<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=utf-8">
</head>
<body bgcolor="#FFFFFF" text="#000000">
Dear Autonians,<br>
<br>
If you are available please come and cheer Ina on her path to
completion.<br>
<br>
Her thesis defense will take place this Monday at 10am in Gates Hall
room 6115.<br>
<br>
See you there!<br>
Artur<br>
<div class="moz-forward-container"><br>
<br>
-------- Forwarded Message --------
<table class="moz-email-headers-table" border="0" cellpadding="0"
cellspacing="0">
<tbody>
<tr>
<th align="RIGHT" nowrap="nowrap" valign="BASELINE">Subject:
</th>
<td>Reminder - Thesis Defense - 8/3/15 - Ina Fiterau -
Discovering Compact and Informative Structures through
Data Partitioning</td>
</tr>
<tr>
<th align="RIGHT" nowrap="nowrap" valign="BASELINE">Date: </th>
<td>Sun, 02 Aug 2015 15:48:43 -0400</td>
</tr>
<tr>
<th align="RIGHT" nowrap="nowrap" valign="BASELINE">From: </th>
<td>Diane Stidle <a class="moz-txt-link-rfc2396E" href="mailto:diane@cs.cmu.edu"><diane@cs.cmu.edu></a></td>
</tr>
<tr>
<th align="RIGHT" nowrap="nowrap" valign="BASELINE">To: </th>
<td><a class="moz-txt-link-abbreviated" href="mailto:ML-SEMINAR@cs.cmu.edu">ML-SEMINAR@cs.cmu.edu</a>, Andreas Krause
<a class="moz-txt-link-rfc2396E" href="mailto:krausea@ethz.ch"><krausea@ethz.ch></a></td>
</tr>
</tbody>
</table>
<br>
<br>
<meta http-equiv="content-type" content="text/html; charset=utf-8">
Thesis Defense<br>
<br>
Date: 8/3/15<br>
Time: 10:00am<br>
Place: 6115 GHC<br>
PhD Candidate: Madalina Fiterau-Brostean<br>
<br>
Title: Discovering Compact and Informative Structures through Data
Partitioning<br>
<br>
<div>Abstract: </div>
<div>In this thesis, we have shown that it is possible to identify
low-dimensional structures in complex high-dimensional data, if
such structures exist. We have leveraged these underlying
structures to construct compact interpretable models for various
machine learning tasks<span style="text-align:justify"> that
benefit practical applications.</span></div>
<div>
<p class="MsoNormal"
style="text-align:justify;text-justify:inter-ideograph">To
start with, I will formalize Informative Projection Recovery,
the problem of extracting a small set of low-dimensional
projections of data that jointly support an accurate model for
a given learning task. Our solution to this problem is a
regression-based algorithm that identifies informative
projections by optimizing over a matrix of point-wise loss
estimators. It generalizes to multiple types of machine
learning problems, offering solutions to classification,
clustering, regression, and active learning tasks. Experiments
show that our method can discover and leverage low-dimensional
structures in data, yielding accurate and compact models. Our
method is particularly useful in applications in which expert
assessment of the results is of the essence, such as
classification tasks in the healthcare domain.</p>
<p class="MsoNormal"
style="text-align:justify;text-justify:inter-ideograph">In the
second part of the talk, I will describe back-propagation
forests, a new type of ensemble that achieves improved
accuracy over existing ensemble classifiers such as random
forests classifiers or alternating decision forests.
Back-propagation (BP) trees use soft splits, such that a
sample is probabilistically assigned to all the leaves. Also,
the leaves assign a distribution across the labels. The
splitting parameters are obtained through SGD by optimizing
the log loss over the entire tree, which is a non-convex
objective. The probability distribution over the leaves is
computed exactly by maximizing a log concave procedure. In
addition, I will present several proposed approaches for the
use of BP forests within the context of compact informative
structure discovery. We have successfully used BP forests to
increase the performance of deep belief network architectures,
with results improving over the state of the art on vision
datasets.<br>
</p>
<p class="MsoNormal"
style="text-align:justify;text-justify:inter-ideograph">Thesis
Committee:<br>
Artur Dubrawski, Chair<br>
Geoff Gordon<br>
Alex Smola<br>
Andreas Krause (ETH Zurich)<br>
</p>
</div>
<pre class="moz-signature" cols="72">--
Diane Stidle
Graduate Programs Manager
Machine Learning Department
Carnegie Mellon University
<a moz-do-not-send="true" class="moz-txt-link-abbreviated" href="mailto:diane@cs.cmu.edu">diane@cs.cmu.edu</a>
412-268-1299</pre>
<br>
</div>
<br>
</body>
</html>