In the second in a series of short, informative videos on privacy, Google offers a closer look at personalization and the privacy tools that are available when you choose to personalize your search. As the video explains, search algorithms that are designed to take your personal preferences into account, including the things you search for and the sites you visit, have better odds of delivering useful results for you. So if you’ve been checking out sites about the Louvre and you search for “Paris,” you’re more likely to get results about the French capital than the celebrity heiress. The privacy tools we’ve designed–such as “pause” and “remove” buttons — help put you in control of personalization.

Speakers: AnnaLee Saxenian, Dean, School of Information and Professor, City and Regional Planning, University of California, Berkeley. Steven Weber, Director, Institute of International Studies and Professor, Political Science

The mobility of skilled labor is transforming the flow of knowledge around the world. As U.S.-educated engineers and professionals return to their home countries, they are turning what once was a brain drain into a two-way process of brain circulation. These professionals are transferring to developing regions the technology and managerial know-how that once resided exclusively in advanced economies like the U.S. This process is fueling the emergence of new centers of technology entrepreneurship and creating new competitors for Silicon Valley and foreshadows persistent global skill shortages in coming decades. Discover Cal Lecture, University of California, Berkeley.

Dr. Kourosh Gharachorloo, head of Google’s Ad Traffic Quality engineering team, speaks at the AdFraud workshop at Stanford University on September 14th.

http://www.google.com/adwords/adtrafficquality/files/adfraud_anecdotes.pdf

Mark Edelman, of Playing At Learning, describes the 2007 FIRST Lego League “Power Puzzle” challenge: the challenge details, changes in this years rules and judging guidelines, and answers questions. Slides for the talk are at: http://www.ncafll.org/2007/kickoff_slides.pdf

Google Tech Talks
September, 5 2007

ABSTRACT

Statistical language processing tools are being applied to an
ever-wider and more varied range of linguistic data. Researchers and
engineers are using statistical models to organize and understand
financial news, legal documents, biomedical abstracts, and weblog
entries, among many other domains. Because language varies so widely,
collecting and curating training sets for each different domain is
prohibitively expensive. At the same time, differences in vocabulary
and writing style across domains can cause state-of-the-art supervised
models to dramatically increase in error.

This talk describes structural correspondence learning (SCL), a method
for adapting models from resource-rich source domains to resource-poor
target domains. SCL uses unlabeled data from both domains to induce a
common feature representation for domain adaptation. We demonstrate
SCL for two NLP tasks: sentiment classification and part of speech
tagging. For each of these tasks, SCL significantly reduces the error
of a state-of-the-art discriminative model.

Speaker: John Blitzer

Complete Conversion University presentation by Jeff Gillis