Let Knowledge Grow

November 20th, 2013 in General News, Knowledge Lab News, Metaknowledge,

Author: Bill Harms for the Division of the Social Sciences

Read the story on Division of the Social Sciences website

Anyone who has used the Internet to hunt down an obscure piece of information knows that human knowledge is more organized and accessible than ever before in human history.

It might not be so obvious that “this same electronic availability is revolutionizing our ability to create knowledge about knowledge, or metaknowledge,” says James Evans, associate professor in Sociology and Director of the Metaknowledge Network project at the University.

Evans’ own research explores how social and technical institutions like teams, universities, markets and the Internet, shape knowledge.  Sociologists have long been interested in processes of knowledge making, and work in metaknowledge provides an opportunity to explore how networks of scientists, techniques, ideas and money channel discoveries and decision making in science.

The interdisciplinary nature of the University of Chicago makes the university the perfect setting for Evan's metaknowledge project. The University and the Computation Institute, in which the project is housed, have a long tradition of supporting interdisciplinary scholarship and collaboration between computer scientists and domain experts in the natural and social sciences.

New tools to make a deep dive into data

Metaknowledge researchers develop tools that allow them to harvest and synthesize data at unprecedented scale and from diverse sources, including scientific journals, online encyclopedias, Twitter and blog posts, news articles, books and patents.  Using this information, they can track ideas over time and follow the work of individual scientists and citizens.

The UChicago Knowledge Lab and the multi-institutional Metaknowledge Research Network that it supports is funded with a $5.2 million grant from the John Templeton Foundation.  Knowledge Lab analyzes text, images and audio files, some hundreds of years old, in the service of understanding how knowledge is constructed and dismantled over time.

“This is different from a Google search, because all of these sources are wired, to the authors, to each other, based on the institutions in which they were sponsored, and how distant they are from each other in the social network of collaboration.  We’re interested in extracting and the linking concepts and claims from these articles,” said Evans, who leads the effort.

Making an impact on science

The goal of the project is not merely to assemble information about information, but to understand both the anatomy and physiology of knowledge—the structure of knowledge and the processes that produce it, he explained. Metaknowledge research looks for trends in the data, such as which ideas were considered worth studying in the past and how that shaped the path of scientific consideration, discovery and certainty.

The promise of metaknowledge is its potential to uncover hidden biases, suggest new questions and help define new hypotheses that may have been missed in the conventional, individual process of preparing and publishing research.

“If we are successful, we will improve the quality of science,” Evans said.  Not only will new questions get asked, but scientists will be better able to track which ideas and which research proves most productive.  That kind of evaluation not only traces what has been studied and how, but models what could have been researched, will help research institutions as well as government and private funders better understand how to invest research funds, Evans said.

A national collaboration

The Metaknowledge Network includes 25 scholars from around the country, who meet as a group twice a year to develop new metaknowledge research projects that address fundamental questions, like “What makes someone a great scientist?” or “How will science change in the age of intelligent machines?” that need further examination.  The Knowledge Lab at UChicago provides support for the research of network members.

A new web portal will provide an opportunity for new scholars to join the conversation. The portal will post research updates from the network, index articles of interest from around the world, and serve as a clearinghouse for funding opportunities, including those supported by the Templeton Foundation to be dispensed by the Metaknowledge Research Network.

Evans and colleagues already have work underway that sheds new light on the scientific process.  By tracking the work of scientists who won major awards, including the Nobel Prize, Evans and network member Jacob Foster, assistant professor at the University of California, Los Angeles, are studying the common characteristics of successful scientific careers. 

“One thing we have found is that the scientists who eventually succeed most performed risky research—they introduced novel entities or concepts and pioneered new connections.  Also there was a receptive audience for that risky research, other people who were prepared to understand how important the new ideas were,” he said.

“For example, Andrew Fire and Craig Mello received the Nobel Prize in 2006 for their discovery of RNA interference. The discovery and its later applications involved the proposal of a novel process, and the silencing of a number of previously unrelated genes and so connected established components in a fundamentally new way.” By understanding the process of breakthrough science, the metaknowledge project also hopes to gain insight into funding and hosting institutions that will support it.