Knowledge retrieval

Abstract: The present invention is directed to an integrated implementation framework and resulting medium for knowledge retrieval, management, capture, sharing, discovery, delivery and presentation. The system is responsible for maintaining semantic information.

Abstract: We present a new kind of question answering dataset, OpenBookQA, modeled after open book exams for assessing human understanding of a subject. The open book that comes with our questions is a set of 1326 elementary level science facts. Roughly 6000 questions probe an understanding of these facts and their application to novel situations. This requires combining an open book fact (e.g., metals conduct electricity) with broad common knowledge (e.g., a suit of armor is made of metal) obtained from other sources. While existing QA datasets over documents or knowledge bases, being generally self-contained, focus on linguistic understanding, OpenBookQA probes a deeper understanding of both the topic—in the context of common knowledge—and the language it is expressed in. Human performance on OpenBookQA is close to 92%, but many state-of-the-art pre-trained QA methods perform surprisingly poorly, worse than several simple neural baselines we develop. Our oracle experiments designed to circumvent the knowledge retrieval bottleneck demonstrate the value of both the open book and additional facts. We leave it as a challenge to solve the retrieval problem in this multi-hop setting and to close the large gap to human performance.

Abstract: The term “mapping knowledge domains” was chosen to describe a newly evolving interdisciplinary area of science aimed at the process of charting, mining, analyzing, sorting, enabling navigation of, and displaying knowledge. This field is aimed at easing information access, making evident the structure of knowledge, and allowing seekers of knowledge to succeed in their endeavors. Although thousands of years old, this area has undergone a sea change in the last 15 years, a change fostered by an explosion of the amount of information available, the accessibility of that information due to electronic storage, and the new techniques of analysis, retrieval, and visualization that are made possible by vast increases in computational storage capacity and processing speed and power. Many of us are so involved in the new ways of accessing knowledge that we have forgotten how recent is the change to computerized knowledge retrieval with search engines operating on the World Wide Web. Remarkable as these changes are to date, they are only a hint of the transformation to come. The Arthur M. Sackler Colloquium on Mapping Knowledge Domains, held at the Arnold and Mabel Beckman Center of the National Academies of Sciences and Engineering in Irvine, CA, May 9-11, 2003, was designed to showcase the ongoing developments in this transformation and provide pointers toward the directions it will move. The changes that are taking place profoundly affect the way we access and use information. Scientists, academics, and librarians have historically worked hard to codify, classify, and organize knowledge, thereby making it useful and accessible. The day is fast approaching when all this knowledge will be coded electronically, but mixed in a vast and largely disorganized and often unreliable sea of mostly recent information. Fishing this sea for desired information is presently no easy task and will continue to increase …