Reinventing Discovery: The New Era of Networked Science

Image of Reinventing Discovery: The New Era of Networked Science
Release Date: 
October 23, 2011
Princeton University Press
Reviewed by: 

“Reinventing Discovery is a survey, an analysis, a how-to, and a harbinger of greater things to come. Kudos to the author for picking a timely and relevant subject perhaps just on the edge of social consciousness and making a great story out of it.”

Networked science changes the nature of scientific discovery, enabling collaboration across the world, converting individual insight into collective insight, and delegating complexity to the computer. Collective insight not only increases the number of questions we can ask of science but also the number of questions we can answer.

To the question, “Is the web making us smarter or dumber?” Michael Nielsen in Reinventing Discovery is firmly in the camp of smarter. And while networked collaborative science may be an almost invisible revolution in the making, Michael Nielsen “gets it” and brings us the fascinating story.

One networked collaboration having tremendous impact on software development occurred in 1991. Linus Torvalds posed a question in an online forum asking for advice on features for a new computer operating system. He ended up creating Linux, kicking off the open source software movement.

Another collaboration pointing out to the world the value of many minds working in conjunction occurred in 1999 when Gary Kasparov played chess against the “world.” The world was comprised of 50,000 chess fans from 50 countries who collaborated through a chess website. The world studied moves, shared analyses, and voted on one move per 24 hours. The world turned out to be pretty smart—and the game was not as easy as Kasparov had imagined. At one point the world came close to winning, although in the end voted to resign after 62 moves.

In Reinventing Discovery, although author Michael Nielsen provides the reader with a survey of various networked collaborations, he principally focuses on networked collaborative science. Through blogs and websites, the Internet permits scientists to collaborate with non-scientists on in projects in math, physics, biology, dinosaurs, bird migration, oceanography, astronomy, tracking comets, and more.

One such collaborative effort in mathematics is a blog called the Polymath Project, run by a Fields Medal winner, Tim Gowers. Tim posted one particular math problem, and over a period of 37 days received 800 comments from 27 participants. Not only was the math problem solved, but the more difficult general case that it was part of was also solved.

There are other institutional websites dedicated to collaborative science. Probably the most recognized genetic website is the Human Genome Project—actually only semi-open as not all genetic sequences are available to all users.

The success of the Human Genome Project led to other collaborative efforts including HapMap. Started in 2007, HapMap is a database of the genetic code—not so much the genes of selected individuals but a map of the entire human race and the common patterns of genetic variation.

Other collaborative efforts on the genetic front include the website GenBank, sponsored by the National Institute of Health (NIH), and the Allen Brain Atlas, which maps out the brain to the level of cells and genes.

Non-scientists can contribute to collaborative science efforts, too. In 1998 Don Swanson used MEDLINE and detected a pattern between magnesium deficiency and migraines. His discovery did not lead to a cure (yet), but provided a valid inference, pointing out a need for further exploration.

FoldIt is another networked collaboration between scientists and non-scientists. FoldIt is a computer game that figures out how DNA gives rise to proteins. The game works by problem solving: giving a higher score to the lowest energy state of a folded model of a protein. To date 75,000 people have signed up. The website also provides a problem solving wiki and hosts online groups to share strategies and create game scripts. FoldIt was recently in the news for resolving the structure of a retrovirus enzyme, something that scientists had been working on for the past decade without success.

There are astronomy collaborations as well. The Sloan Digital Sky Survey (SDSS) is an open accessible database that holds data taken from robot-controlled telescopes using special wide-angle lenses. The SDSS database holds not just images but also database query tools and tutorials. It currently holds images of 230 million celestial objects (of which 930,000 are galaxies). This is only one quarter of the sky! Computer algorithms are used to search images for known types, for example dwarf galaxies and black holes (by their gravitational effect on objects nearby). So far, SDSS has had a tremendous impact on astronomy. But there is a limit on what computer analysis can do. GalaxyZoo to the rescue!

The GalaxyZoo encourages volunteers to help astronomers classify galaxy images that computer algorithms can’t. GalaxyZoo currently has 200,000 participants and can even make new discoveries, such as identifying new classes of galaxies—green galaxies were discovered by volunteers. Another amateur, a Dutch schoolteacher named Hanny van Arkel, was the first to discover a quasar mirror (Hanny’s Voorwerp). A quasar mirror is a light echo of a long gone object reflected in interstellar gas. Since her discovery, more quasar mirrors have been found.

If you are more interested in the ocean than in astronomy, there are ocean collaborations. The Ocean Observatories Initiative was started in 2009 with NSF funding to bring the Internet into ocean studies by use of sensors powered by undersea cables deep in the ocean.

There are online collaborations for solving all kinds of problems. Open source collaborations give people freedom to improve and modify others’ work, enabling groups to create projects that are more complex than what any one individual could create on his or her own. The Open Architecture Network was started in 2006 and is an online community for architecture. By 2010 the Open Architecture Network held 4,000 projects: floor plans for buildings, homes, and schools.

Successful collaborative websites restructure how expert attention is applied by helping find the right expert for the right problem at the right time. Michael Nielsen calls design that successfully matches experts to problems, “designed serendipity.” Networked collaborations can be markets as well. InnoCentive is a marketplace for science and technology problems that has 160,000 participants from 175 countries and presents 200 challenges.

An example of a challenge to InnoCentive that resulted in a successful collaboration occurred when ASSET—a nonprofit dedicated to educating Indian women in technology—needed but could not find low cost solar-powered wireless routers. They issued a challenge to InnoCentive using $20,000 in prize money put up by the Rockefeller Foundation. They found a collaborator and got their router design.

In describing collaborative efforts, Michael Nielsen provides the reader with more than a survey. He also lays down a philosophy and an instructional how to of conducting collaborative science. He points out myths and fallacies, identifies the various ways to organize a collaborative effort, and points out necessary scaling and administrative elements such as the division of tasks, component reuse, and signaling and coordination.

The author also notes that some kinds of problems—those with answers that can be tested—are better suited to scientific collaborative efforts than other types of problems. Science and math problems are better suited to collaboration than value-laden problems. In this, the author cautions that collaborative efforts may not be able to solve all the world’s problems, especially those in which science and politics intersect.

Written for the popular press as well as for scientists looking for ways to make their work more open and visible to the public, this book attempts many things and (surprisingly) does them all quite well. Reinventing Discovery is a survey, an analysis, a how-to, and a harbinger of greater things to come. Kudos to the author for picking a timely and relevant subject perhaps just on the edge of social consciousness and making a great story out of it.