Information Heuristics & Finding Knowledge In Information

By Michael A. Keller, Stanford University Librarian, Publisher Stanford University Press

Words, like objects in the real world, can be helpful in carrying multiple uses. The word heuristic is a nice example of words carrying multiple uses and meanings. The literal meaning in ancient Greek is “I find” or even, “I discover.”

The modern use of heuristics describes the study of methods of innovation or discovery in problem solving. Because experience is unique to each person, problem solving can be riddled with unhelpful bias toward certain types of solutions, it can be hobbled by ignorance of truly helpful solutions, or limited by illogical choice reinforced by culture and language. Finding or inventing solutions to problems can also be difficult because there is an inadequate tool set.

One area of heuristics that is perhaps under-researched is in the area of knowledge seeking. There isn’t really a science of Information Heuristics, but perhaps there should be.

Information professionals such as librarians are in a unique position to advance a practice of Information Heuristics because they spend much of their time helping solve real problems in the world of knowledge seeking. People bring to libraries widely diverse needs and librarians respond to them with a more extensive search and discover tool set than most people have.

On the one hand libraries help people solve problems in the midst of a medical crisis, personal and business financial challenges, or research on topics of personal interest in current events — understanding what’s happening in the news. On the other hand, libraries help professionals solve problems of a deeply technical or intellectual nature; finding connections between fields of study in interdisciplinary research or locating highly relevant data in large bodies of data such as scientific or humanities research. In both cases, a rigorous model for problem solving involving a commitment to evidence, credibility, and source reliability are all essential to results that the information seeker can rely on.

Librarians can and should be creative editors and advocates for a rigorous approach to the evaluation of the results delivered by catalogs and discovery tools that aggregate and index the results of scientific and humanities research. Their training, experience, and role puts them in an ideal position to do that. However, the tool set is limited by search and discovery methods that evolved over recent decades (keyword and key phrase searching) and over centuries (document description, classification, and subject designation).

The systems that gather information should reveal highly relevant resources without the limitations of keyword searching and the blinders enforced by the limitations of subject terms and abstracts in languages the searcher does not comprehend. By focusing on context and the concepts inside the full text of data, we should be exposed to the widest variety of highly relevant results and then apply the rigors of Information Heuristics to the knowledge seeking process. The neural network approach of Yewno discover along with its freedom from the boundaries of languages of sources reveals concepts of relevance and correlations among such concepts from a corpus of over 125 million academic sources. That conceptual approach adds dramatically to the librarian’s tool set.

Jun Ge