It's 2007, good time to resume our weekly seminar! There are some exciting speakers lined up for you this year, stay tuned. Before this weeks announcement, I would like to point out that we are currently actively hiring interns (algorithms, systems, design and more) for this term as well as the summer. Drop me a line if interested, or see more on our blog at http://whyrb.com .

This week in Yahoo Research Berkeley Brain Jam:

Stephen W Smoliar
A Socio-Technical Approach to Content Mining

Friday, Jan 5th, 3pm
Yahoo! Research Berkeley
1950 University Ave., Berkeley.
http://upcoming.org/event/133460/

Abstract:
Content mining technology is based on the ability to identify andrecognize PATTERNS in large information repositories. Whether itinvolves data, metadata, or a combination of the two, the concept of"pattern" is a mathematical one; and it is because of its mathematicalnature that it can be readily manipulated by powerful software.Nevertheless, this is a dangerously limited interpretation of theconcept of "pattern," which can only impede efforts to engage softwareto facilitate the management of those repositories, whether theyconstitute a digital library or someone's favorite collection ofphotographs. If we wish to develop more powerful software, we mustbegin by acknowledging that patterns are not strictly OBJECTIVE (andthus susceptible to mathematical modeling) but also SUBJECTIVE andSOCIAL. As SUBJECTIVE phenomena, patterns are FLUID, their treatment oftime is qualitatively different from that of space, and they deal notjust with ARTIFACTS but also with the ACTS OF MAKING those artifacts.As SOCIAL phenomena, patterns are CONSTRUCTED THROUGH INTERPERSONALCOMMUNICATIVE ACTS. Consequently, effective content mining demandsanalytic techniques that embrace the subjective and social worlds aswell as the objective world of mathematics. While this constitutes anintimidating challenge to software development, it will be demonstratedthat socio-technical systems in areas such as knowledge management canprovide guidelines for taking a similar approach to content mining.

Bio:
Stephen William Smoliar obtained his PhD in Applied Mathematics and hisBSc in Mathematics from MIT. He has taught Computer Science at both theTechnion, in Israel, and the University of Pennsylvania. He has workedon problems involving specification of distributed systems at GeneralResearch Corporation and has investigated expert systems development atboth Schlumberger and the Information Sciences Institute (University ofSouthern California). Dr. Smoliar also has extensive background inmusic, having composed 36 works between 1969 and 1975, and is a formermember of the Society for Music Theory. His main areas of researchinterest are in knowledge representation, perceptual categorization,and cognitive models. From May of 1991 until August of 1994, he led aproject on video classification at the Institute of Systems Science atthe National University of Singapore. From August of 1995 until May of1999, he managed research programs in Multimedia and Communication andCollaboration at the FX Palo Alto Research Laboratory, after which heserved as Knowledge Solutions Coordinator until May of 2004. During2005 he was a member of the InfoBiz team at PARC with the mission ofdeveloping strategies for private-sector support of sensemakingresearch and technologies. He is currently an independent consultantand blogger, working on a book that addresses the conflict betweenscientific dogma and the human conditions.

Upcoming Brain Jams:
12 Jan 2007 – Jon Foote – Analyzing Music and Audio by Self-Similarity
19 Jan 2007 – Scott Golder - TBD

--
To get regular (not more often than once a week) Brain Jamannouncements, please join the yrb-bj mailing list by sending email toyrb-bj-subscribe@yahoogroups.com or by visiting the group page athttp://groups.yahoo.com/group/yrb-bj/

Yahoo Research Berkeley
http://whyrb.com