Where do ideas come from?  How does science solve problems?  My research suggests that it is usually  more than just some scientist in a lab somewhere with a eureka moment.  In recent years I have become fascinated with the field of informatics, and specifically, bioinformatics.  bluekeyboardAt a level of abstraction, informatics is the articulation of my organizational learning research at a much more broad level.  The key question driving my research on organizational learning is that of understanding how organizations create breakthrough new knowledge.  I see these types of breakthroughs as the driver of of most business and science, those ideas that transform industries, create new markets, and improve the human condition. Informatics aims to make the process of finding solutions / breakthroughs much more systematic… driving the questions we ask to be more precise and efficient. 

Who decides the direction of science?  On the ‘supply-side’ of course there are several answers, including individual scientists, research grants, and private funding.  Certainly the ‘demand-side’ of the equation has some influence too….  when an epidemic rages, science quickly jumps to find solutions.  Pharma companies very carefully choose their projects based on projected market sizes for potential drugs.  So, supply and demand control the direction of science.  I think its more idiosyncratic than that.  People have a wide variety of  motivations, and some of these motivations lead them to think about grad school.  Some do biology, some do engineering.  Maybe they even come  into grad school with some sort of idea of the type of work they would like to do.  Maybe.  More likely they read bios for people they might work with, decide which sound most interesting, and eventually work with one of these people.  At the end of their graduate studies they are a fresh-minted scientist, maybe a Ph.D., but what will they work on?  Something deriving out of their work in grad school.  This path dependency constrains science in existing directions.  Why are most major math breakthroughs created by people under the age of 30?  Because they are free of the rigidities of their experience.  The same goes for science.  Some break the mold, and all scientists morph over time to create their own identity, but their work, their version of science, their paths are significantly influenced by their experience. 

Informatics has the potential to change the questions people ask.  Why study the properties of one particular enzyme and not another?  People assign probabilities to certain directions, after engaging in local search around what they already know.  If informatics can provide tools for analyzing the big-picture, or make search generally more broad, or make search generally more effective by using more data more effectively to shape the questions we ask, the potential is crazy.  Everything could be solved faster.  Science would still be incremental, but the gains could be so much more systemic. Labs become exponentially more productive.  Collaborations can be driven by likelihood models instead of just social relations or geographic proximity. 

Informatics could be the solution to solutions.  Or at least the solution to asking better questions, and attacking better problems.  I think that is fascinating.

Apparently Microsoft thinks so too… see this doc on Microsoft’s vision of science in 2020 and the roadmap to get there (driving straight through informatics).