFORE IT'S PUBLISHED, CHECK OFF, CHECK NO, THAT WILL WORK. IT'S RIDICULOUS, RIGHT? AS THE WEB GROWS THE IDEA OF A PERSON OR HUGE GROUP OF PEOPLE READING IN ADVANCE AND FILTERING MANUALLY IS ABSURD. YOU NEED TO FILTER ALGORITHMICALLY. PAGE RANK POWERS GOOGLE. LARRY PAGE WAS WORKING ON NSF-FUNDED RESEARCH WHEN HE CAME UP WITH, THAT HAD HE A GRANT TO STUDY THE CITATION GRANT, TRYING TO DISTILL MEANING. HECK, I COULD DO THIS FOR THE INTERNET, IT'S MORE LUCRATIVE. BUT I THINK WHAT'S REALLY COOL IS THIS IS PART OF OUR LEGACY IN SCIENCE, IN SCHOLARSHIP, WE HAVE THIS REALLY COOL LINK GRAPH THAT'S BEEN ESTABLISHED, MACHINE READABLE, YOU CAN DISTILL MEANING. GOOGLE USES THAT TO DISTILL MEANING OUT OF THE WEB. THIS WEBSITE IS LINKED BY PAGES, THAT'S ACTUALLY A DISTRIBUTIVE QUALITY JUDGMENT BY THE ENTIRE INTERNET, I CAN PUT A WEIGHT ON THAT, ESPECIALLY IF VOTES ARE FROM CREDIBLE SOURCES. WE CAN POTENTIALLY DO THE SAME THING FOR SCHOLARSHIP, WE NEED TO MAKE CONNECTIONS, THINK ABOUT INSTEAD OF WRITING STORIES TO EACH OTHER THAT ARE RESEARCH PAPERS, WHICH MADE A LOT OF SENSE IN THE 17th CENTURY, WE NEED INSTEAD TO THINK ABOUT SMALLER TERMS, MAYBE THINKING ABOUT USING MACHINES TO MAKE SOME OF THOSE RELEVANT JUDGMENTS FOR US. SO, THAT'S ALL KIND OF CONCEPTUAL AND EVERYTHING. AND I THINK I COULD PROBABLY DESCRIBE -- I WAS TALKING TO HEATHER WHEN I DID THE TALK, THINKING THAT IT WOULD BE COOL TO DESCRIBE LIKE THE OPEN SCIENCE FUTURE. WOULDN'T IT BE NEAT TO SAY THIS IS WHAT WE COULD EXPECT, A DAY IN THE LIFE OF AN AVERAGE SCIENTIST AT LET'S SAY THE NIH TO BE IN THE WORLD OF OPEN SCIENCE, 10 OR 20 YEARS, WHEN THIS HAPPENS, THIS IS WHAT LIFE WOULD LOOK LIKE. REALLY COOL, LIKE SCIENCE FICTION, RIGHT? YOU DESCRIBE WHAT IT WILL LOOK LIKE. REALISTICALLY I CAN'T DESCRIBE THAT BECAUSE I DON'T KNOW WHAT IT'S GOING TO LOOK LIKE. NOBODY KNOWS WHAT IT'S GOING TO LOOK LIKE. THAT'S THE THING ABOUT REVOLUTION, IT'S HARD TO PREDICT. AMERICAN REVOLUTION, ARE YOU GOING TO GET WASHINGTON, D.C. OR REIGN OF TERROR, IT'S HARD TO TELL GOING IN. WHO KNOWS? IT'S A DISCONTINUITY IN HISTORY, HARD TO TELL AT THE BEGINNING WHAT'S GOING TO HAPPEN. WE DON'T KNOW WHAT'S GOING TO HAPPEN TO OPEN SCIENCE. ALL WE KNOW IS WE NEED MACHINES TO AUGMENT THE HUMAN BRAIN BECAUSE THROWING MORE BRAINS AT THE PROBLEM ISN'T WORKING, BECAUSE WE'VE STARTED TO HIT THE ERA OF DIMINISHING RETURNS. WE STARTED TO HIT PLACE WHERE THE NETWORK IS GROANING UNDER THE WEIGHT OF ALL OF THIS ADDITIONAL SCIENTISTS WE'RE ADDING AROUND ADDITIONAL PUBLICATIONS. THAT'S THE DIRECTION WE NEED TO GO. TALKING ABOUT HILL CLIMBING, IN COMPUTER SCIENCE OR COMPUTER PROGRAMMING A LOT OF TIMES YOU DON'T NECESSARILY KNOW WHAT SOLUTION YOU'RE LOOKING FOR BUT YOU KNOW WHICH WAY IS UP. IF I'M STANDING ON THE SIDE OF A HILL, I DON'T KNOW WHAT THE TOP OF THE HILL LOOKS LIKE BUT IF I LOOK DOWN I'LL GET LOWER, IF I GO HIGH I'LL GET HIGHER. I DON'T THINK WE NEED DETAILED EXTENSIVE MAPS BECAUSE I DON'T THINK ANYONE WHO HAS A DETAILED EXTENSIVE MAP HAS MUCH CREDIBILITY. HOW WOULD THEY KNOW? BUT WE DO HAVE AN IMPERATIVE TO MOVE FORWARD IN THE DIRECTION OF OPEN SCIENCE, IN THE DIRECTION OF MACHINE MEDIATED SCIENCE. SAND . AND IN ORDER TO DO THAT WE NEED A HILL TO CLIMB. I PROPOSE TWO ATTRIBUTES TO MEASURE THE SLOPE OF THE HILL. FIRST IS OPENNESS. BECAUSE IF AS A PERSON, I CAN'T READ YOUR ARTICLE, OF COURSE MY BRAIN DOESN'T GET TO CONTRIBUTE TO THAT MODEL BUILDING. SIMILARLY IN THE WORLD OF MACHINE-MEDIATED SCIENCE, AS A MACHINE, IF I CAN'T GET AT YOUR DATA, OR YOUR PAPER, OR WHATEVER IT IS YOU'VE PRODUCED, THAT'S USELESS TO ME, IN MY ALGORITHM. OPENNESS WILL HAVE TO BE A CORE PRINCIPLE OF FUTURE SCIENCE, FUTURE OPEN SCIENCE, MACHINE-MEDIATED SCIENCE WORLD. A SECOND CORE PRINCIPLE NEEDS TO BE CONNECTION. AGAIN, THIS IS TRUE OF SCIENCE IN THE PAST AS WELL, IT'S ALWAYS BEEN TRUE OF SCIENCE, NEWTON WAS CITING PEOPLE BEFORE. HEY, A LOT OF PEOPLE HAVE HAD SOME OF THESE IDEAS ABOUT GRAVITY BEFORE ME, RIGHT? SO CITATION HAS ALWAYS BEEN AN IMPORTANT PART OF SCIENCE. IT'S A LINK. IT SAYS I'VE GOT THIS IDEA, I'M GOING TO MAKE A CONNECTION DO THIS IDEA. GARFIELD HAD THIS PROFOUND IDEA WE COULD TRAVERSE THIS NETWORK OF CITATIONS AND FIND REALLY INTERESTING KNOWLEDGE. SIMILARLY IN THE ERA OF MACHINE-MEDIATED SCIENCE WE'RE GOING TO NEED TO TRAVERSE NETWORKS, WE NEED LINKING TO DO THAT. I SUGGEST WE NEED OPENNESS AND LINKS, TWO THINGS WE SHOULD LOOK FOR WHEN PEOPLE ARE TALKING ABOUT DOING NEW PROJECTS IN SCIENCE, IMPROVING SCIENCE, DOING SOMETHING THAT'S GOING TO CHANGE THE WAY WE DO SCIENCE, LOOK, OPEN IS A CONNECTOR. I'M GOING TO GO THROUGH THIS AND GIVE EXAMPLES OF STUFF THAT IS OPEN AND CONNECTED, KIND OF COOL AND I'VE GOT TEN MINUTES AND WE'LL HAVE TIME FOR QUESTIONS. ONE OF THE ONES I'M EXCITED ABOUT IS ARGUE MENTATION NETWORKS, VANNEVAR BUSH, AN INTERESTING GUY, HAD THIS IDEA OF A MEMEX, WHEN YOU SAW THE IDEA LIGHT, YOU COULD PUT A SCROLL THERE AND ZOOM IN TO OTHER PEOPLE, INVENTED BASICALLY THE WORLDWIDE WEB IN 1945, JUST DIDN'T HAVE THE TECHNOLOGY TO DO IT. HE SAID WE SHOULD HAVE THIS WEB DEAL, HE SAID WE SHOULD BE ABLE TO REGISTER IDEAS. THAT'S A COOL IDEA, RIGHT? IN SCIENCE WE WORK IN THE REALM OF PAPERS, BUT WHAT WE'RE INTERESTED IN IS IDEAS. I DON'T WANT TO CITE A PAPER. I WANT TO CITE AN IDEA. I DON'T WANT TO READ PAPERS. I WANT TO READ IDEAS. MACHINES CAN EXTRACT THE IDEAS, THIS IS BEING DONE IN A LOT OF FIELDS RECOVER. ALREADY. IN LAW THEY HAVE TOLLMAN NETWORKS, WE CAN TAKE ANY LEGAL ARGUMENT, BREAK IT DOWN INTO ITS CONSTITUENT PARTS, AND WE CAN USE THAT TO HELP US UNDERSTAND WHAT'S BEING ARGUED. SO THIS IS A TOLLMAN STRUCTURE. OTHER PEOPLE HAVE DONE THIS IN ACADEMIA, IN ACADEMIC DISCOURSE, WE CAN DO THE SAME THING. THIS IS AN EXAMPLE OF AN ARGUEMENTATION NETWORK. IN WE CAN EXTRACT FROM THE LITERATURE OR PUBLISH THEM AUTOMATICALLY, INSTEAD OF AS PAPERS, AS NANOPUBLICATIONS WE COULD HAVE A COOL NET WOK OF I COULD DISPLAY THEM IN ANAPERS. INTERESTING WAY. THIS IS AN ONLINE SYSTEM THAT'S MEANT TO GUIDE ARGUMENTS ABOUT NON-SCIENTIFIC, GENERAL INTEREST TOPICS. I COULD HAVE THIS INTERFACE, REPRESENTING EVERY ARGUMENT AS A BOX, COULD HAVE PROS AND CONS, AND FOR ANY PROS AND CONS I COULD DIVE INTO THE SUBPROS AND SUBCONS AND RATE THEM AND BROWSE ARGUMENTS INSTEAD OF BROWSING A PAPER, HELPFUL WHEN I'M DOING SOMETHING LIKE A LIT REVIEW AND VISUALIZE IT AND SAY EVERY ARGUMENT HAS ITS OWN SUBARGUMENT, SUB-SUBARGUMENT, THIS WOULD BE AN INTERESTING WAY TO DIVE INTO THE LITERATURE AND INSTEAD OF VISUALIZE I COULD ASK. SAY, HEY, COMPUTER, WHAT ARE THE GENERAL SCHOOLS OF THOUGHT ON THIS PARTICULAR ISSUE? I COULD GET THE SCHOOLS OF THOUGHT. IF I WANTED TO DIVE IN, I COULD DIVE INTO THEM. ANOTHER I THINK REALLY INTERESTING OPEN SCIENCE, AND I DON'T WANT TO LOSE MY THEME, WE'RE LOOKING FOR OPEN AND CONNECTED, RIGHT? WHAT'S COOL HERE IS THAT WE CAN HAVE A CONNECTED WEB OF ARGUMENTS, BUT IT'S VERY IMPORTANT WHEN ANYBODY IMPLEMENTS THIS STUFF IT'S DONE IN AN OPEN WAY. I THINK WHEN PEOPLE ARE TALKING ABOUT ARGUE MENTATION RESEARCH, THE MAJORITY SAID WE'LL MAKE THIS OPEN. AT SOME POINT I GUARANTEE A MAJOR PUBLISHER WILL SEE WE'VE GOT AN AMAZING NETWORK BUILT ON A THREE MILLION ARTICLES. WE NEED TO SAY AWESOME, GOOD WORK ON THE TECHNICAL SIDE OF THAT. COME BACK WHEN YOU'VE GOT THAT ON THE NETWORK OF 200 MILLION SCIENTIFIC ARTICLES, ALL THE SCIENTIFIC ARTICLES, THAT'S WHAT WE'RE INTERESTED IN. SCIENCE IS NOT A CHOPPED-UP ENTERPRISE. SCIENCE, THERE'S ONLY ONE UNIVERSE. THERE'S ONLY ONE SET OF ACTUAL TRUTHS ABOUT THE UNIVERSE. LIKE WE AS HUMANS MAYBE CAN PERCEIVE ONE WAY OR ANOTHER. THERE'S GOING TO BE DISCOURSE, DIFFERENCES, BUT WE'RE ALL WORKING TOGETHER AS A SPECIES TO SOLVE THESE PROBLEMS. WE DON'T WANT A SILO, WE WANT A FULLY CONNECTED VERSION. ALT-METRICS, WHAT I WAS WORKING ON. HEY, WHY DON'T WE LOOK AT DIFFERENT PLACES A PERSON COULD TALK ABOUT AN ARTICLE INSTEAD OF CITATIONS, LIST THE NAME WHEN A PERSON WAS IN THE CONFERENCE, HAVE A CONVERSATION ABOUT THIS ARTICLE, THIS IDEA. WE CAN'T NECESSARILY DO THAT BUT CAN LISTEN IN ON LINE. A LOT OF THAT CONVERSATION HAPPENS ONLINE NOW. THIS WOULD BE REALLY COOL. THERE'S A LOT OF RESEARCH ON ALT-METRICS. THIS IS A COOL VERSION, SOCIAL BOOK MARKING, DELICIOUS, REMEMBER THAT? BIG BACK IN THE DAY. WE COULD LOOK AT A CITE-YOU-LIKE AND SAY WHAT'S THE TREND? IT'S BIG ON MAYBE THE LIGHT WEIGHT THINGS AT FIRST, AND EVENTUALLY BIG ON CITATION, MAYBE TRIGGERING OTHER ACTIVITY. WHAT IF WE HAD A PERFECT SPECK-O-GRAPH. IT'S INTERESTING TO THE GENERAL PUBLIC, BUT REALLY INTERESTING TO PEOPLE IN THIS FIELD, AND A LOT OF PEOPLE ARE TALKING ABOUT DOING EXPERIMENTS HERE, BECAUSE THEY THINK THE METHOD IS VALUABLE BUT THE RESULTS ARE RUBBISH. WHAT IF WE COULD GET THAT PERFECT SPECTOGRAPH? WE COULD ASK THE MACHINE WHO IS TAKING MUST BEING IT? FOR A PARTICULAR SCHOLAR WHAT IS THEIR IMPACT ON METHODS OR IMPACT ON THE SCHOLARS AROUND THEM? LEO ZOLLARD IS A HUNGARIAN PHYSICIST, HE WORKS ON THE MANHATTAN PROJECT, NEVER HAD A HIGH PRODUCTIVE OUTPUT IN TERMS OF NUMBERS OF THE PAPERS BUT YOU CAN LOOK AT THE PLACES HE WAS AT, WHEREVER HE WAS SEEMED TO HAVE A HIGHER BUMP IN OUTPUT. SOMETHING ABOUT HIM BEING. THERE NOT EVEN A PLEASANT GUY TO BE AROUND, BUT SOMETHING ABOUT HIS PRESENCE, MAYBE HE PROMPTED OTHER PEOPLE TO DO MORE WORK. WE COULD DO THIS IN ALT-METRICS, HAVE A NUANCED AND TEXTURED IDEA OF IMPACT MEDIATED THROUGH MACHINES, BECAUSE AS HUMANS THERE'S TOO MUCH DATA FOR US TO DEAL WITH. IF WE THINK ABOUT THIS BEING A REDUCED BY MACHINES, WE LOOK AT FACTS, HEY, WE CAN BUILD FACTS MACHINES, LET'S BUILD AROUND THE FACTS OF SCIENCE, THE FACTS GENERATED BY SCIENCE, BY THE PROCESS EVER DOING SCIENCE, TURNING THE SCIENTIFIC METHOD ON THE SCIENCE WE'RE DOING. THAT'S WHEN YOU GET COOL SECOND ORDER EFFECTS. ANOTHER ONE THAT I'M EXCITED ABOUT IS FAIR DATA. THIS IS A COOL INITIATIVE TO ENCOURAGE PEOPLE TO SHARE DATA IN A MEANINGFUL WAY. A LOT OF US ARE PROBABLY AWARE EVER DATA SHARING, NIH HAS BEEN A REAL PIONEER, A REAL LEADER IN DATA SHARING. AND A LOT OF FOLKS WHO ARE TRYING TO BUILD DATA INTO WORKFLOW SAID JUST PUTTING AN EXCEL SPREADSHEET OF UNLABELED COLUMNS ON YOUR PERSONAL WEB SERVER SOMEWHERE WHILE IT'S MUCH BETTER THAN HIDING YOUR DATA, IT'S ACTUALLY NOT THAT USEFUL FOR A MACHINE-MEDIATED WORK FLOW. I DON'T KNOW WHAT'S IN THE COLUMNS. I'VE GOT TO TALK TO YOUR POSTDOC WHO IS GONE BEFORE I CAN UNDERSTAND IT. YOUR WEB SERVER GOES DOWN, I CAN'T GET IT, IT'S NOTING AGGREGATED OR CONNECTED. I DON'T HAVE TIME TO GO THROUGH THESE, CRITERIA TO SAY THESE ARE THE ATTRIBUTES OF DATA THAT WOULD MAKE IT REALLY TRANSFORMATIVE. NOT JUST OPEN, NOT JUST SHARED, BUT ALSO CONNECTED, WE'VE GOT THIS OPEN AND CONNECTED THING GOING AGAIN. ORCID IS ANOTHER ONE THAT'S FANTASTIC. HOW MANY PEOPLE HEARD OF ORCID? THAT'S AWESOME, I'M HAPPY. I THINK ORCID HAS DONE A LOT OF OUTREACH. IT'S A VALUABLE -- THEY HAVE SEEN GOOD BENEFITS FROM THAT, GOOD RETURNS ON THAT, IT DOES HAVE A LOT OF AWARENESS. I THINK THE NEXT STEP IS TO MAKE SURE THAT EVERY SINGLE SCHOLARLY PRODUCT IS LINKED TO AN ORCID. IT'S POWERFUL BECAUSE WHEN I'M BUILDING A THING, I WANT TO OF COURSE CONNECT IT TO THE CREATOR OF THE THING, RIGHT? BECAUSE THAT'S WHERE A LOT OF CREDIBILITY COMES FROM. THAT'S WHERE A LOT OF MY ABILITY TO DO ASSESSMENT, REWARD THEM FOR SHARING DATA OR SOFTWARE, SOMETHING WITH ORCID ALLOWS TO US DO THAT, FITTING WELL IN THE OPEN AND CONNECTED WORLD, OPEN BECAUSE ORCID IS -- THE CODE IS OPEN, WHICH IS ADMIRABLE, AND CONNECTED BECAUSE ORCIDs, HUMBLE SETS OF NUMBERS ARE POWERFUL BECAUSE THEY ALLOW TO US CONNECT DIFFERENT PRODUCTS TO ADD DIFFERENT PRODUCTS. MACHINES NEED CONNECTIONS TO GENERATE KNOWLEDGE. THIS IS SOMETHING THAT WE BUILT AT IMPACTSTORY, DEPSY, SHORT FOR DEPENDENCY, SHORT FOR SCIENTIFIC SOFTWARE. I THINK OF THREE LEGS TO OPEN SCIENCE STOOL. WE NEED OPEN PUBLICATIONS, OPEN ACCESS. WE NEED OPEN DATA, WHICH WE TALKED ABOUT. WE NEED OPEN SOFTWARE, BECAUSE AS A SCIENTIST FOR A GROWING AMOUNT OF FIELDS YOUR WORKFLOW LOOKS LIKE I GENERATE DATA, ANALYZE WITH SOFTWARE, CREATE A PUBLICATION, ALL THE STEPS, THREE STEPS, IF THOSE ARE ALL OPEN WE CAN HAVE A FULLY REPRODUCIBLE PIPELINE. IF WE HAVE A FULLY REPRODUCIBLE PIPELINE, MACHINES CAN DO IT. WE ADD MORE PEOPLE, INSTEAD ADD MORE MACHINES. THE DOWN SIDE OF OPEN SOFTWARE IS A HARD -- A LOT OF WORK, NO ONE WANTS TO DO IT, PARTICULARLY IF THEY DON'T GET CREDIT THEY ARE NOT GOING TO DO IT. WITH DEPSY WE SAY HOW CAN WE GIVE CREDIT TO PEOPLE FOR SOFTWARE, SHOW IT BEING VALUABLE, BUILT ALL THREE INTO DEPSY, BEING DOWNLOADED IF WE CAN COMPARE TO OTHER SIMILAR SOFTWARE, IT'S A START. IS IT BEING CITED? I THINK THAT'S -- NOT EVERYBODY CITES BUT SOME PEOPLE DO. IT'S AN INDICATOR. WE'RE WORKING ON A AN ALFRED SLOAN FUNDED GRANT TO GET MORE. FINALLY IS IT BEING REUSED BY OTHER PROJECTS. CAN YOU IMPORT SOFTWARE SOMEONE HAS DONE, WE WERE ABLE TO CRAWL THROUGH GITHUB AND SEE WHERE PEOPLE ARE IMPORTING SOFTWARE. THAT'S PRETTY COOL. THIS IS WORK IS SUPER COOL, THE IDEA OF PUBLISHING INSTEAD OF A PAPER, PUBLISHING ASSERTIONS. AN EXAMPLE, ASSERTION OF THIS PARTICULAR -- IS THAT A CHEMICAL OR GENE? DOES SOMEBODY KNOW? THIS PARTICULAR SYMBOL. I DON'T KNOW ENOUGH ABOUT SCIENCE TO KNOW WHAT IT IS, HAS A PARTICULAR FREQUENCY AND PARTICULAR POPULATION IN SARDINIA, A PROVENANCE ATTACHED TO IT, WHO FOUND THIS. INSTEAD OF PUBLISHING A PAPER WHAT IF I PUBLISHED 50 THIS TO OF THESE TRIPLES, PUBLISH 50,000 CLAIMS, MADE THIS OBSERVATION, FOUND THIS THING. UNDER THESE CONDITIONS, HERE'S WHERE IT CAME FROM. IF I PUBLISH 50,000 OF THOSE, EVERY ONE OF US PUBLISH 50,000, OR 5 MILLION, IT'S SOMETHING THE MACHINE LOVES. THAT'S MARSHMALLOW CANDY TO A MACHINE, SO EASY TO BUILD A GRAPH OUT OF THOSE AND SAY THESE TRIPLES SAY THIS, THESE TRIPLES SAY THIS, NOW I CAN MAKE INFERENCE. SAME THING YOU AND ME AS A HUMAN DO BUT WE DO IT MORE ABSTRACTLY. THE MACHINE NEEDS HELP TO MAKE THAT HAPPEN. WE NEED TO PUBLISH THE TRIPLES OR WE CAN TAKE AN EXISTING PAPER, THIS IS REALLY IMPORTANT, WE CAN TAKE SOME THAT ALREADY EXIST, THAT WORKS, START THE TRIPLES FROM THE PAPER. BOTH ARE POWERFUL. OR WE WANT TO BUILD EVENTUALLY, WHAT WE CAN BUILD CONCEPTUALLY IS WHAT'S HIS NAME, KUKULA, THE GUY WHO DISCOVERED THE BENZENE RIG. >> KEKULE. >> THANK YOU, EXCELLENT. >> (INAUDIBLE). >> HE DIDN'T DISCOVER IT. HE READ IT SOMEWHERE. >> (INAUDIBLE). >> EXCELLENT. WELL, I THINK THAT'S AN EXCELLENT DEMONSTRATION OF WHAT I'M EXCITED ABOUT, THE MACHINES WILL BE ABLE TO DO. KEKULA FAMOUSLY GETS CREDIT FOR DISCOVERING BENZENE RING, HE SAID HAD HE A DAY DREAM OF A SNAKE EATING ITS TAIL AND SAID, WOW, LIKE THAT MUST BE THE STRUCTURE OF BENZENE, THAT'S REALLY COOL. AND HE DIDN'T HAVE THAT IDEA IN ISOLATION. AND THE POINT THAT EXCITES ME ABOUT THAT IS NOT THAT HE HAD SOME ALL OF A SUDDEN FLASH OF BRILLIANCE BUT RATHER HE SAID THIS WAS BASED ON LOTS OF RESEARCH THAT HE HAD ALREADY BEEN DOING, IDEAS ORGANIZED IN MY HEAD BASED THAT I READ A LOT OF THESE THINGS OR OUT OF A BOOK. WE CAN PROGRAM A MACHINE TO HAVE THOSE SAME KIND OF INSIGHTS IN THEORY, AS LONG AS WE'RE ABLE TO FEED DATA IN, THE MACHINE CAN CONNECT THESE TWO DISPARATE DATA POINTS, AND CREATE AN EXPLANATION FOR THAT, SOMETHING HUMANS, TRADITIONALLY ONLY PEOPLE COULD DO. ONE OF THE EXAMPLES OF THAT IS AUTOMATED SCIENCE. HERE'S A LOT OF SCIENCE PROJECTS BEING DONE NOW TO GET A ROBOT TO DO ALL OF THE LAB WORK THAT NORMALLY A PERSON WOULD DO. WHAT'S COOL, CAN YOU FEED DATA INTO THE MACHINE. THE MACHINE WITH FORM A THOUSAND HIGH HYPOTHESES A DAY WITH A ROBOT IN A LAB. IT'S THE MODEL PART, NOT FAST THE FACT PART, MACHINE CAN TEST ITS OWN HYPOTHESES. WHAT PERCENT IN THE BIOMEDICAL LITERATURE COULD HAVE BEEN DONE BY ROBOT LABS, 85 TO 90%, A HUGE AMOUNT COULD BE DONE IN AN AUTOMATED WAY, ENCOURAGING FOR OPEN SCIENTISTS. FINALLY I WANT TO TALK ABOUT OPEN ACCESS, IN MANY WAYS THE POSTER CHILD OF OPEN SCIENCE BECAUSE IT'S HAD HIGH PROFILE, AND ONE OF THE MOST IMPORTANT COMPONENTS ESPECIALLY IN THE EARLY DAYS, THE MAJORITY -- BECAUSE THE MAJORITY OF SCIENCE LIVES IN THE LITERATURE, WE NEED TO GET THAT LITERATURE OPEN SO MACHINES CAN EXTRACT FACTS FROM THEM AND SO PEOPLE CAN READ THEM. SO THIS IS A STUDY THAT WE DID WITH A NUMBER OF CO-AUTHORS FROM SEVERAL UNIVERSITIES, USING UNPAYWALL DATABASE THAT WE BUILT TO BRING OPEN ACCESS INTO ONE PLACE. SOME IS OPEN, NOT CONNECTED, WE'RE BRINGING IT AND CONNECTING WITH VALUABLE METADATA. WE'VE GOT 20 MILLION OPEN ACCESS ARTICLES, 95 MILLION ALTOGETHER, IT'S IN 1500 LIBRARIES. WE LOOKED AT THAT DATABASE. HEY, IS OPEN ACCESS GETTING MORE OR LESS POPULAR, MORE OR LESS SUCCESSFUL. WE'RE HAPPY TO SEE, I DON'T HAVE TIME, WE WANT TO SEE THE LINE GO UP. I THOUGHT, I TURNED THE LINEUP SIDE DOWN, SO THE RED LINE, 1990 TO 2015 IS OPEN ACCESS GRAPH, BUT FLIPPED UPSIDE DOWN. THE RED LINE IS OPEN ACCESS. THE BLUE LINE IS NUMBER OF HORSES AND MULES PER CAPITA IN THE UNITED STATES OVER THOSE DATES. I WAS HAPPY TO SEE, HEY, MAYBE TOTAL ACCESS IS FOLLOWING THE SAME TRAJECTORY AS HORSES AND MULES IN THE UNITED STATES, BECOMING LESS IMPORTANT, ECONOMICALLY THAT'S WHAT WE WANT TO SAY. UNPAYWALL.ORG, IT'S FREE, WE'RE A NON-PROFIT. WE'RE TRYING TO GET THAT INFORMATION OUT THERE. SO I THINK I'M GOING TO END ON THAT BECAUSE WE'RE RUNNING SHORT ON TIME. I WOULD LOVE TO GET SOME QUESTIONS. WE THANK THE FOLKS WHO AT ONE TIME OR ANOTHER PROVIDED FUNDING FOR WHAT WE'RE DOING. YEAH, HEAR WHAT YOU GUYS HAVE TO SAY. [APPLAUSE] >> WE HAVE TIME FOR QUESTIONS. YEAH, PLEASE STEP UP TO THE MIC. GO AHEAD. >> VERY INTERESTING TALK. >> THANKS. >> SORT OF THINKING LIKE ALTERNATIVE DOOMSDAY APPROACH, LIKE WITH THE RISE OF SOCIAL NETWORKS WE SAW SIMILAR MODEL WHEREAS MORE PEOPLE COMMUNICATE WITH MORE PEOPLE IT BRINGS OUT VOICES, CONNECTIONS THAT WOULDN'T EXIST BEFORE. BUT THEN KIND OF OVER TIME, THINGS GET CONSOLIDATED, AND, YOU KNOW, ONCE THE ESTABLISHED ORGANIZATIONS KIND OF ARE PART OF THAT SYSTEM, YOU DON'T GET THAT SAME BENEFIT. SO LIKE YOU DO A PAGE RANK SEARCH ON A RANDOM TOPIC, WIKIPEDIA IS THE FIRST ENTRY. IT'S NOT QUITE AS DEMOCRATIZING AS IT WOULD BE. RIGHT NOW IN SCIENTIFIC PUBLICATION THERE'S A BUNCH OF YOUNG RESEARCHERS WHO ARE, YOU KNOW, PUNCHING ABOVE, THEY KNOW HOW TO ACCESS THESE SYSTEMS WHEN ESTABLISHED PEOPLE DON'T. BUT, YOU KNOW, WE GO 20 YEARS IN THE FUTURE, AND THEN THE PEOPLE WHO ARE THE HUBS ARE GOING TO ESSENTIALLY BE CONSOLIDATED EVEN MORE, AND ALL THESE ALGORITHMS ARE GOING TO BE, YOU KNOW, IF SOMEONE WHO HAS, YOU KNOW, LIKE A HUB IN A SCIENTIFIC NETWORK LINKS TO AN ARTICLE OR SOMETHING THAT'S GOING TO SHOOT TO THE TOP NO MATTER WHAT AND REPLICATE SORT OF THE OLD SCHOOL NETWORKS WE HAVE BEFORE. HOW DO WE DESIGN A SYSTEM THAT'S GOING TO ACTUALLY FACILITATE, YOU KNOW, WIDER COMMUNICATION AS OPPOSED TO CHANGE THE WAY THAT, YOU KNOW, THE PEOPLE WHO ARE ALREADY CONNECTED GET MORE CONNECTIONS? >> YEAH, GREAT QUESTION. I THINK IT'S A TWO-PHASE THING, RIGHT? RIGHT NOW PEOPLE WHO ARE NOT SO CONNECTED ARE ABLE TO EXPLOIT SOME OF THE ADDITIONAL CONNECTIVITY OF THE WEB FOR INSTANCE AND BENEFIT BUT THAT WON'T ALWAYS BE THE CASE. THAT'S AN INHERENT PROPERTY OF NETWORKS. SO IT'S CALLED PREFERENTIAL ATTACHMENT IN NETWORK SCIENCE, THE MATTHEW EFFECT IN BIBLIOMETRICS, THE RICH GET RICHER, POOR GET POORER. OVERALL OPEN SCIENCE IS STILL SCIENCE. I DON'T THINK THAT IT'S MEANT TO BE SOMETHING THAT WILL REPLACE THE CORE NORMS AND CORE ENGINES OF HOW SCIENCE IS DONE. I THINK IT REMAINS A COLLABORATIVE ENTERPRISE OF AS MANY PEOPLE AS WE CAN GET INVOLVED. MODELING AND TRYING TO UNDERSTAND AS MANY FACTS AS WE CAN FIND. I THINK SOME OF THESE NETWORK EFFECTS ARE STILL GOING TO HAPPEN. I DON'T THINK THAT'S NECESSARILY THE GOAL OF OPEN SCIENCE TO STOP THAT. I THINK TO THE EXTENT THAT WE POLITICALLY WANT TO STOP THAT, WHICH I THINK WE DEFINITELY DO, I THINK THAT THERE ARE -- WE NEED TO LOOK FOR MORE POLITICAL SOLUTIONS FOR THAT. WHEN THE POLITICAL WILL IS THERE, I MEAN IN THE BROAD SENSE, RIGHT, A BUNCH OF PEOPLE WORKING TOGETHER, RIGHT? WHERE WE SOCIALLY OR POLITICALLY WANT TO CHANGE THAT, I THINK WE CAN CERTAINLY BUILD THE ALGORITHMS TO SERVE AS DIFFERENT CONTENT IN DIFFERENT WAYS. PEOPLE TALK ABOUT FILTER BUBBLE, MAY I WOULD ONLY SEE THINGS I ALREADY LIKE, IF I GET A STREAM OF STUFF INTO MY MAGIC, OPEN SCIENCE WORK STATION, IT'S ALL BEEN FILTERED BY ALT-METRICS, WHAT IF IT ONLY SHOWS ME WHAT I LIKE? WE SEE THAT IN PEOPLE'S NEWS FEEDS OR SOMETHING. WE CAN BUILD A SERENDIPITY BUTTON, HEY, SHOW ME STUFF FROM THE OTHER SIDE OF THE ARGUMENT. THE CHALLENGE IS GETTING PEOPLE TO TURN THAT BUTTON, THAT I THINK IS THE SOCIAL CHALLENGE THAT I'M NOT SURE IS INSIDE THE BORDER OF OPEN SCIENCE, A GREAT QUESTION, SOMETHING WE WANT TO THINK ABOUT AS WE MOVE INTO THE SECOND SCIENTIFIC REVOLUTION ARE THERE WAYS THAT WE CAN STRUCTURE THE TOOLS THAT PEOPLE USE TO IMPROVE DIVERSITY AND VALUE DIVERSITY. >> I ENJOYED THE ANALOGY ABOUT HILL CLIMBING. WE DON'T KNOW WHAT THE FUTURE IS GOING TO LOOK LIKE BUT KNOW WHICH WAY UP IS, ONE THING WE WANT TO OPTIMIZE IS OPENNESS. FOR PEOPLE MAKING MAKING FUNDING DECISIONS TODAY, REWARDING INVESTIGATORS WHO ARE BEING MORE OPEN WHAT METRICS ARE AVAILABLE TO THEM NOW THAT THEY CAN MAKE THOSE DECISIONS? >> YEAH, GREAT QUESTION. IT'S SOMETHING THAT WE ARE HOPING TO BUILD IN THE NEXT PROBABLY SIX MONTHS OR SOMETHING, WITH UNPAYWALL BECAUSE WE HAVE A GOOD DATABASE, PROBABLY AUTHORITATIVE DATABASE OF WHAT EXISTS THAT OPEN, WE'LL BUILD TOOLS YOU CAN USE AT THE NIH OR ANYWHERE ELSE TO TYPE IN PARTICULARLY ORCID, WE CAN FIND ALL OF THEIR PUBLICATIONS, SAY THIS IS OPEN, FOR INSTANCE THE KIND OF THING YOU COULD USE TO MAKE A DECISION, THERE'S INTERESTING, I DON'T KNOW IF YOU HEARD ABOUT CHORUS, I THINK NIH HAS SOME COMPLIANCE DASHBOARD TYPE OF THINGS, THAT I THINK MAYBE I'M PROBABLY LIKE -- I THINK SCIENCE (INDISCERNIBLE) HAS STUFF, BASIC SIGNALS THAT SOMEONE VALUES OPENNESS, ARE THEY WRITING IN GRANT APPLICATIONS, YOU KNOW, GIVING A POSITION OF PRIVILEGE, HEY, ALL THE CODE WILL BE OPEN SOURCE, RIGHT? HERE IS THE REPOSITORY WHERE THE CODE IS LOCATED. ALL OF THE DATA WILL BE HERE. HERE IS MY DATA SHARING PLAN, MY DATA MANAGEMENT PLANNING, IS THAT PERFUNCTORY OR HAVE I PUT TIME INTO IT? ON A MANUAL LEVEL WHEN YOU LOOK AT APPLICATION, FOR ME WHEN I LOOK AT, YOU KNOW, AS A REVIEWER, IT DEFINITELY JUMPS OUT TO ME HOW MUCH A PERSON HAS PUT EFFORT INTO, HEY, YOU KNOW LIKE INTEGRATING WHAT I'M CREATING, THE PRODUCTS I'M CREATING, AT EVERY LEVEL, THE DATA, SOFTWARE, THE CODE, THE PAPERS. THAT'S A PRIORITY TO ME. SO THAT'S ONE OF THE THINGS WHEN I PEER REVIEW STUFF THAT'S WHAT I LOOK FOR. MAYBE THAT'S A PARTIAL ANSWER. IT'S A GREAT ANSWER., A GREAT QUESTION. >> CHORUS. >> TO MONITOR PUBLIC ACCESS POLICIES. >> (INAUDIBLE). >> YEAH, THANKS, HEATHER. HEATHER WAS SAYING NIH IS NOT ASSOCIATED WITH CHORUS, A FEW AGENCIES ARE. THEY HAVE NOT CONTRACTED BUT THEY HAVE SORT OF ALIGNED THEMSELVES WITH USING THIS APPROACH. IT'S WORTH LOOKING INTO. >> YEAH, GREAT TALK. >> THANK YOU. >> GREAT IDEAS AS FAR AS -- YEAH, TRYING TO -- I'M ALL FOR OPENNESS AND CONNECTEDNESS, COMPUTERS AND ALGORITHMS FOR FINDING MODELS AND THINGS LIKE THAT. ONE PROBLEM THAT I POTENTIALLY SEE, IT SEEMS LIKE FOR INSTANCE IN THE FIELD OF NEUROSCIENCE, THERE'S THE WHOLE BRAIN INITIATIVE, TRYING TO CATALYZE RESEARCH, TRYING TO CATALYZE ALSO INTERDISCIPLINARY COMMUNICATION, BUT IT SEEMS THAT YOU NEED ONE MORE THING TOO. IT SEEMS THAT YOU NEED A WAY TO FOCUS THE RESEARCH. IT SEEMS LIKE A LOT OF PEOPLE ARE DOING THEIR THING, TRYING TO ADVANCE HYPOTHESIS, TESTING MODELS, BUT IT WOULD BE NICE TO HAVE A MECHANISM IN BETWEEN THAT SORT OF, YOU KNOW, SOMEHOW GENERATES THE BIG QUESTIONS, SORT OF -- YOU KNOW IT ALL COMES DOWN TO HUMAN BRAINS TRYING TO PROCESS THIS, AND SORT OF A BIG QUESTION LIKE, OKAY, WHAT ARE THE MODELS BEING TESTED? AND WHAT DO WE KNOW AND NOT KNOW ABOUT THEM, LIKE AN INTERMEDIARY. NOT LIKE AT THE VERY LEVEL DOING YOUR OWN SCIENCE, NOT AT THE VERY HIGH LEVEL WHERE EVERYONE SOMEHOW KNOWLEDGE IS BEING GENERATED BUT SOMEWHERE IN BETWEEN WHERE IT'S BEING FOCUSED, AND IT WOULD BE INTERESTING TO EXPLORE MECHANISMS FOR LIKE BRINGING DISPARATE RESEARCH TOGETHER TO FOCUS IT MORE. SORT OF, YOU KNOW, CAUSE PEOPLE TO DECIDE WHAT RESEARCH THEY WANT TO DO BASED ON THAT. >> I COULDN'T AGREE WITH YOU MORE. I THINK THAT'S -- I THINK IN THE EARLY DAYS OF OPEN SCIENCE, GENERATIVE SCIENCE, THAT'S WHERE A LOT OF FOCUS NEEDS TO BE. THE SAME WAY EARLY DAYS OF COMPUTER SCIENCE, BIG MAINFRAMES, BIG COMPANIES, IT TOOK A LONG TIME TO GET THE COMPUTER IN MY POCKET, RIGHT? IT'S GOING TO TAKE A LONG TIME TO GET THESE TOOLS. TO RESEARCHERS INDIVIDUALLY THEY ARE GOING TO START COMPLICATED, EXPENSIVE AND EXPERTS ONLY. AND I THINK THAT PART OF WHY I'M REALLY EXCITED TO BE HERE, GIVING THIS TALK HERE, WE'RE TALKING TO FOLKS AT THE OTHER AGENCIES AS WELL, IS THAT'S A TERRIFIC FIRST-USE CASE FOR A LOT OF STUFF WE CAN DO WITH MACHINE-AIDED UNDERSTANDING OF SCIENCE, AT THE AGENCY AND FUNDING LEVEL. WHAT I WOULD LOVE IS A PROGRAM DIRECTOR TO BE ABLE TO LOOK AT SOME KIND OF A DASHBOARD WITH ALL ALT-METRICS FROM EVERY PERSON, ALL THE FACTS DISTILED OUT OF ALL THE PAPERS. HEY, LOOK, THESE ARE EMERGING RESEARCH FUNDS. MACHINES ARE SUPER GOOD AT PREDICTING, THIS IS A PREDICTION NOT OF WHO IS HAVING SUCCESS NOW OR NOT WHO HAD SUCCESS TEN YEARS AGO, WHAT WE LOOK AT WHEN WE EVALUATE GRANTS, IT TAKES THAT LONG TO LEARN WHETHER IT WAS WORKING OR NOT, BUT WHO IS GOING TO BE SUCCESSFUL WITH THIS LINE OF QUESTIONING IN FIVE OR TEN YEARS. THERE'S ALREADY BEEN TREMENDOUS AMOUNT OF WORK DONE ON DETECTING RESEARCH FRONTS IN THE LITERATURE. I WOULD ARGUE BECAUSE DATA ISN'T THERE AT A NUANCED ENOUGH LEVEL, AGAIN THE MACHINES NEED A LOT OF DATA FOR THEM TO GENERATE ANYTHING VALUABLE. SO THE POINT OF OPEN DATA AGAIN IS TO GET THAT, ALL OF THAT DATA OR OPEN SCIENCE, GET THAT DATA OUT THERE, MACHINES WILL BE MORE SUCCESSFUL. I'LL REALLY EXCITED IN THE NEXT FIVE YEARS TO SEE GOOD SYSTEMS OR AGENCY OR PROGRAM LEVEL TO DETECT WHERE SHOULD WE BE SPENDING OUR MONEY, WHERE IS THE BIG FIND, NEXT FIND GOING TO HAPPEN. IT'S A GREAT QUESTION. >> SO REALLY INSPIRING TALK. >> THANK YOU. >> TWO QUESTIONS. FIRST, I'M SURE YOU'VE HEARD IT SAID ABOUT MODELS THAT EVERYONE HAS ONE, IT'S LIKE A TOOTHBRUSH, NO ONE WANTS TO USE SOMEONE ELSE'S MODEL. SEEMS LIKE PART OF THE REASON IS THAT MODEL CREATORS ARE INVESTED IN THEIR MODEL. AND SCIENCE IS SOCIAL ACTIVITY. SO IT'S HARD FOR PEOPLE TO BE TORN APART FROM THAT. ONE QUESTION I HAVE FOR YOU, THESE EFFORTS YOU'VE TAKEN, HOW ARE THEY GOING TO -- THAT SORT OF ROAD BLOCK, THAT OPENNESS. >> GREAT QUESTION. TO ME IT COMES DOWN TO SOMETHING THAT'S BEEN UNDEREMPHASIZED IN OPEN SCIENCE, I TALKED A LITTLE IN THE TALK ABOUT MODELING ARGUE MENTATION NETWORKS. YOU SAID SCIENCE IS A SOCIAL ENTERPRISE. I HIT THE POSITIVIST THINKING HARD IN THE PRESENTATION, A BIG BELIEVER IN TOUR AND FTS-LIKE WAYS OF UNDERSTANDING SCIENCE, A HUGE PART IS RHETORIC. ARGUING TO ANOTHER TRYING TO CONVINCE ONE ANOTHER, HOPEFULLY WITH FACTS, SOMETIMES WITH PERSONALITY, WHO IS SHOUTING LOUDER, HOPEFULLY WITH FACTS, THAT'S PART OF COST OF BRINGING EXTRA BRAINS TO THE PLAY. THE INTERFACE BETWEEN YOUR BRAIN AND MY BRAIN IS LOW BANDWIDTH THAT INVOLVES SOMETIMES SHOUTING UNTIL WE CAN AGREE. AND I DON'T THINK THAT THE FUTURE OF SCIENCE, FUTURE OPEN SCIENCE WORLD, IS GOING TO -- AGAIN, I DON'T WANT TO TRANSFORM THE RULES BY WHICH WE DO SCIENCE. THEY WORK QUITE WELL. I WANT TO MAKE THEM MASSIVELY MORE EFFICIENT. SO I THINK IF WE CAN -- HOPEFULLY THIS IS A LONG WAY TO ANSWER YOUR QUESTION, I THINK IF WE CAN MODEL THE ARGUMENT MORE EFFICIENTLY, I THINK WE'RE GOING TO GET TO A CONSENSUS MORE EFFICIENTLY AS WELL. THAT'S BEEN MY EXPERIENCE WHEN I TALK, YELLING OVER A RUSTY CONNECTION, IF I'M IN PERSON OR SHARING A MEAL THAT'S ANOTHER THING. THE LATTER VERSION IS WHAT WE'RE TRYING TO WORK TOWARDS, MODEL ARGUMENTS, BROWSE THEM, INSTEAD OF WASTING TIME ON SHOUTING AND CONFUSION MAYBE WE CAN MORE QUICKLY GET TO THE HEART. THESE ARE WHY THESE CAMPS AGREE OR DISAGREE, THIS IS ALL OF THE HISTORY THAT'S GONE INTO THIS THIS CAMP. IS ALL THE FACTS IN THE MACHINE CAN SAY, FUNDAMENTALLY EVEN THOUGH THERE'S A THOUSAND OR A MILLION FACTS, IF WE WERE ABLE TO SOLVE THIS ONE LITTLE PART, THAT WOULD PROBABLY RECONCILE THE TWO MODELS AND WE COULD FOCUS ON THAT. AGAIN, IN MY MIND, IT'S A WAY OF QUALITATIVELY IMPROVING THE EFFICIENCY OF THE ARGUMENT, NOT ELIMINATING. THE ARGUMENT IS WHERE THE KNOWLEDGE HAPPENS, RIGHT? GREAT QUESTION. >> MY SECOND QUESTION IS ABOUT FACTS, AND AS LONG AS WE'RE DOING HYPOTHESIS TESTING HOW DOES THAT RELATE WITH THE FACTS THAT YOU'RE TALKING ABOUT? WHEN YOU SAY FACTS, LAWS OF GRAVITY, THE HARD LAWS. >> OKAY, OKAY. >> AS OPPOSED TO JUST SAYING SOMETHING IS BETTER THAN THIS, .05. EVEN IF YOU HAVE BIBLIOMETRICS, THOSE ARE HYPOTHESIS TESTED UNDER CONSTRAINED CIRCUMSTANCES AND SO ON, RIGHT? SO HOW DOES THAT REALLY HELP, RIGHT? >> IT'S A GOOD QUESTION, I APOLOGIZE FOR LACK OF CLARITY. I SAY FACTS, INFORMATION SCIENCE, LITERATURE, A LOT OF TIMES WE LOOK AT THINGS IN HIERARCHY, WE'VE GOT FACTS, INFORMATION, KNOWLEDGE. ENACTS FACTS IS CLAIMS. THE CHAIR IS BLUE, THE CHAIR IS RED. I'VE GIVEN A PROPOSITION. WE HAVE A BUNCH OF INFORMATION. I COULD GIVE A FACT BECOME EVERY CHAIR OR PERSON IN THE ROOM. I'M NOT SURE WHAT I WOULD DO WITH IT. EVERYBODY IN THE ROOM IS OVER THE AGE OF 30, THAT'S INFORMATION. THEN I COULD HAVE A HYPOTHESIS OR BUILD KNOWLEDGE. OH, PEOPLE WHO ARE OVER THE AGE OF 30 ARE MORE LIKELY TO GO TO TALK AT NIH, MAYBE THAT'S BECAUSE IT TAKES CAREER PREPARATION TO BE AT THE NIH. I'M MAKING THIS UP AND IT'S STUPID. BUT MY POINT IS, YOU KNOW, WE START WITH THE SMALLEST GRAINS OF OBSERVATIONS AND GRADUALLY WORK OUR WAY UP TO MORE AND MORE ABSTRACTION, WHERE THE KNOWLEDGE IS. EVENTUALLY WE'RE GETTING INCREDIBLE INK THAT EXPLAINS ENORMOUS PARTS OF THE SOLAR SYSTEM, WHAT HAPPENS WHEN I DROP A ROCK, WHEN AN APPLE HITS MY HEAD, WE'VE GOT TO START WITH GRANULAR FACTS. BRAINS HAD THE ONLY THING WE HAVE THAT CAN DO THAT. WE WANT MACHINES TO DO THAT, THEY ARE MORE EFFECTIVE. >> IT WAS JOSEPH LUESH WROTE THE BOOK, NOT CLEAR IF KEKULE READ IT. >> OKAY. >> INTERESTING TALK. I THINK MORE EMPHASIS IN THE FUTURE ON THE FOLLOWING. GRESHAM'S LAW, TRENDS, FADS, NOMENCLATURE, MISTAKES, FRAUD, FAILURE, PREDATORY PUBLISHING, HOW TO AVOID THESE KINDS OF DELETERIOUS EFFECTS IN THE NEXT GENERATION, NEXT SCIENTIFIC REVOLUTION. >> GREAT, THANKS. POINT WELL TAKEN. HOPEFULLY AFTER THE TALK WE CAN DELVE MORE INTO THAT. >> ALL RIGHT, LET'S THANK JASON AGAIN.