tag:blogger.com,1999:blog-14311460780255010172024-03-08T06:46:17.415-08:00UCSD AI seminarThis is the schedule of the AI seminar in the CSE department at UC San Diego, for the winter quarter of 2014. Talks cover machine learning and applications. All talks are at 12pm on a Monday, in room 4140, in the CSE building on the UCSD campus.
Later talks are at the top, and may be in the future. Please scroll down to find the full list of talks.Charles Elkanhttp://www.blogger.com/profile/10639994516997092663noreply@blogger.comBlogger20125tag:blogger.com,1999:blog-1431146078025501017.post-37715576179410660752013-12-29T16:42:00.004-08:002014-02-19T17:38:04.958-08:00<h3 style="background-color: white; color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; margin: 0px; position: relative;">
<span style="font-family: Arial, Helvetica, sans-serif;">3/10/14: </span><span style="background-color: transparent;"><span style="font-family: Arial, Helvetica, sans-serif;">Shlomo Dubnov</span></span></h3>
<div style="background-color: white; color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 13px; line-height: 18px;">
<span style="font-family: Arial, Helvetica, sans-serif;"><br /></span></div>
<div class="post-body entry-content" id="post-body-260607834746700431" itemprop="description articleBody" style="background-color: white; position: relative; width: 570px;">
<div dir="ltr" trbidi="on">
<div align="center" class="MsoNormal" style="text-align: center;">
<div align="center" class="MsoNormal">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 22px; line-height: 23.997394561767578px;"><b>Science of Story</b></span></span></div>
<div align="center" class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 15px; line-height: 1.4;">
<span style="font-size: 21.66666603088379px; line-height: 23.997394561767578px;"><b><br /></b></span></div>
</div>
<div align="center" class="MsoNormal" style="margin-bottom: 0.0001pt; text-align: center;">
<div align="center" class="MsoNormal" style="margin-bottom: 0.0001pt;">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 18.333332061767578px;">Shlomo Dubnov</span></span></div>
<div align="center" class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 15px; line-height: 1.4; margin-bottom: 0.0001pt;">
<span style="line-height: 18.333332061767578px;">Department of Music</span></div>
<div align="center" class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 15px; line-height: 1.4; margin-bottom: 0.0001pt;">
<span style="line-height: 18.333332061767578px;">UCSD</span></div>
</div>
<div class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 13px; line-height: 1.4;">
<br /></div>
<div class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 15px; line-height: 1.4; text-align: justify;">
<div class="MsoNormal">
<span style="line-height: 17.99479103088379px;">Abstract: Story is a delicate balance. Can a computer appreciate a good story? Probably not, at least not yet today. But it can help weigh the options and make cold-blooded calculations of how well story elements combine according to some commonly accepted script-writing formulas. In the talk I will describe an ongoing research on using NLP to track the structure of narrative in film scripts by embedding scenes in a semantic space and tracing their evolution over time. The method allows matching theoretical elements of story structure, such as theme, turning points, B-story, climax and resolution, and many other elements outlined in so called "beat-sheet" formulas to actual changes in word statistics over the duration of a film script. This automated analysis can be compared to previous research on green lighting movie scripts that uses human evaluations as predictors for commercial success of movies. Some speculations on universality of story structure in relation to human perception of epic / myth and musical form will be discussed.</span></div>
</div>
<div class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 13px; line-height: 1.4; text-align: justify;">
<span style="font-size: 12pt; line-height: 18px;"><br /></span></div>
<div class="MsoNormal" style="margin-bottom: 0.0001pt; text-align: justify;">
<div class="MsoNormal" style="margin-bottom: 0.0001pt;">
<div class="MsoNormal" style="margin-bottom: 0.0001pt;">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 18.333332061767578px;"><b>Shlomo Dubnov</b></span></span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 15px; line-height: 18.333332061767578px;"><b> </b></span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 15px; line-height: 18.333332061767578px;">is a Professor in music technology at UCSD. He received his PhD in Computer Science from Hebrew University and was a researcher in IRCAM, Paris and faculty in Communication Systems Engineering in Ben-Gurion University, Israel. Among his main contributions are new methods for statistical audio analysis/synthesis, modeling of emotions and aesthetics, and machine learning systems for musical improvisation. He co-edited a book “The Structure of Style: algorithmic approaches to understanding manner and meaning” and served as a secretary of IEEE Technical Committee on Computer Generated Music. Currently he serves as a co-lead editor of ACM Computers in Entertainment and directs Qualcomm Institute's Center for Research on Entertainment and Learning (CREL).</span></span></div>
</div>
</div>
</div>
</div>
Charles Elkanhttp://www.blogger.com/profile/10639994516997092663noreply@blogger.com0tag:blogger.com,1999:blog-1431146078025501017.post-18148093009309252462013-12-29T16:39:00.004-08:002014-02-19T17:35:07.209-08:00<h3 style="background-color: white; margin: 0px; position: relative;">
<span style="color: #222222; font-family: Arial, Helvetica, sans-serif;">3/3/14: </span><span style="background-color: transparent;"><span style="color: #222222; font-family: Arial, Helvetica, sans-serif;">Ery Arias-Castro</span></span></h3>
<div style="background-color: white; color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 13px; line-height: 18px;">
<span style="font-family: Arial, Helvetica, sans-serif;"><br /></span></div>
<div class="post-body entry-content" id="post-body-260607834746700431" itemprop="description articleBody" style="background-color: white; position: relative; width: 570px;">
<div dir="ltr" trbidi="on">
<div align="center" class="MsoNormal" style="text-align: center;">
<div align="center" class="MsoNormal">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 22px; line-height: 23.997394561767578px;"><b>Community detection in a random network</b></span></span></div>
<div align="center" class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;">
<span style="font-size: 21.66666603088379px; line-height: 23.997394561767578px;"><b><br /></b></span></div>
</div>
<div align="center" class="MsoNormal" style="margin-bottom: 0.0001pt; text-align: center;">
<div align="center" class="MsoNormal" style="margin-bottom: 0.0001pt;">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 18.333332061767578px;">Ery Arias-Castro</span></span></div>
<div align="center" class="MsoNormal" style="margin-bottom: 0.0001pt;">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 18.333332061767578px;">Department of Mathematics</span></span></div>
<div align="center" class="MsoNormal" style="margin-bottom: 0.0001pt;">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 18.333332061767578px;">UCSD</span></span></div>
</div>
<div class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 13px; line-height: 1.4;">
<br /></div>
<div class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 15px; line-height: 1.4; text-align: justify;">
<div class="MsoNormal">
<span style="line-height: 17.99479103088379px;">Abstract: We formalize the problem of detecting a community in a network into testing whether in a given (random) graph there is a subgraph that is unusually dense. Specifically, we observe an undirected and unweighted graph on N nodes. Under the null hypothesis, the graph is a realization of an Erdös-Rényi graph with probability p_0. Under the (composite) alternative, there is an unknown subgraph of n nodes where the probability of connection is p_1 > p_0. We derive detection lower bounds for detecting such a subgraph in terms of (N, n, p_0, p_1) in various regimes, and exhibit a number of tests that achieve that lower bound in some particular regime: the scan statistic and variants, the size of the largest connected component, the number of triangles, the eigengap of the adjacency matrix, etc. We also consider the problem of testing in polynomial-time. Our detection bounds are sharp, except in the Poisson regime where we were not able to fully characterize the constant arising in the bound. Joint work with Nicolas Verzelen (INRA, France).</span></div>
</div>
<div class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 13px; line-height: 1.4; text-align: justify;">
<span style="font-size: 12pt; line-height: 18px;"><br /></span></div>
<div class="MsoNormal" style="margin-bottom: 0.0001pt; text-align: justify;">
<div class="MsoNormal" style="margin-bottom: 0.0001pt;">
<div class="MsoNormal" style="margin-bottom: 0.0001pt;">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 18.333332061767578px;"><b>Ery Arias-Castro</b></span></span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 15px; line-height: 18.333332061767578px;"><b> </b>is an associate professor of statistics at UCSD. He </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 18.333332061767578px;">received his PhD in Statistics from Stanford University in 2004. His M.A. is in artificial intelligence, from the Ecole Normal Superieure de Cachan (France) and Washington University, Saint Louis, while his </span></span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 14.999999046325684px; line-height: 18.333332061767578px;">B.S. is in mathematics from the Ecole Normal Superieure de Cachan.</span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 14.999999046325684px; line-height: 18.333332061767578px;"> He joined the faculty in the mathematics department at UCSD in 2005. His research interests are in high-dimensional statistics, machine learning, spatial statistics, image processing, and applied probability.</span></div>
</div>
</div>
</div>
</div>
Charles Elkanhttp://www.blogger.com/profile/10639994516997092663noreply@blogger.com0tag:blogger.com,1999:blog-1431146078025501017.post-4456395829164054092013-12-29T16:37:00.004-08:002014-02-19T17:24:01.070-08:00<h3 style="background-color: white; margin: 0px; position: relative;">
<span style="color: #222222; font-family: Arial, Helvetica, sans-serif;">2/24/14: </span><span style="color: #222222; font-family: Arial, Helvetica, sans-serif;">Chun-Nan Hsu </span></h3>
<div style="background-color: white; color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 13px; line-height: 18px;">
<br /></div>
<div class="post-body entry-content" id="post-body-260607834746700431" itemprop="description articleBody" style="background-color: white; position: relative; width: 570px;">
<div dir="ltr" trbidi="on">
<div align="center" class="MsoNormal" style="text-align: center;">
<div align="center" class="MsoNormal">
<h2>
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 23px; line-height: 31.5px;">Identifying Transformative Research</span></span></h2>
</div>
</div>
<div align="center" class="MsoNormal" style="margin-bottom: 0.0001pt; text-align: center;">
<div align="center" class="MsoNormal" style="margin-bottom: 0.0001pt;">
<div align="center" class="MsoNormal" style="margin-bottom: 0.0001pt;">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 18.333332061767578px;">Chun-Nan Hsu</span></span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 15px; line-height: 18.333332061767578px;"> </span></div>
</div>
<div align="center" class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 15px; margin-bottom: 0.0001pt;">
<span style="line-height: 18.333332061767578px;">UCSD</span></div>
</div>
<div class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 13px; line-height: 1.4;">
<br /></div>
<div class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 15px; line-height: 1.4; text-align: justify;">
<div class="MsoNormal">
<span style="line-height: 17.99479103088379px;"><br /></span></div>
<div class="MsoNormal">
<span style="line-height: 17.99479103088379px;">Abstract. Transformative research refers to research that shifts or disrupts established scientific paradigms. Identifying potential transformative research early and accurately is important for funding agencies to maximize the impact of their investments. It also helps scientists identify and focus their attention on promising emerging works. In this talk, I will present a data-driven approach where citation patterns of scientific papers are analyzed to quantify how much a potential challenger idea shifts an established paradigm. I will present experimental results showing that some successful transformative research works disrupt established paradigms in Physics, Biomedical Sciences and Computer Science, regardless of whether the challenger paradigm is an instant hit or a classic whose contribution is formally recognized with a Nobel Prize decades later.</span></div>
</div>
<div class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 13px; line-height: 1.4; text-align: justify;">
<span style="font-size: 12pt; line-height: 18px;"><br /></span></div>
<div class="MsoNormal" style="margin-bottom: 0.0001pt; text-align: justify;">
<div class="MsoNormal" style="margin-bottom: 0.0001pt;">
<div class="MsoNormal" style="margin-bottom: 0.0001pt;">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 18.333332061767578px;"><b>Chun-Nan Hsu </b></span><span style="font-size: 15px; line-height: 18.333332061767578px;">has been Associate Professor in the Division of Biomedical Informatics, UC San Diego since November 2013. He is interested in biomedical data mining, text mining and cell image analysis. He has developed several widely-used bioinformatics tools and services, including top-performing text mining systems in BioCreative international contest series. His recent work focuses on advanced text-mining algorithms to establish knowledge-bases of phenotypes and genetic diseases. He was elected as the president of the Taiwanese Association for Artificial Intelligence in 2009 and a IBM faculty award recipient. He is a senior member of ACM.</span></span></div>
</div>
</div>
</div>
</div>
Charles Elkanhttp://www.blogger.com/profile/10639994516997092663noreply@blogger.com0tag:blogger.com,1999:blog-1431146078025501017.post-82901262685367536502013-12-29T16:31:00.004-08:002014-02-08T16:20:28.080-08:00<h3 style="background-color: white; margin: 0px; position: relative;">
<span style="color: #222222; font-family: Arial, Helvetica, sans-serif;">2/17/14: </span><span style="color: #222222; font-family: Arial, Helvetica, sans-serif;">No seminar</span></h3>
<div style="background-color: white; color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 13px; line-height: 18px;">
<br /></div>
<div class="post-body entry-content" id="post-body-260607834746700431" itemprop="description articleBody" style="background-color: white; position: relative; width: 570px;">
<div dir="ltr" trbidi="on">
<div align="center" class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 15px; line-height: 1.4; text-align: center;">
<div align="center" class="MsoNormal">
<h2>
<br /></h2>
</div>
</div>
<div class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 13px; line-height: 1.4;">
<i style="font-size: 15px; line-height: 17.99479103088379px; text-align: justify;">Monday February 17 is Presidents' Day and an official holiday at UCSD. </i></div>
</div>
</div>
Charles Elkanhttp://www.blogger.com/profile/10639994516997092663noreply@blogger.com0tag:blogger.com,1999:blog-1431146078025501017.post-76521510673191114452013-12-29T16:26:00.000-08:002014-02-08T16:22:08.330-08:00<h3 style="background-color: white; margin: 0px; position: relative;">
<span style="color: #222222; font-family: Arial, Helvetica, sans-serif;">2/10/14: </span><span style="color: #222222; font-family: Arial, Helvetica, sans-serif;">Amit K. Roy Chowdhury </span></h3>
<div>
<br /></div>
<div class="post-body entry-content" id="post-body-260607834746700431" itemprop="description articleBody" style="background-color: white; position: relative; width: 570px;">
<div dir="ltr" trbidi="on">
<div align="center" class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 15px; line-height: 1.4; text-align: center;">
<div align="center" class="MsoNormal">
<h2>
Situation Awareness from Wide-Area Vision Networks</h2>
</div>
</div>
<div align="center" class="MsoNormal" style="margin-bottom: 0.0001pt; text-align: center;">
<div align="center" class="MsoNormal" style="margin-bottom: 0.0001pt;">
<div align="center" class="MsoNormal" style="margin-bottom: 0.0001pt;">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 18.333332061767578px;">Amit K. Roy Chowdhury </span></span></div>
</div>
<div align="center" class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 15px; line-height: 1.4; margin-bottom: 0.0001pt;">
<span style="line-height: 18.333332061767578px;">University of California, Riverside</span></div>
</div>
<div class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 13px; line-height: 1.4;">
<br /></div>
<div class="MsoNormal" style="text-align: justify;">
<div class="MsoNormal">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 15px; line-height: 17.99479103088379px;">Abstract. </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 17.99479103088379px;">Over the past decade, large-scale camera networks have become increasingly prevalent in a wide range of applications, such as security and surveillance, disaster response, and environmental modeling. However, the analysis of the acquired videos has been largely manual and post-facto. Thus, the development of algorithms capable of analyzing a scene covering a wide area is extremely important. In this talk, we will focus on three inter-related problems in this domain.</span></span><br />
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 17.99479103088379px;">i) The performance of the analysis algorithms often suffers because of the inability to effectively acquire the desired images. We will discuss our recent work on integrated sensing and analysis in a distributed camera network so as to maximize various scene-understanding performance criteria (e.g., tracking accuracy, best shot, and image resolution). We will show how the existing work in autonomous multiagent systems can be leveraged for this purpose - more specifically, game theory-based distributed optimization algorithms for dynamic camera network reconfiguration.</span></span><br />
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 17.99479103088379px;">ii) In application domains where there is no central processor accumulating all the data (e.g., disaster response), inferences have to be drawn through local decisions at the camera nodes and negotiations with the neighbors. We will present our work on distributed reasoning in vision networks, especially the recently proposed Information Weighted Consensus Filter (ICF). The application of ICF to multi-target tracking will then be presented.</span></span><br />
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 17.99479103088379px;">iii) Finally, we will address the issue of higher-level scene understanding. The recognition of activities in video is an essential step in this regard. Activities happening over a wide-area are often related in space and time. We will show how graphical inference methods can be used to robustly recognize such activities, specifically taking into account the contextual relationships between them.</span></span><br />
<div style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 15px; line-height: 1.4;">
<br /></div>
</div>
</div>
<div class="MsoNormal" style="margin-bottom: 0.0001pt; text-align: justify;">
<div class="MsoNormal" style="margin-bottom: 0.0001pt;">
<div class="MsoNormal" style="margin-bottom: 0.0001pt;">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 18.333332061767578px;"><b>Amit K. Roy Chowdhury </b></span><span style="font-size: 14.999999046325684px; line-height: 18.333332061767578px;">received his undergraduate degree in electrical engineering from Jadavpur University, Calcutta, India, his Masters degree in systems science and automation from the Indian Institute of Science, Bangalore, India, and the Ph.D. degree in electrical engineering from the University of Maryland, College Park. He is a Professor of Electrical Engineering and a Cooperating Faculty in the Department of Computer Science at the University of California, Riverside. His broad research interests include the areas of image processing and analysis, computer vision, and statistical signal processing and pattern recognition. Together with his students and collaborators, he has over 100 technical publications in these areas, including one best student paper award. His current research projects include intelligent camera networks, wide-area scene analysis, motion analysis in video, activity recognition and search, video-based biometrics (face and gait), and biological video analysis. He is the first author of the book - Camera Networks: The Acquisition and Analysis of Videos over Wide Areas - the first research monograph on this topic. He has been on the organizing and program committees of multiple computer vision and image processing conferences and is serving on the editorial boards of multiple journals.</span></span></div>
</div>
</div>
</div>
</div>
<br />Charles Elkanhttp://www.blogger.com/profile/10639994516997092663noreply@blogger.com0tag:blogger.com,1999:blog-1431146078025501017.post-33378539233037090102013-12-29T10:49:00.002-08:002014-01-02T13:43:17.644-08:00<h3 style="background-color: white; color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; margin: 0px; position: relative;">
<span style="font-family: Arial, Helvetica, sans-serif;">2/3/14: Christian Shelton</span></h3>
<div style="background-color: white; color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 13px; line-height: 18px;">
<span style="font-family: Arial, Helvetica, sans-serif;"><br /></span></div>
<div class="post-body entry-content" id="post-body-260607834746700431" itemprop="description articleBody" style="background-color: white; position: relative; width: 570px;">
<div dir="ltr" trbidi="on">
<div align="center" class="MsoNormal" style="text-align: center;">
<div align="center" class="MsoNormal">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 21.66666603088379px; line-height: 23.997394561767578px;"><b>Machine Learning and Critical Care Pediatrics</b></span></span></div>
</div>
<div align="center" class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 13px; line-height: 1.4; text-align: center;">
<br /></div>
<div align="center" class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 15px; margin-bottom: 0.0001pt; text-align: center;">
<div align="center" class="MsoNormal" style="margin-bottom: 0.0001pt;">
<span style="line-height: 18.333332061767578px;">Christian Shelton</span></div>
<div align="center" class="MsoNormal" style="margin-bottom: 0.0001pt;">
<span style="line-height: 18.333332061767578px;">University of California, Riverside</span></div>
</div>
<div class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 13px; line-height: 1.4;">
<br /></div>
<div class="MsoNormal" style="text-align: justify;">
<div class="MsoNormal">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 17.99479103088379px;">Electronic health records provide the opportunity for data-driven medical </span></span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 14.999999046325684px; line-height: 17.99479103088379px;">discovery, even with all of their current flaws. Intensive care units are </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 14.999999046325684px; line-height: 17.99479103088379px;">particularly interesting microcosm as their data are relatively frequent </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 17.99479103088379px;">and some types of outcomes are more quickly known. </span></span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 14.999999046325684px; line-height: 17.99479103088379px;">In this talk, I will first outline the type of data, scientific questions, </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 14.999999046325684px; line-height: 17.99479103088379px;">and domain challenges Children's Hospital Los Angeles and my group have </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 14.999999046325684px; line-height: 17.99479103088379px;">been working on to improve critical care. Then I will describe one project </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 14.999999046325684px; line-height: 17.99479103088379px;">on estimating blood gas levels for children on mechanical ventilation. </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 14.999999046325684px; line-height: 17.99479103088379px;">This project aims to remove invasive tests and provide faster weaning </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 14.999999046325684px; line-height: 17.99479103088379px;">off of ventilation to decrease costs and improve health.</span><br />
<b style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 14.999999046325684px; line-height: 18.333332061767578px;"><br /></b>
<b style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 14.999999046325684px; line-height: 18.333332061767578px;">Christian Shelton </b><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 18.333332061767578px;">is an Associate Professor of Computer Science at the </span></span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 14.999999046325684px; line-height: 18.333332061767578px;">University of California at Riverside. He joined the faculty in 2003. His </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 14.999999046325684px; line-height: 18.333332061767578px;">research interest is in statistical approaches to artificial intelligence, </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 14.999999046325684px; line-height: 18.333332061767578px;">mainly in the areas of machine learning and dynamic processes. He has </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 14.999999046325684px; line-height: 18.333332061767578px;">been the Managing Editor of the Journal of Machine Learning Research and </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 14.999999046325684px; line-height: 18.333332061767578px;">on the editorial board of the Journal of Artificial </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 18.333332061767578px;">Intelligence Research. </span></span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 14.999999046325684px; line-height: 18.333332061767578px;">Dr. Shelton received his B.S. in Computer Science from Stanford University </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 14.999999046325684px; line-height: 18.333332061767578px;">in 1996 and his Ph.D. from MIT in 2001. From 2001 to 2003, he was a post</span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 14.999999046325684px; line-height: 18.333332061767578px;">doctoral scholar back at Stanford. He has been a visiting researcher at</span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 14.999999046325684px; line-height: 18.333332061767578px;">Intel Research (2003-2004) and Children's Hospital Los Angeles (2012-2013).</span><br />
<div>
<br /></div>
</div>
</div>
</div>
</div>
Charles Elkanhttp://www.blogger.com/profile/10639994516997092663noreply@blogger.com0tag:blogger.com,1999:blog-1431146078025501017.post-39810460434968381602013-12-29T10:46:00.000-08:002013-12-29T10:46:00.604-08:00<h3 style="background-color: white; color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; margin: 0px; position: relative;">
<span style="font-family: Arial, Helvetica, sans-serif;">1/27/14 </span></h3>
<div style="background-color: white; color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 13px; line-height: 18px;">
<span style="font-family: Arial, Helvetica, sans-serif;"><br /></span></div>
<div class="post-body entry-content" id="post-body-260607834746700431" itemprop="description articleBody" style="background-color: white; color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 15px; position: relative; width: 570px;">
<div dir="ltr" trbidi="on">
<div align="center" class="MsoNormal" style="line-height: 1.4; text-align: center;">
<div align="center" class="MsoNormal">
<span style="font-size: 21.66666603088379px; line-height: 23.997394561767578px;"><b><br /></b></span></div>
<div align="center" class="MsoNormal">
<span style="font-size: 21.66666603088379px; line-height: 23.997394561767578px;"><b>Reports of Interesting Research from NIPS</b></span></div>
</div>
<div align="center" class="MsoNormal" style="font-size: 13px; line-height: 1.4; text-align: center;">
<br /></div>
<div align="center" class="MsoNormal" style="line-height: 1.4; margin-bottom: 0.0001pt; text-align: center;">
<div align="center" class="MsoNormal" style="margin-bottom: 0.0001pt;">
<br /></div>
</div>
<div class="MsoNormal" style="text-align: justify;">
<div class="MsoNormal">
<span style="line-height: 17.99479103088379px;">In this session, UCSD grad students will describe especially interesting research presented at NIPS in December 2013.</span></div>
</div>
</div>
</div>
Charles Elkanhttp://www.blogger.com/profile/10639994516997092663noreply@blogger.com0tag:blogger.com,1999:blog-1431146078025501017.post-44446142967184300752013-12-29T10:43:00.002-08:002013-12-29T10:43:40.836-08:00<h3 style="background-color: white; margin: 0px; position: relative;">
<span style="color: #222222; font-family: Arial, Helvetica, sans-serif;">1/20/14: </span><span style="background-color: transparent;"><span style="color: #222222; font-family: Arial, Helvetica, sans-serif;">No seminar.</span></span></h3>
<div class="post-body entry-content" id="post-body-260607834746700431" itemprop="description articleBody" style="background-color: white; position: relative; width: 570px;">
<div dir="ltr" trbidi="on">
<div align="center" class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; text-align: center;">
<div align="center" class="MsoNormal">
<br /></div>
</div>
<div class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 15px; text-align: justify;">
<div class="MsoNormal">
<span style="line-height: 17.99479103088379px;">No meeting because of Martin Luther King day.</span></div>
</div>
</div>
</div>
Charles Elkanhttp://www.blogger.com/profile/10639994516997092663noreply@blogger.com0tag:blogger.com,1999:blog-1431146078025501017.post-24879403820435062152013-12-29T10:42:00.004-08:002014-01-08T21:50:36.367-08:00<h3 style="background-color: white; margin: 0px; position: relative;">
<span style="color: #222222; font-family: Arial, Helvetica, sans-serif;">1/13/14: </span><span style="background-color: transparent;"><span style="color: #222222; font-family: Arial, Helvetica, sans-serif;">Sumithra Velupillai</span></span></h3>
<div style="background-color: white; color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 13px; line-height: 18px;">
<span style="font-family: Arial, Helvetica, sans-serif;"><br /></span></div>
<div class="post-body entry-content" id="post-body-260607834746700431" itemprop="description articleBody" style="background-color: white; position: relative; width: 570px;">
<div dir="ltr" trbidi="on">
<div align="center" class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; text-align: center;">
<div align="center" class="MsoNormal">
<h2>
<span style="font-family: arial, sans-serif; text-align: start;"><span style="font-size: small;">Shades of Certainty -- Working with Swedish Medical Records</span></span></h2>
</div>
</div>
<div align="center" class="MsoNormal" style="margin-bottom: 0.0001pt; text-align: center;">
<div align="center" class="MsoNormal" style="margin-bottom: 0.0001pt;">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 18.333332061767578px;">Sumithra Velupillai</span></span></div>
<div align="center" class="MsoNormal" style="margin-bottom: 0.0001pt;">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 18.333332061767578px;">Division of Biomedical Informatics</span></span></div>
<div align="center" class="MsoNormal" style="margin-bottom: 0.0001pt;">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 18.333332061767578px;">UCSD</span></span></div>
</div>
<div class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 13px; line-height: 1.4;">
<br /></div>
<div class="MsoNormal" style="text-align: justify;">
<div class="MsoNormal">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 15px; line-height: 17.99479103088379px;">Abstract. </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 14.999999046325684px; line-height: 17.99479103088379px;">Different levels of knowledge certainty, or factuality levels, are expressed in clinical health record documentation. This information is currently not fully exploited, as the subtleties expressed in natural language cannot easily be machine analyzed. </span><br />
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 14.999999046325684px; line-height: 17.99479103088379px;"><br /></span>
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 14.999999046325684px; line-height: 17.99479103088379px;">Two annotated corpora have been created for capturing speculations and uncertainties in Swedish medical records. </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 14.999999046325684px; line-height: 17.99479103088379px;">One model distinguishes certain and uncertain expressions on a sentence level, and is applied on medical documentation from several clinical departments. Differences between clinical practices are also studied. More fine-grained certainty level distinctions are presented in a second model, with two polarities along with three levels of certainty, and is applied on a diagnostic statement level from an emergency department. Overall agreement results for both models are promising, but differences are seen depending on clinical practice, the definition of the annotation task and the level of domain expertise among the annotators. </span><br />
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 17.99479103088379px;"><br /></span></span>
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 17.99479103088379px;">Using annotated resources for automatic classification of certainty levels is also studied by employing machine learning techniques. Encouraging overall results using local context information are obtained. The fine-grained certainty level model is also used for building classifiers for coarser-grained, real-world e-health scenarios, showing that fine-grained annotations can be used for several e-health scenario tasks. </span></span><br />
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 17.99479103088379px;"><br /></span></span>
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 17.99479103088379px;">This talk will also present ongoing research on Swedish medical records and the Stockholm EPR Corpus from the Clinical Text Mining Group at the Department of Computer and Systems Sciences, Stockholm University. </span></span><br />
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 17.99479103088379px;"><br /></span></span></div>
</div>
<div class="MsoNormal" style="margin-bottom: 0.0001pt; text-align: justify;">
<div class="MsoNormal" style="margin-bottom: 0.0001pt;">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 14.999999046325684px; line-height: 18.333332061767578px;"><b>Sumithra Velupillai</b>, Ph.D.,</span></span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 14.999999046325684px; line-height: 18.333332061767578px;"> is a postdoctoral researcher at UCSD, coming from the Department of Computer and Systems Sciences at Stockholm University. She has been awarded an international postdoctoral fellowship from the Swedish Research Council along with a Fulbright scholarship, in which she will base a majority of her research at UCSD during 2014-2015. She successfully defended her Thesis "Shades of Certainty – Annotation and Classification of Swedish Medical Records" on April 27th, 2012. Velupillai has participated in several national and international research projects, among others the Interlock project - a research collaboration between Stockholm University and DBMI at UCSD. Velupillai has a background in Computational Linguistics and specializes in research covering Language Technology, Information Access and Extraction, and Health Informatics.</span></div>
</div>
</div>
</div>
Charles Elkanhttp://www.blogger.com/profile/10639994516997092663noreply@blogger.com0tag:blogger.com,1999:blog-1431146078025501017.post-33919767318213909192013-12-29T10:38:00.000-08:002013-12-29T10:38:46.933-08:00<h3>
<span style="font-family: Arial, Helvetica, sans-serif;">1/6/14: Steve Gallant</span></h3>
<div>
<span style="font-family: Arial, Helvetica, sans-serif;"><br /></span></div>
<div class="post-body entry-content" id="post-body-260607834746700431" itemprop="description articleBody" style="background-color: white; position: relative; width: 570px;">
<div dir="ltr" trbidi="on">
<div align="center" class="MsoNormal" style="text-align: center;">
<div align="center" class="MsoNormal">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="font-size: 21.66666603088379px; line-height: 23.997394561767578px;"><b>Representing Free Text for Machine Learning </b></span></span><b style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 21.66666603088379px; line-height: 23.997394561767578px;">Using High-Dimensional, Distributed Vectors</b></div>
</div>
<div align="center" class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 13px; line-height: 1.4; text-align: center;">
<br /></div>
<div align="center" class="MsoNormal" style="margin-bottom: 0.0001pt; text-align: center;">
<div align="center" class="MsoNormal" style="margin-bottom: 0.0001pt;">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="line-height: 18.333332061767578px;">Steve Gallant</span></span></div>
<div align="center" class="MsoNormal" style="margin-bottom: 0.0001pt;">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="line-height: 18.333332061767578px;">MultiModel Research</span></span></div>
<div align="center" class="MsoNormal" style="margin-bottom: 0.0001pt;">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="line-height: 18.333332061767578px;">Cambridge, MA</span></span></div>
</div>
<div class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 13px; line-height: 1.4;">
<br /></div>
<div class="MsoNormal" style="text-align: justify;">
<div class="MsoNormal">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="line-height: 17.99479103088379px;">We want a way to represent natural language, including sentence structure, to </span></span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; line-height: 17.99479103088379px;">make it easy for machine learning (back propagation, perceptron learning). This </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; line-height: 17.99479103088379px;">places certain constraints upon the Representation. Here the Binding problem is </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; line-height: 17.99479103088379px;">especially important; for example, we need to represent the binding of adjectives </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; line-height: 17.99479103088379px;">to the nouns they modify, and nouns to their roles (actor, agent). </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; line-height: 17.99479103088379px;">The talk will present a neurally-inspired representation, MBAT (matrix binding </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; line-height: 17.99479103088379px;">of additive terms), that represents a sentence by a single high-dimensional </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; line-height: 17.99479103088379px;">distributed vector. We argue that this representation satisfies the constraints. It </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; line-height: 17.99479103088379px;">also permits us, to prove that certain concepts can be learned, rather than relying </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; line-height: 17.99479103088379px;">on simulations. </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="line-height: 17.99479103088379px;">A project is underway to test MBAT techniques with real-world problems, such as </span></span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; line-height: 17.99479103088379px;">sentiment analysis. The architecture also has implications for Cognitive Science.</span></div>
</div>
<div class="MsoNormal" style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; font-size: 13px; line-height: 1.4; text-align: justify;">
<span style="font-size: 12pt; line-height: 18px;"><br /></span></div>
<div class="MsoNormal" style="margin-bottom: 0.0001pt; text-align: justify;">
<div class="MsoNormal" style="margin-bottom: 0.0001pt;">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="line-height: 18.333332061767578px;"><b>Steve Gallant </b>developed several neural network learning algorithms when </span></span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; line-height: 18.333332061767578px;">Associate Professor of Computer Science at Northeastern University. He showed </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; line-height: 18.333332061767578px;">how to interpret a neural network as an expert system, and was one of the first to </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; line-height: 18.333332061767578px;">extract rules from a neural network. </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; line-height: 18.333332061767578px;">He led a document retrieval project at HNC in San Diego based upon context </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; line-height: 18.333332061767578px;">vectors (which he invented), a precursor to current research that could </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; line-height: 18.333332061767578px;">not represent sentence structure. (HNC created a spinoff based upon this </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; line-height: 18.333332061767578px;">technology.)</span></div>
<div class="MsoNormal" style="margin-bottom: 0.0001pt;">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="line-height: 18.333332061767578px;"><br /></span></span></div>
<div class="MsoNormal" style="margin-bottom: 0.0001pt;">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="line-height: 18.333332061767578px;">Currently Steve leads an NSF supported research project on representation and </span></span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; line-height: 18.333332061767578px;">machine learning at a Cambridge, MA startup: MultiModel Research. </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; line-height: 18.333332061767578px;">He has over 40 publications and a book on “Neural Network Learning.”</span></div>
<div class="MsoNormal" style="margin-bottom: 0.0001pt;">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; line-height: 18.333332061767578px;"><br /></span></div>
<div class="MsoNormal" style="margin-bottom: 0.0001pt;">
<span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif;"><span style="line-height: 18.333332061767578px;">Education: </span></span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; line-height: 18.333332061767578px;">MIT: Undergrad (Math), </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; line-height: 18.333332061767578px;">Stanford: Ph.D. (Operations Research), </span><span style="color: #222222; font-family: Arial, Tahoma, Helvetica, FreeSans, sans-serif; line-height: 18.333332061767578px;">University of Waterloo: Post Doc (Combinatorics & Optimization).</span></div>
</div>
</div>
</div>
Charles Elkanhttp://www.blogger.com/profile/10639994516997092663noreply@blogger.com0tag:blogger.com,1999:blog-1431146078025501017.post-2606078347467004312011-11-27T17:50:00.008-08:002012-03-06T15:37:20.373-08:003/12/12: Tamara Sipes (UCSD ECE)<div dir="ltr" style="text-align: left;" trbidi="on"><br />
<div align="center" class="MsoNormal" style="text-align: center;"><b><span style="font-size: 16pt; line-height: 115%;">Multivariate Time Series Classification Using Temporal Metafeature Abstractions<o:p></o:p></span></b></div><div align="center" class="MsoNormal" style="text-align: center;"><br />
</div><div align="center" class="MsoNormal" style="margin-bottom: 0.0001pt; text-align: center;"><span style="font-size: 12pt;">Tamara B. Sipes, Ph.D.<o:p></o:p></span></div><div align="center" class="MsoNormal" style="margin-bottom: 0.0001pt; text-align: center;"><span style="font-size: 12pt;">Space Plasma Physics Lab, UCSD<o:p></o:p></span></div><div align="center" class="MsoNormal" style="margin-bottom: 0.0001pt; text-align: center;"><span style="font-size: 12pt;">SciberQuest, Inc.</span><span style="font-size: 12pt;"><o:p></o:p></span></div><div class="MsoNormal"><br />
</div><div class="MsoNormal" style="text-align: justify;"><span style="font-size: 12pt; line-height: 115%;">Extraction of knowledge from massive and complex data sets poses a major obstacle to scientific progress, even more so when the data is in the form of time series. We demonstrate a new approach to the classification of multivariate time series data by utilizing an innovative feature extraction technique in combination with a specialized data mining algorithm. The technique extracts global features and metafeatures in order to capture the necessary time-lapse information. The features are then used to create a static, intermediate data set that includes all the important time-varying information and is suitable for analysis using the standard supervised data mining techniques. The viability of the new algorithm called MineTool-TS is demonstrated through its application to the problem of automatic detection of flux transfer events in spacecraft data and mining of simulation data. The technique has also been successfully applied to a variety of medical, biomedical, environmental and space physics data.<o:p></o:p></span></div><div class="MsoNormal" style="text-align: justify;"><span style="font-size: 12pt; line-height: 115%;"><br />
</span></div><div class="MsoNormal" style="margin-bottom: 0.0001pt; text-align: justify;"><b><span style="font-size: 12.0pt; mso-bidi-font-family: Calibri; mso-bidi-theme-font: minor-latin;">Tamara B. Sipes, Ph.D.</span></b><span style="font-size: 12.0pt; mso-bidi-font-family: Calibri; mso-bidi-theme-font: minor-latin;"> </span><span style="font-size: 12.0pt; mso-bidi-font-family: Calibri; mso-bidi-theme-font: minor-latin; mso-fareast-font-family: "Times New Roman";">is a researcher at the Space Physics lab at UCSD specializing in data mining, predictive modeling and computational algorithms. She has a substantial industry and research experience in building highly innovative predictive models and creative solutions to complex problems. Her data mining expertise has been applied to a variety of data, industries and areas, including space physics, biotechnology, scientific, financial, robotic, educational and signal processing data. Her work has led to several patent applications and published research in the areas of data mining and learning technologies. In addition to her position at SciberQuest, Inc. Dr. Sipes is also an instructor at the University of California San Diego Extension where she created and teaches several Data Mining courses. </span><span style="font-size: 12.0pt; mso-bidi-font-family: Calibri; mso-bidi-theme-font: minor-latin;"><o:p></o:p></span></div><div class="MsoNormal" style="margin-bottom: 0.0001pt; text-align: justify;"><span style="font-size: 12.0pt; mso-bidi-font-family: Calibri; mso-bidi-theme-font: minor-latin; mso-fareast-font-family: "Times New Roman";"><br />
</span></div><div class="MsoNormal" style="margin-bottom: 0.0001pt; text-align: justify;"><span style="font-size: 12.0pt; mso-bidi-font-family: Calibri; mso-bidi-theme-font: minor-latin;">Besides her appointment at U<a href="" name="_GoBack"></a>CSD, Dr. Sipes is a Vice President of Analytics as SciberQuest, Inc. a company specializing in providing advanced solutions to the most complex computational and data analysis challenges facing the scientific world, including the development of self-adaptive algorithms for modeling of complex, multi-scale problems, computational infrastructure for NASA's magnetospheric virtual observatory , specialized scientific simulations for the Air Force including a first-ever, multi-resolution, 3D, parallel, object-oriented, electromagnetic PIC (particle in cell) code capable of handling complex boundaries, as well as advanced data mining methods for predictive analytics of static, time series, image and simulation data. </span><span style="font-size: 12.0pt; mso-bidi-font-family: Calibri; mso-bidi-theme-font: minor-latin; mso-fareast-font-family: "Times New Roman";">Dr. Sipes earned her Ph.D. in Computer Science from Vanderbilt University.</span><span style="font-size: 12.0pt; mso-bidi-font-family: Calibri; mso-bidi-theme-font: minor-latin;"><o:p></o:p></span></div></div>Charles Elkanhttp://www.blogger.com/profile/10639994516997092663noreply@blogger.com0tag:blogger.com,1999:blog-1431146078025501017.post-72987986025508456042011-11-27T17:50:00.007-08:002012-03-01T16:30:47.380-08:003/5/12: Doug White (UC Irvine Anthropology)<div dir="ltr" style="text-align: left;" trbidi="on"><b style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; text-align: -webkit-auto;">Networks, Causality and Evolution of Cooperation</b><br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; text-align: -webkit-auto;"><b><br />
</b></span><br />
Abstract: The 1927 Menger theorem proves the equivalence of a maximal k-cohesive set of nodes in a network to a maximal set of nodes in which all pairs of nodes are at least k-connected. We call these equivalent concepts ''stru-cohesion.'' It has a record of powerful causal predictions in social networks. This talk will present six types of examples, and end with a discussion of how stru-cohesion outperforms existing rules for the evolution of cooperation in human groups.<br />
<br />
<br />
Bio: Doug White is a mathematical anthropology faculty member at UCI and complexity scientist on the external faculty at SFI. He designed one of the major cross-cultural databases in the social sciences (SCCS), founded the prototype for the Kinsources datasite for the study of community social structure, leads a causality research group at SFI working on evolutionary causality, co-authored algorithms for statistical entailment analysis, regular equivalence network role analysis, and stru-cohesion. He has published spates of articles and numbers of books on social networks, mathematical sociology, the comparative network study of human kinship communities, world system economic networks, and the dynamics of global inter-urban networks. He founded the eScholarship World Cultures and Structure and Dynamics ejournals and continues to edit the latter.<br />
<br />
<a href="http://eclectic.ss.uci.edu/~drwhite/" style="font-family: arial, sans-serif; font-size: small;">http://eclectic.ss.uci.edu/~drwhite/</a><br />
<div><br />
</div><br />
<div><span class="Apple-style-span" style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px;"><br />
</span></div></div>Charles Elkanhttp://www.blogger.com/profile/10639994516997092663noreply@blogger.com0tag:blogger.com,1999:blog-1431146078025501017.post-51821081716278180412011-11-27T17:50:00.001-08:002012-02-05T14:16:34.378-08:002/27/12: Lorenzo Torresani (Dartmouth)<div dir="ltr" style="text-align: left;" trbidi="on"><br />
<b><span style="font-size: large;">Learning a Compact Image Code for Efficient Recognition of Novel Classes</span></b><br />
<br />
Lorenzo Torresani<br />
<br />
Assistant Professor of Computer Science<br />
Dartmouth College<br />
<br />
<a href="http://www.cs.dartmouth.edu/~lorenzo/home.html">http://www.cs.dartmouth.edu/~lorenzo/home.html</a><br />
<div><br />
</div>Abstract: In this talk I will discuss methods enabling efficient object-class recognition in large image collections. We are specifically interested in scenarios where the classes to be recognized are not known in advance. The motivating application is "object-class search by example" where a user provides at query time a small set of training images defining an arbitrary novel category and the system must retrieve images belonging to this class from a large database. This application scenario poses challenging requirements on the system design: the object classifier must be learned efficiently at query time from few examples; recognition must have low computational cost with respect to the database size; finally, compact image descriptors must be used to allow storage of large collections in memory.<br />
<br />
We propose to address these requirements by learning a compact image code optimized to yield good categorization accuracy with linear (i.e., efficient) classifiers: even when the representation is compressed to less than 300 bytes per image, linear classifiers trained on our descriptor yield accuracy matching the state-of-the-art but at orders of magnitude lower computational cost.<br />
<br />
<br />
<br />
</div>Charles Elkanhttp://www.blogger.com/profile/10639994516997092663noreply@blogger.com0tag:blogger.com,1999:blog-1431146078025501017.post-83968630389270168782011-11-27T17:42:00.001-08:002012-01-03T16:00:56.171-08:002/20/12: No seminar--Presidents' day<div dir="ltr" style="text-align: left;" trbidi="on"><br />
</div>Charles Elkanhttp://www.blogger.com/profile/10639994516997092663noreply@blogger.com0tag:blogger.com,1999:blog-1431146078025501017.post-1623393793439770652011-11-27T17:41:00.001-08:002012-01-26T13:19:17.792-08:002/13/12: Lilia Iakoucheva (UCSD Psychiatry): Prediction of protein post-translational modifications using machine learning approaches<div dir="ltr" style="text-align: left;" trbidi="on"><b><span style="font-size: large;"><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; text-align: -webkit-auto;">Prediction of protein post-translational modifications using machine learning approaches</span></span></b><br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;"><br />
</span><br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">Abstract: </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">I will describe how machine learning approaches could help biologists to predict the sites of posttranslational modifications in proteins using two examples - phosphorylation and ubiquitination. I will also briefly summarize the ongoing systems biology projects that we are currently working on in my lab.</span><br />
<br />
<div style="text-align: left;"><br />
</div><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">Bio: </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">Dr. Lilia Iakoucheva received her PhD degree from the Institute of Immunology, Moscow, Russia. After completing postdoctoral training in protein biochemistry and protein structure/intrinsic disorder analysis, she joined The Rockefeller University (New York, NY) as a Research Assistant Professor, and then the faculty of the UCSD Department of Psychiatry as an Assistant Professor. Dr. Iakoucheva is applying her experience in protein structure and protein-protein interactions analysis towards investigation of psychiatric disorders. Her research focuses on understanding molecular basis of psychiatric diseases using systems biology approaches. Dr. Iakoucheva has been the principal investigator on research grants from NSF, NCI, NICHD, and NIMH.</span><br style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;" /><div style="text-align: -webkit-auto;"><span style="color: #222222; font-family: arial, sans-serif; font-size: x-small;"><br />
</span></div><br />
<a href="http://psychiatry.ucsd.edu/faculty/lIakoucheva.html">http://psychiatry.ucsd.edu/faculty/lIakoucheva.html</a></div>Charles Elkanhttp://www.blogger.com/profile/10639994516997092663noreply@blogger.com0tag:blogger.com,1999:blog-1431146078025501017.post-30011839486137802192011-11-27T17:40:00.001-08:002012-01-15T09:53:07.967-08:002/6/12: Joseph Barr: Risk Scoring and Future Directions (Id Analytics, Inc.)<div dir="ltr" style="text-align: left;" trbidi="on"><div class="MsoNormal" style="text-align: center;"><div style="text-align: left;"><b><span style="font-family: Arial, sans-serif; font-size: large;">Risk Scoring and Future Directions<o:p></o:p></span></b></div></div><div align="center" class="MsoNormal" style="text-align: center;"><div style="text-align: left;"><br />
</div></div><div class="MsoNormal"><span style="font-family: inherit;"><u>Part 1:</u> ID Analytics main business is scoring applications (for credit/services) for risks including identity/authenticity & credit. By definition an application is a vector of identity elements (SSN, Name, Address, Phone, DOB, more), a vector known as “SNAPD”, as well as additional fields. ID Analytics process the data, extract pertinent features and calculate risk score on the fly. The entire process has a sub-second latency. At the basis of our analytics is the <i>ID Network</i> – a virtual graph with SNAPD-vectors as nodes. One can envision making a connection between two nodes if they share some identity element. The weight of the edge is the strength of the connection. As one can imagine various graphical parameters are the predominant inputs to our risk models. At the time I write this, the ID network has 1.5 billion nodes (corresponding to number of transactions); this of course means that the graph is too large to be stored in memory, and needless to say, how we do it is a trade secret, but I will indicate some principles behind the ideas.<br />
<o:p></o:p></span></div><div class="MsoNormal"><span style="font-family: inherit;"><u><br />
Part 2:</u> The risk ID Analytics is scoring falls under the more general rubric of consumer behavior. We are interested in the spatial / temporal aspects of our network and how it related to macroeconomic and social data including demographics, geography, housing, census, interest rates, unemployment, federal deficit, foreign balance of trade and whatnot. Under certain conditions, we will avail our data to an outside organization to participate in publishable research. <o:p></o:p></span></div><div class="MsoNormal"><br />
</div><div class="MsoNormal"><span style="font-family: inherit;">Introducing <i>id:a labs</i>, a research-oriented organization which promotes collaborations with academia and other research institutions. <br />
<u><br />
</u></span></div><span style="font-family: inherit;"><u>Bio<br />
</u>Joseph Barr is the Chief Scientist at ID Analytics (<a href="http://www.idanalytics.com/">www.idanalytics.com</a>). After a few years in academia (as assistant professor at California Lutheran University,) he has spent the past 16+ years in industry as a risk & consumer behavior (analytics) professional. He was awarded a Ph.D. in mathematics from the University of New Mexico in 1991 on his work on graph colorings. His current interests include the application of statistics, machine-learning and combinatorial algorithms to risk management and consumer behavior. <br />
<br />
</span><a href="http://www.linkedin.com/in/barranalytics"><span style="font-family: inherit;">http://www.linkedin.com/in/barranalytics</span></a></div>Charles Elkanhttp://www.blogger.com/profile/10639994516997092663noreply@blogger.com0tag:blogger.com,1999:blog-1431146078025501017.post-35851374645183279022011-11-27T17:39:00.005-08:002012-01-29T14:37:05.063-08:001/30/12: Vaclav Petricek: Data-driven Matchmaking at Scale (eHarmony, Inc.)<div dir="ltr" style="text-align: left;" trbidi="on"><span style="font-size: large;"><b><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; text-align: -webkit-auto;">Data-driven Matchmaking at Scale</span></b></span><br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;"><br />
</span><br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">Note start time is 12:15pm, because of the 11am talk by Andrew Ng; see </span><span style="color: #222222; font-family: arial, sans-serif;"><span style="font-size: 14px;"><a href="http://www.cs.ucsd.edu/node/2096">http://www.cs.ucsd.edu/node/2096</a></span></span><br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;"><br />
</span><br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">Abstract: </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">Nearly 5% of all US marriages are created by eHarmony. I will talk about the </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">tech that stands behind this and how eHarmony is different from a typical </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">dating site. I will describe the three main components of eHarmony's approach. </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">First I will discuss the models for predicting deep psychological </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">compatibility. I will then show how we use large scale machine learning to </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">learn models of affinity based on user behavior, demographics, interests etc </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">and show some insights into what makes a match more likely to succeed. Finally </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">I will demonstrate how we use graph optimization to choose the best matches to deliver </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">every single day.</span><br style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;" /><br style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;" /><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">eHarmony iPad app demo:</span><br style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;" /><a href="http://www.youtube.com/watch?v=zQE-ILMmqDs" style="background-color: rgba(255, 255, 255, 0.917969); color: #1155cc; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;" target="_blank">http://www.youtube.com/watch?<wbr></wbr>v=zQE-ILMmqDs</a><br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;"><br />
</span><br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">Bio: </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">Vaclav Petricek is a Principal Data Scientist at Santa Monica-based</span><br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">eHarmony where he is responsible for optimization and machine learning</span><br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">for eHarmony's core matchmaking algorithms. He also runs a series of</span><br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">invited ML talks at eHarmony, part of the Los Angeles Machine Learning Meetup.</span><br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">Prior to eHarmony, Vaclav was Visiting Researcher at University College, London</span><br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">where his research spanned recommender systems, social networks, web structure</span><br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">and online auctions. Prior to that he has worked at several Czech internet startups.</span><br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">Vaclav earned his PhD in Computer Science and a Masters in Distributed Systems</span><br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">from Charles University in Prague.</span><br />
<br />
<a href="http://www.occamslab.com/petricek/">http://www.occamslab.com/petricek/</a></div>Charles Elkanhttp://www.blogger.com/profile/10639994516997092663noreply@blogger.com0tag:blogger.com,1999:blog-1431146078025501017.post-80787154223859837262011-11-27T17:39:00.003-08:002012-01-15T10:00:23.243-08:001/23/12: Lars Kai Hansen (Danish Technical University)<div dir="ltr" style="text-align: left;" trbidi="on"><b><span style="font-size: large;"><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; text-align: -webkit-auto;">Learning from small samples in high dimensions</span></span></b><br />
<span class="il" style="background-attachment: initial; background-clip: initial; background-color: rgba(255, 255, 255, 0.917969); background-image: initial; background-origin: initial; color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;"><br />
</span><br />
<span class="il" style="background-attachment: initial; background-clip: initial; background-color: rgba(255, 255, 255, 0.917969); background-image: initial; background-origin: initial; color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">Abstract</span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">: </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">I will discuss recent progress in coping with variance inflation in </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">high-dimensional unsupervised learning (PCA and kPCA). </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">Small sample high-dimensional principal component analysis (PCA) </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">suffers from variance inflation and lack of generalizability. It has </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">earlier been pointed out that a simple leave-one-out variance </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">renormalization scheme can cure the problem. We have generalized the </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">cure in two directions: First, we propose a computationally less </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">intensive approximate leave-one-out estimator, secondly, we show that </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">variance inflation is also present in kernel principal component </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">analysis (kPCA) and we provide a non-parametric renormalization scheme </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">which can quite efficiently restore generalizability in kPCA. As for </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">PCA our analysis also suggests a simplified approximate </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">expression. Finally, I present evidence that these ideas may be </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">relevant also for supervised high-dimensional supervised learning </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">with support vector machines.</span><br />
<br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">Reference: </span><i><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">A Cure for Variance Inflation in High Dimensional Kernel Principal </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">Component Analysis</span></i><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;"> by T. J. Abrahamsen and L.K. </span><span class="il" style="background-attachment: initial; background-clip: initial; background-color: rgba(255, 255, 255, 0.917969); background-image: initial; background-origin: initial; color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">Hansen, </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">Journal of Machine Learning Research </span><b style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">12</b><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">:2027-2044 (2011).</span><br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;"><br />
</span><br />
<span class="il" style="background-attachment: initial; background-clip: initial; background-color: rgba(255, 255, 255, 0.917969); background-image: initial; background-origin: initial; color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">Bio: Professor Lars</span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;"> Kai </span><span class="il" style="background-attachment: initial; background-clip: initial; background-color: rgba(255, 255, 255, 0.917969); background-image: initial; background-origin: initial; color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">Hansen is the d</span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">irector of the THOR Center for Neuroinformatics and the </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">Head of the Section for Cognitive Systems at </span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">DTU Informatics at the Technical University of Denmark.</span><br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;"><br />
</span><br />
<a href="http://www.imm.dtu.dk/~lkh" style="background-color: rgba(255, 255, 255, 0.917969); color: #1155cc; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;" target="_blank">http://www.imm.dtu.dk/~lkh</a></div>Charles Elkanhttp://www.blogger.com/profile/10639994516997092663noreply@blogger.com0tag:blogger.com,1999:blog-1431146078025501017.post-9837226326516782062011-11-27T17:36:00.001-08:002012-01-03T15:59:31.682-08:001/16/12: No seminar--MLK day<div dir="ltr" style="text-align: left;" trbidi="on"><br />
</div>Charles Elkanhttp://www.blogger.com/profile/10639994516997092663noreply@blogger.com0tag:blogger.com,1999:blog-1431146078025501017.post-9367619944961511712011-11-27T17:32:00.000-08:002012-01-15T09:51:21.502-08:001/9/12: Dhruv Batra: Focused Inference in Markov Random Fields with Local Primal-Dual Gaps (Toyota Technological Institute at Chicago)<div dir="ltr" style="text-align: left;" trbidi="on"><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: large; text-align: -webkit-auto;"><b>Focused Inference in Markov Random Fields with Local Primal-Dual Gaps</b></span><br />
<br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">A large number of problems in computer vision, computational biology and robotics can formulated as the search for the most probable state under a discrete probabilistic model -- known as the MAP inference problem in Markov Random Fields (MRFs). </span><br />
<br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">While a lot of progress has been made on the "static" version of this problem, a number of situations require dynamic inference algorithms that must adapt and reorder computation to focus on "important" parts of the problem. In this talk I will describe one measure for identifying such important parts of the problem -- called Local Primal Dual Gaps (LPDG). LPDG is based on complementary slackness conditions in the Primal-Dual pair of Linear Programs (LP) in the LP relaxation of MAP inference. We have found LPDG to be useful in a number of situations -- speeding-up message-passing algorithms by re-ordering message computations (Tarlow et al. ICML '11), speeding up alpha-expansion by re-ordering label sweeps (Batra & Kohli CVPR '11) and adaptive tightening of the standard LP relaxation by choosing important constraints to add (Batra et al. AISTATS '11). </span><br />
<br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">Time permitting, I will also talk about our recent work on the M-Best-Mode problem, which involves extracting not just the most probable solution, but also a /diverse/ set of top M most probable solutions in discrete graphical models. </span><br />
<br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">The talk is meant to be accessible to a broad audience. No background in MRFs or discrete optimization is assumed. </span><br />
<br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">Joint work with Pushmeet Kohli (MSRC), Vladimir Kolmogorov (IST), Sebastian Nowozin (MSRC), Greg Shakhnarovich (TTIC), Daniel Tarlow (UToronto) and Payman Yadollahpour (TTIC).</span><br />
<br />
<br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;"><b>Bio: </b></span><span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">Dhruv Batra is a Research Assistant Professor at Toyota Technological Institute at Chicago (TTIC), a philanthropically endowed academic computer science institute affiliated with the University of Chicago. He received his M.S. and Ph.D. degrees from Carnegie Mellon University in 2007 and 2010 respectively, advised by Tsuhan Chen. In the past, he has held visiting positions at Cornell University and MIT. </span><br />
<br />
<span style="background-color: rgba(255, 255, 255, 0.917969); color: #222222; font-family: arial, sans-serif; font-size: 13px; text-align: -webkit-auto;">His research interests include machine learning, computer vision and applications of combinatorial optimization algorithms to learning and vision tasks. Specifically, he is interested in structured prediction, MAP inference in MRFs, max-margin methods, co-segmentation in multiple images, and interactive 3D modelling. </span><br />
<a href="http://ttic.uchicago.edu/~dbatra/">http://ttic.uchicago.edu/~dbatra/</a></div>Charles Elkanhttp://www.blogger.com/profile/10639994516997092663noreply@blogger.com