Abstract:
Over the last few years, machine learning approaches to natural language processing have made significant strides in numerous applications, including: automatic summarization, sentiment detection, personalized news feed generation, natural language user interfaces, and more. Among these various applications, there is a unifying theme -- reliance on a broad mathematical foundation, with special emphasis on linear algebra, probability, and statistics.

In this talk, I'll introduce the fundamentals of machine learning and delve into some of the natural language problems listed above, with the goal of sketching solutions using only techniques that I introduce, motivate, demonstrate, and rigorously justify during the talk.

While the talk is targeted at a mathematically sophisticated audience familiar with standard concepts from linear algebra and probability theory, no prior exposure to machine learning or natural language processing is required or expected.

Bio:
David Montague is part of the Quantitative Execution and Development team at Palantir, where he works on a variety of mathematical modeling problems across Palantir and its customers. Some of the domains in which he has worked include health insurance, local law enforcement, media, oil & gas, and pharmaceutical research. David holds a PhD in mathematics from Stanford University, where he studied covariance estimation and graphical models for infinite collections of random variables. Speaker(s): David David Monteque (Palantir)