Colloquium Series

Matthew Andrews of Alcatel-Lucent Bell Labs

Friday, September 27, 2013 - 11:00am to 2:00pm
EE Conference Room – Mudd 1300 Suite
Columbia University
New York, NY 10027
United States

Electrical Engineering & Data Science Institute Seminar

Abstract: In recent years there has been growing interest in using Call Detailed Records (CDRs) to understand user behavior in wireless networks. Typical studies examine user mobility patterns or social interactions via mobile call graphs. In this talk we discuss ways in which we can use CDRs to understand how users respond to the pricing features of their wireless service plan. Such information allows network operators to better tailor plans to individual needs.

Never-ending Learning Machine by Dr. Tom M. Mitchell

Wednesday, May 7, 2014 - 6:00pm
Columbia University
Davis Auditorium, Room 412, Shapiro SEPSR
New York, NY 10027
United States

Dr. Tom M. Mitchell

E. Fredkin University Professor and Chair Machine Learning Department 
Carnegie Mellon University
 
Abstract:
We will never really understand learning until we can build machines that learn many different things, over years, and become better learners over time. 
 
This talk describes our research to build a Never-Ending Language Learner (NELL) that runs 24 hours per day, forever, learning to read the web. Each day NELL extracts (reads) more facts from the web, and integrates these into its growing knowledge base of beliefs. Each day NELL also learns to read better than yesterday, enabling it to go back to the text it read yesterday, and extract more facts, more accurately, today. NELL has been running 24 hours/day for four years now. The result so far is a collection of 70 million interconnected beliefs, that NELL is considering at different levels of confidence, along with millions of learned phrasings, morphological features, and web page structures that NELL uses to extract beliefs from the web. 

Challenges for Machine Learning in Computational Sustainability by Dr. Thomas G. Dietterich

Thursday, March 13, 2014 - 6:00pm to 7:00pm
Columbia University
Davis Auditorium, Room 412, Shapiro SEPSR
New York, NY 10027
United States

Research in computational sustainability seeks to develop and apply methods from computer science to the many challenges of managing the earth's ecosystems sustainably. Viewed as a control problem, ecosystem management is challenging for two reasons. First, we lack good models of the function and structure of the earth's ecosystems. Second, it is difficult to compute optimal management policies because ecosystems exhibit complex spatio-temporal interactions at multiple scales.

This talk will discuss some of the many challenges and opportunities for machine learning research in computational sustainability. These include sensor placement, data interpretation, model fitting, computing robust optimal policies, and finally executing those policies successfully. Examples will be discussed on current work and open problems in each of these problems.

Signal Processing for Networked Data

Tuesday, November 19, 2013 - 11:00am to 12:00pm
Columbia University
Davis Auditorium
New York, NY 10027
United States

Electrical Engineering & Data Science Institute Distinguished Lecture:
Signal Processing for Networked Data

Speaker: Dr. José M. F. Moura 

Abstract: In the era of Big Data are traditional analysis tools and concepts drawn from signal processing any useful? In social contexts, or the web, or enterprises, the relations and dependencies among data are often conveniently represented by graphs, with the data becoming functions or signals on a graph–a point of view structurally di erent from the one traditionally adopted with time series. This talk extends the basic concepts of signal processing to graph signals: lters and ltering, shifting, frequency, low-, high-pass graph signals, frequency response, linear transforms, Fourier and z-transforms. We then illustrate signal processing on graphs with datasets from social networks, a service provider, and the web. Work with Dr. Aliaksei Sandryhaila and graduate student Stephen Kruzick

Dr. Judea Pearl of UCLA

Monday, November 4, 2013 - 6:00pm
Davis Auditorium, Room 412, Shapiro SEPSR
500 West 120th Street
New York, NY 10027
United States

The Mathematics of Cause and Effect

 

Abstract: Recent advances in graphical models and the logic of counterfactuals have had a marked effect on the way scientists treat problems involving cause-effect relationships. Paradoxes and controversies have been resolved, slippery concepts have been demystified, and practical problems requiring causal information, which long were regarded as either metaphysical or unmanageable can now be solved using elementary mathematics.

I will review concepts, principles, and mathematical tools that were found useful in this transformation, and will demonstrate their applications in several data-intensive sciences.

These include questions of confounding control, policy analysis, misspecification tests, missing data, mediation, heterogeneity and the integration of data from diverse studies.

Biography: Judea Pearl is a professor of computer science and statistics at UCLA. He is a graduate of the Technion, Israel and has joined the faculty of UCLA in 1970, where he currently directs the Cognitive Systems Laboratory and conducts research in artificial intelligence, causal inference and philosophy of science.

Read more to see our video of Dr. Judea Pearl's Colloquium at Columbia University.

On the Computational and Statistical Interface and Big Data by Michael Jordan

Tuesday, October 15, 2013 - 3:00pm to 4:00pm
313 Fayerweather
Columbia University
New York, NY 10027
United States

Professor pictureMichael I. Jordan (born 1957) is an American scientist, Professor at the University of California, Berkeley and leading researcher in machine learning and artificial intelligence presented "On the Computational and Statistical Interface and Big Data" at a Seminar series event for the Data Science Institute at Columbia University on October 15, 2013.

Dr. David Blei of Princeton University

Thursday, September 12, 2013 - 6:00pm
Davis Auditorium, Room 412, Shapiro SEPSR
500 West 120th Street
New York, NY 10027
United States

Probabilistic topic models provide a suite of tools for analyzing large document collections. Topic modeling algorithms discover the latent themes that underlie the documents and identify how each document exhibits those themes. Topic modeling can be used to help explore, summarize, and form predictions about documents.

Pages

Subscribe to RSS - Colloquium Series
Back to Top