Progress in Search

A Panel Discussion

Accelerating Change 2005
46 minutes, 21.1mb, recorded 2005-09-17
Ronald Kaplan, Marti Hearst and Sibley Verbeck

Do you want a "dinner party" robot who will chat with your guests making small talk and providing entertainment? Or do you just want to be able to ask a computer a real question, like "who came second in the 1986 World Cup" and get a real answer? Research into intelligent language recognition could provide solutions to both scenarios.

At this talk from Accelerating Change 2005 moderated by Sibley Verbeck, Ronald Kaplan and Marti Hearst discuss the current state of conversational interfaces as well as offer their predictions for the future. Today most interfaces rely on typewritten input, which even between humans is a less than perfect communication medium. Context and inflection are difficult to convey, and ambiguity is often a problem, especially with short statements.

In the future, however, search tools could respond to full questions, and offer real answers rather than simply links to possibly relevant pages. Search will rely on an understanding of meaning rather than an aggregate of whatever happens to be popular. We may not all have our own party-bot, but we may be able to search more quickly, robustly and intuitively.


Ronald Kaplan is a Research Fellow at the Xerox Palo Alto Research Center and leader of the linguistic research group at Xerox. He is also a Consulting Professor of Linguistics at Stanford University. As a co-creator of the theory of Lexical Functional Grammar, he was responsible for many of its formal and conceptual characteristics and has investigated its mathematical and computational properties. He received a Ph.D. in 1975 from Harvard University.

Marti Hearst is an associate professor in SIMS, the School of Information Management and Systems at UC Berkeley, with an affiliate appointment in the Computer Science Division. She has done extensive research on search user interfaces. Her primary research interests are user interfaces, visualization for information retrieval, empirical computational linguistics, and text data mining. She received BA, MS, and PhD degrees in Computer Science from the University of California at Berkeley, and she was a Member of the Research Staff at Xerox PARC from 1994 to 1997.

Sibley Verbeck (Moderator) is a leading researcher in advanced computational linguistic and statistical techniques for analyzing audio, video, and text. He has received and led multiple R&D grants and contracts from leading research organizations such as NIST, the NSF, the US Army, the US Air Force, the Missile Defense Agency, and the Lemelson Foundation to conduct research into natural language understanding techniques, machine translation, and artificial intelligence. He is responsible for continuing to expand the state-of-the-art through StreamSage's automated rich media indexing platform and related applications.

Resources:

This free podcast is from our Accelerating Change series.

For The Conversations Network: