<html><head><meta http-equiv="content-type" content="text/html; charset=utf-8"></head><body dir="auto"><div dir="ltr">If anyone is interested in joining a workshop on network inference, this would be IDEAL!</div><div dir="ltr"><br></div><div dir="ltr">Best,</div><div dir="ltr">Madhur <br><br>Begin forwarded message:<br><br></div><blockquote type="cite"><div dir="ltr"><b>From:</b> Aravindan Vijayaraghavan <aravindv@northwestern.edu><br><b>Date:</b> June 27, 2020 at 11:08:32 AM CDT<br><b>To:</b> Lev Reyzin <lreyzin@uic.edu>, "Perkins, William Frank Cox" <willp@uic.edu>, Madhur Tulsiani <madhurt@ttic.edu><br><b>Cc:</b> Varun Gupta <guptav@uchicago.edu><br><b>Subject:</b> <b>IDEAL Workshop: Computational vs Statistical Tradeoffs in Network Inference [June 29, 11-4 Central]</b><br><br></div></blockquote><blockquote type="cite"><div dir="ltr">
<meta http-equiv="Content-Type" content="text/html; charset=Windows-1252">
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
Hi Lev, Madhur, Will, </div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<span style="color: rgb(0, 0, 0); font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt;">I hope you are all doing well and keeping safe in these strange times. </span><span style="color: rgb(0, 0, 0); font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt;">Sorry
for the late notice. </span><br>
</div>
<div>
<div dir="ltr">
<div style="font-family:Calibri,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
<br>
</div>
<div>
<div dir="ltr">
<div dir="ltr">
<div style="font-family:Calibri,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
<div style="margin:0px; font-size:15px; background-color:rgb(255,255,255)"><span style="font-family:Arial,Helvetica,sans-serif">We have our final </span><span style="margin:0px; font-family:Arial,Helvetica,sans-serif">workshop</span><span style="font-family:Arial,Helvetica,sans-serif"> </span><span style="font-family:Arial,Helvetica,sans-serif">as
part of the IDEAL special quarter on <i>Inference and Data Science on Networks </i>
on <span style="margin:0px; font-weight:normal; font-size:16px; font-family:Arial,Helvetica,sans-serif; background-color:rgb(255,255,255); text-align:left; color:rgb(0,124,137); text-decoration:underline">
<a href="https://www.ideal.northwestern.edu/events/workshop-computational-vs-statistical-tradeoffs-in-network-inference/" target="_blank" rel="noopener noreferrer" style="margin:0px; font-weight:normal; font-size:16px; font-family:Helvetica; background-color:rgb(255,255,255); text-align:left; color:rgb(0,124,137); text-decoration:underline">Computational
vs Statistical Tradeoffs in Network Inference</a></span>. The workshop will take place on Monday, June 29th. There will be </span><span style="color:rgb(32,32,32); font-family:Arial,Helvetica,sans-serif; font-size:16px; text-align:left; background-color:rgb(255,255,255); display:inline!important">talks
by Andrea Montanari, Ankur Moitra and Liza Levina from 11am-3:15pm Central Time (CT) and a panel discussion with the speakers from 3:25-4:00 pm CT. </span><span style="font-family:Arial,Helvetica,sans-serif; font-size:15px">Please see the details below or
on </span><a href="https://www.ideal.northwestern.edu/events/workshop-computational-vs-statistical-tradeoffs-in-network-inference/" title="https://www.ideal.northwestern.edu/events/workshop-computational-vs-statistical-tradeoffs-in-network-inference/" style="font-size:15px"><span style="font-family:Arial,Helvetica,sans-serif">this</span></a><span style="font-family:Arial,Helvetica,sans-serif; font-size:15px">
page</span><span style="font-family:Arial,Helvetica,sans-serif; font-size:15px">. </span></div>
<div style="margin:0px; font-size:15px; background-color:rgb(255,255,255)"><span style="font-size:15px"><br>
</span></div>
<div style="margin:0px; font-size:15px; background-color:rgb(255,255,255)"><span style="font-family:Arial,Helvetica,sans-serif; font-size:15px">It would be great if you could forward this announcement to your group and others who may be interested in it. </span></div>
<div style="margin:0px; font-size:15px; background-color:rgb(255,255,255)"><br>
</div>
<div style="margin:0px; font-size:15px; background-color:rgb(255,255,255)"><span style="font-family:Arial,Helvetica,sans-serif">Best,</span></div>
<div style="margin:0px; font-size:15px; background-color:rgb(255,255,255)"><span style="font-family:Arial,Helvetica,sans-serif">Aravindan (on behalf of the w</span><span style="font-family:Arial,Helvetica,sans-serif; font-size:15px">orkshop</span><span style="font-family:Arial,Helvetica,sans-serif; font-size:15px"> o</span><span style="font-family:Arial,Helvetica,sans-serif; font-size:15px">rganizers)</span></div>
<div style="margin:0px; font-size:15px; background-color:rgb(255,255,255)"><br>
</div>
</div>
<div>
<div id="x_x_x_divRplyFwdMsg" dir="ltr">
<div> </div>
</div>
<div dir="ltr">
<div style="font-family:Calibri,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
</div>
<div>
<div dir="ltr">
<div style="font-family:Calibri,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
<br>
</div>
<div style="font-family:Calibri,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
<table width="100%" style="border-collapse:collapse; min-width:100%">
<tbody>
<tr>
<td style="padding-top:9px"><br>
<br>
---------------------------------------------------------------------------------------------------<br>
<br>
<table align="left" width="100%" style="border-collapse:collapse; max-width:100%; min-width:100%">
<tbody>
<tr>
<td style="word-break:break-word; color:rgb(32,32,32); font-family:Helvetica; line-height:24px; text-align:left; padding:0px 18px 9px">
<h1 style="display:block; margin:0px; font-size:26px; font-weight:bold; line-height:32.5px; text-align:center">
Upcoming IDEAL Workshop: <br>
Workshop: Computational vs Statistical Tradeoffs in Network Inference</h1>
<p style="margin-top: 0px; margin-bottom: 0px;margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin:10px 0px">
The final workshop of the quarter is on<span> </span><a href="https://www.ideal.northwestern.edu/events/workshop-computational-vs-statistical-tradeoffs-in-network-inference/" target="_blank" rel="noopener noreferrer" style="margin:0px; font-weight:normal; color:rgb(0,124,137); text-decoration:underline">Computational
vs Statistical Tradeoffs in Network Inference</a><span> </span>takes place on June 29 with talks from 11am-3:15pm Central and a panel discussion from 3:25-4:00 pm. Participants can register to join on Zoom, or live-stream the video (details below). We are
looking forward to seeing everyone there!</p>
<h4 style="display:block; margin:0px; font-size:18px; font-weight:bold; line-height:22.5px">
<strong>About the Series</strong></h4>
<p style="margin-top: 0px; margin-bottom: 0px;margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin:10px 0px">
The IDEAL workshop series brings in experts on topics related to the foundations of data science to present their perspective and research on a common theme. Different components of the workshop will also allow for continued discussion and interaction between
attendees and the speakers.</p>
<h4 style="display:block; margin:0px; font-size:18px; font-weight:bold; line-height:22.5px">
<br>
<strong>Synopsis</strong></h4>
<p style="margin-top: 0px; margin-bottom: 0px;margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin:10px 0px">
Network models have been used as a tool to understand the role of interconnections between entities, by diverse communities such as sociology, biology, meteorology, economics, and computer science, to name a few. Moreover emerging technological developments
allow collecting data on increasingly larger networks. This leads to both computational and statistical challenges when inferring or learning the structure of such networks. This workshop will cover some of the advances in the last decade on understanding
trade-offs between statistical and computational efficiency for many inference problems on large networks. The workshop speakers are Andrea Montanari, Ankur Moitra, and Liza Levina.</p>
<h4 style="display:block; margin:0px; font-size:18px; font-weight:bold; line-height:22.5px">
<br>
<strong>Logistics</strong></h4>
<ul>
<li><strong>Date: </strong>Monday, June 29, 2020.</li><li><strong>Location:</strong> Zoom participation (register below). Panopto<span> </span><a href="https://www.ideal.northwestern.edu/events/workshop-computational-vs-statistical-tradeoffs-in-network-inference/streaming" target="_blank" rel="noopener noreferrer" style="margin:0px; font-weight:normal; color:rgb(0,124,137); text-decoration:underline">live-streaming</a>.</li><li><strong>Registration:</strong> <a href="https://www.ideal.northwestern.edu/events/workshop-computational-vs-statistical-tradeoffs-in-network-inference/registration" target="_blank" rel="noopener noreferrer" style="margin:0px; font-weight:normal; color:rgb(0,124,137); text-decoration:underline">Registration
form</a>. Registered participants will get a Zoom link to the workshop by email.</li></ul>
<h4 style="display:block; margin:0px; font-size:18px; font-weight:bold; line-height:22.5px">
<br>
<strong>Schedule</strong></h4>
<ul>
<li><strong>11:00-11:05</strong>: Opening Remarks</li><li><strong>11:05-11:50</strong>: Andrea Montanari, (Stanford University, EE)<br style="font-family:Helvetica; background-color:rgb(255,255,255)">
<span style="font-family:Helvetica; background-color:rgb(255,255,255); display:inline!important">Optimization of mean field spin glasses.</span><br>
</li><li><strong>11:50-1:45</strong>: Lunch Break</li><li><strong>1:45-2:30</strong>: Ankur Moitra, (MIT, Mathematics)<br>
Learning with Massart Noise</li><li><strong>2:30-3:15</strong>: Liza Levina, (University of Michigan, Statistics)<br>
Hierarchical community detection by recursive partitioning</li><li><strong>3:25-4:00</strong>: Panel Discussion with the Speakers</li></ul>
<p style="margin-top: 0px; margin-bottom: 0px;margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin:10px 0px">
We will also have an online social mixer on Gather.town during the break. </p>
<p style="margin-top: 0px; margin-bottom: 0px;margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin:10px 0px">
<span style="color: rgb(32, 32, 32); font-family: Helvetica;"> </span></p>
<h4 style="display:block; margin:0px; font-size:18px; font-weight:bold; line-height:22.5px">
<strong>Titles and Abstracts</strong></h4>
<h4 style="display:block; margin:0px; font-size:18px; font-weight:bold; line-height:22.5px">
<br>
</h4>
<h4 style="display:block; margin:0px; font-size:18px; font-weight:bold; line-height:22.5px">
<div style="margin:10px 0px; font-weight:400; font-size:16px; font-family:Helvetica; background-color:rgb(255,255,255)">
<strong style="background-color:rgb(255,255,255)"><br>
Speaker:</strong><span style="margin:0px; background-color:rgb(255,255,255); display:inline!important"> Andrea Montanari, Stanford</span><br style="background-color:rgb(255,255,255)">
<strong style="background-color:rgb(255,255,255)">Title and abstract:</strong><span style="margin:0px; background-color:rgb(255,255,255); display:inline!important"> Optimization of mean field spin glasses.<br>
<b>Abstract:</b><span> </span>Let G be a random graph over N vertices generated according to a statistical model, such as the stochastic block model. In order to analyze such a graph, it would be interesting to be able to partition its vertices into k balanced
sets as to minimize the number of edges across the partition. Even for k=2, this problem is hard to approximate. Although we know of semidefinite programming relaxations with optimal worst-case guarantees, these can be sub-optimal for random instances. I will
consider the case k=2 when G is an Erd\”os-Renyi random graph. This problem is equivalent to the one of finding the ground-state of the Sherrington-Kirkpatrick Hamiltonian in statistical physics.<br>
<br>
I describe an algorithm that achieves a (1-\eps) approximation of the latter problem in near-linear time, conditional on a certain conjecture in mathematical physics. The algorithms exploits certain subtle properties of the non-convex energy landscape of this
model. I will then discuss a more general class of Hamiltonians for mean field spin glasses and put forward a precise picture of the average-case tractability of optimization in this context.<br>
[Based on arXiv:1812.10897 and arXiv:2001.00904 (with Ahmed El Alaoui and Mark Sellke)]</span><br style="background-color:rgb(255,255,255)">
<br>
</div>
<div style="margin:10px 0px; font-weight:400; font-size:16px; font-family:Helvetica; background-color:rgb(255,255,255)">
<hr style="background-color:rgb(255,255,255)">
<div style="margin:10px 0px; background-color:rgb(255,255,255)"><strong><br>
</strong></div>
<div style="margin:10px 0px; background-color:rgb(255,255,255)"><strong style="font-size:inherit; font-style:inherit; font-variant-ligatures:inherit; font-variant-caps:inherit">Speaker:</strong><span style="color:rgb(32,32,32); font-family:Helvetica"> Ankur
Moitra, MIT</span><br>
</div>
</div>
</h4>
<p style="margin-top: 0px; margin-bottom: 0px;margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin:10px 0px">
<strong>Title:</strong> Learning with Massart Noise<br>
<strong>Abstract: </strong>In supervised learning, there is often a tension between being robust against increasingly more powerful adversaries and computational complexity. We will be interested in the Massart noise model, where one interpretation is that
an adversary can arbitrarily control the labels on a random subset of the data. This seems to be a comfortable middle ground where algorithms cannot overtune to the noise distribution, but it is nevertheless possible to give algorithms with strong provable
guarantees.</p>
<p style="margin-top: 0px; margin-bottom: 0px;margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin:10px 0px">
We give a new game theoretic framework for learning under Massart noise, and our algorithms are derived from low regret online algorithms for playing relaxations of this game. Along the way we give an algorithm for properly learning halfspaces under Massart
noise, improving upon the recent work of Diakonikolas, Goulekakis and Tzamos that gives an improper learning algorithm, as well as extensions to generalized linear models. Finally we evaluate our algorithms empirically and find that they exhibit some appealing
fairness properties, perhaps as a byproduct of their robustness guarantees.</p>
<br>
This is based on joint work with Sitan Chen, Frederic Koehler and Morris Yau.<br>
<br>
<hr>
<div style="margin:10px 0px"><strong style="color:inherit; font-family:inherit; font-size:inherit; font-style:inherit; font-variant-ligatures:inherit; font-variant-caps:inherit"><br>
</strong></div>
<div style="margin:10px 0px"><strong style="color:inherit; font-family:inherit; font-size:inherit; font-style:inherit; font-variant-ligatures:inherit; font-variant-caps:inherit">Speaker:</strong><span style="font-family:Helvetica"> Liza Levina, Michigan</span><br>
</div>
<div style="margin:10px 0px"><strong>Title:</strong> Hierarchical community detection by recursive partitioning<br>
<strong>Abstract: </strong>Community detection in networks has been extensively studied in the form of finding a single partition into a “correct” number of communities. In large networks, however, a multi-scale hierarchy of communities is much more realistic.
We show that a hierarchical tree of communities, obviously more interpretable, is also potentially more accurate and more computationally efficient. We construct this tree with a simple top-down recursive algorithm, at each step splitting the nodes into two
communities with a non-iterative spectral algorithm, until a stopping rule suggests there are no more communities. The algorithm is model-free, extremely fast, and requires no tuning other than selecting a stopping rule. We propose a natural model for this
setting, a binary tree stochastic block model, and prove that the algorithm correctly recovers the entire community tree under relatively mild assumptions. As a by-product, we obtain explicit and intuitive results for fitting the stochastic block model under
model misspecification. We illustrate the algorithm on a statistics papers dataset constructing a highly interpretable tree of statistics research communities, and on a network based on gene co-occurrence in research papers on anemia. <br>
<br>
This is joint work with Tianxi Li, Lihua Lei, Sharmodeep Bhattacharyya, Koen van de Berge, Purnamrita Sarkar, and Peter Bickel.</div>
<div style="margin:10px 0px">
<div style="margin:10px 0px; font-family:Helvetica; background-color:rgb(255,255,255)">
<span style="margin:0px; background-color:rgb(255,255,255); display:inline!important"><br>
<hr style="font-family:Helvetica; background-color:rgb(255,255,255)">
<div style="margin:10px 0px; font-family:Helvetica; background-color:rgb(255,255,255)">
<br>
</div>
</span></div>
<br>
</div>
</td>
</tr>
</tbody>
</table>
</td>
</tr>
</tbody>
</table>
<table width="100%" style="border-collapse:collapse; min-width:100%">
<tbody>
<tr>
<td style="padding:9px">
<table align="left" width="100%" style="border-collapse:collapse; min-width:100%">
<tbody>
<tr>
<td style="padding:0px 9px; text-align:center"><img align="center" alt="" width="562.75" style="margin:0px; vertical-align:bottom; height:auto; outline:none; max-width:1200px; display:inline!important" data-unique-identifier=""></td>
</tr>
</tbody>
</table>
</td>
</tr>
</tbody>
</table>
<table width="100%" style="border-collapse:collapse; min-width:100%">
<tbody>
<tr>
<td style="padding-top:9px">
<table align="left" width="100%" style="border-collapse:collapse; max-width:100%; min-width:100%">
<tbody>
<tr>
<td style="word-break:break-word; color:rgb(101,101,101); font-family:Helvetica; font-size:12px; line-height:18px; text-align:center; padding:0px 18px 9px">
<br>
<em> <br>
Copyright © 2020 IDEAL, Northwestern University, All rights reserved.</em><br>
You are receiving this email because you opted in via our website.<br>
<br>
<strong>Our mailing address is:</strong><br>
<div style="margin:0px"><span style="margin:0px">IDEAL, Northwestern University</span>
<div style="margin:0px">
<div style="margin:0px">2233 Tech Drive</div>
<span style="margin:0px">Evanston</span>,<span> </span><span style="margin:0px">IL</span><span> </span><span style="margin:0px">60208</span></div>
<br>
<a href="https://urldefense.com/v3/__https://northwestern.us20.list-manage.com/vcard?u=3fc8e0df393510ea0a5b018e1&id=c68ca2c43c__;!!Dq0X2DkFhyF93HkjWTBQKhk!FjWiUjVVB17ifbD4IAG0uRdcYVg_6iTMAARJ6nNTIYMKeDcFkyX_DlvfxF7vM9uyk56FlA$" target="_blank" rel="noopener noreferrer" style="margin:0px; font-weight:normal; text-decoration:underline">Add
us to your address book</a></div>
<br>
<br>
Want to change how you receive these emails?<br>
You can<span> </span><a href="https://urldefense.com/v3/__https://northwestern.us20.list-manage.com/profile?u=3fc8e0df393510ea0a5b018e1&id=c68ca2c43c&e=__;!!Dq0X2DkFhyF93HkjWTBQKhk!FjWiUjVVB17ifbD4IAG0uRdcYVg_6iTMAARJ6nNTIYMKeDcFkyX_DlvfxF7vM9vz9kGd7g$" target="_blank" rel="noopener noreferrer" style="margin:0px; font-weight:normal; text-decoration:underline">update
your preferences</a><span> </span>or<span> </span><a href="https://urldefense.com/v3/__https://northwestern.us20.list-manage.com/unsubscribe?u=3fc8e0df393510ea0a5b018e1&id=c68ca2c43c&e=&c=1a79085507__;!!Dq0X2DkFhyF93HkjWTBQKhk!FjWiUjVVB17ifbD4IAG0uRdcYVg_6iTMAARJ6nNTIYMKeDcFkyX_DlvfxF7vM9unixhMeA$" target="_blank" rel="noopener noreferrer" style="margin:0px; font-weight:normal; text-decoration:underline">unsubscribe
from this list</a>.<br>
<br>
<a href="https://urldefense.com/v3/__http://www.mailchimp.com/email-referral/?utm_source=freemium_newsletter&utm_medium=email&utm_campaign=referral_marketing&aid=3fc8e0df393510ea0a5b018e1&afl=1__;!!Dq0X2DkFhyF93HkjWTBQKhk!FjWiUjVVB17ifbD4IAG0uRdcYVg_6iTMAARJ6nNTIYMKeDcFkyX_DlvfxF7vM9vi6BnSXg$" target="_blank" rel="noopener noreferrer" style="margin:0px; font-weight:normal; text-decoration:underline"><img alt="Email Marketing Powered by Mailchimp" title="Mailchimp Email Marketing" width="139" height="54" style="margin:0px; height:auto!important; outline:none" data-unique-identifier=""></a></td>
</tr>
</tbody>
</table>
</td>
</tr>
</tbody>
</table>
<table width="100%" style="border-collapse:collapse; table-layout:fixed!important; min-width:100%">
<tbody>
<tr>
<td style="min-width:100%; padding:10px 18px 25px">
<table width="100%" style="border-collapse:collapse; min-width:100%; border-top:2px solid rgb(238,238,238)">
<tbody>
<tr>
<td><span style="margin:0px"></span></td>
</tr>
</tbody>
</table>
</td>
</tr>
</tbody>
</table>
<br>
</div>
<div id="x_x_x_x_x_appendonsend"></div>
<hr tabindex="-1" style="display:inline-block; width:98%">
<div id="x_x_x_x_x_divRplyFwdMsg" dir="ltr"><font face="Calibri, sans-serif" color="#000000" style="font-size:11pt"><b>From:</b> Jason Hartline <jasonhartline@gmail.com> on behalf of Jason Hartline <hartline@eecs.northwestern.edu><br>
<b>Sent:</b> Wednesday, June 24, 2020 8:35 AM<br>
<b>To:</b> Kristi Hubbard <kristi.hubbard@northwestern.edu><br>
<b>Cc:</b> Aravindan Vijayaraghavan <aravindv@northwestern.edu>; Pamela Marie Villalovoz <pmv@northwestern.edu><br>
<b>Subject:</b> Fwd: [IDEAL] Workshop: Computational vs Statistical Tradeoffs in Network Inference [June 29, 11-4 Central]</font>
<div> </div>
</div>
<div class="" style="word-wrap:break-word; line-break:after-white-space">Dear Kristi,
<div class=""><br class="">
</div>
<div class="">I think your network may find this workshop on networks interesting. </div>
<div class=""><br class="">
</div>
<div class="">
<blockquote type="cite" class=""><span class="" style="color:rgb(32,32,32); font-size:16px; background-color:rgb(255,255,255)">The Spring 2020 Special Quarter on Inference and Data Science on Networks is wrapping up with a final workshop. The final workshop
of the quarter is on</span> <a href="https://urldefense.com/v3/__https://northwestern.us20.list-manage.com/track/click?u=3fc8e0df393510ea0a5b018e1&id=bee5522cbc&e=cc6e98141b__;!!Dq0X2DkFhyF93HkjWTBQKhk!GSO9DfJNZHBoPACsaggNOr5ToawtFBv83CebCEW0btJjUphvIYZNNiUW33HA37TRK9n6fQ$" target="_blank" class="" style="font-size:16px; color:rgb(0,124,137)">Computational
vs Statistical Tradeoffs in Network Inference</a> <span class="" style="color:rgb(32,32,32); font-size:16px; background-color:rgb(255,255,255)">takes place on June 29 with short talks from 11am-3:15pm Central and a panel discussion from 3:25-4:00 pm. Participants
can register to join on Zoom, or stream the video on Panopto. We are looking forward to seeing everyone there!</span><br class="">
</blockquote>
</div>
<div class=""><br class="">
</div>
<div class="">Jason<br class="">
<div><br class="">
<br class="">
<div class="">
<center class="" style="font-family:Helvetica; font-size:12px; font-style:normal; font-variant-caps:normal; font-weight:normal; letter-spacing:normal; text-indent:0px; text-transform:none; white-space:normal; word-spacing:0px; background-color:rgb(250,250,250); text-decoration:none">
<table align="center" border="0" cellpadding="0" cellspacing="0" height="100%" width="100%" id="x_x_x_x_x_x_bodyTable" class="">
<tbody class="">
<tr class="">
<td align="center" valign="top" id="x_x_x_x_x_x_bodyCell" class="" style="height:3174px; margin:0px; padding:10px; width:973px; border-top-width:0px">
<table border="0" cellpadding="0" cellspacing="0" width="100%" class="x_x_x_x_x_x_templateContainer" style="border-collapse:collapse; border:1px groove rgb(78,42,132); max-width:600px!important; width:600px!important">
<tbody class="">
<tr class="">
<td valign="top" id="x_x_x_x_x_x_templatePreheader" class="" style="background-color:rgb(250,250,250); background-image:none; background-size:cover; border-top-width:0px; border-bottom-width:0px; padding-top:9px; padding-bottom:9px; background-position:center center; background-repeat:no-repeat no-repeat">
</td>
</tr>
<tr class="">
<td valign="top" id="x_x_x_x_x_x_templateHeader" class="" style="background-color:rgb(255,255,255); background-image:none; background-size:cover; border-top-width:0px; border-bottom-width:0px; padding-top:9px; padding-bottom:0px; background-position:center center; background-repeat:no-repeat no-repeat">
</td>
</tr>
<tr class="">
<td valign="top" id="x_x_x_x_x_x_templateBody" class="" style="background-color:rgb(255,255,255); background-image:none; background-size:cover; border-top-width:0px; border-bottom-width:2px; border-bottom-style:solid; border-bottom-color:rgb(234,234,234); padding-top:0px; padding-bottom:9px; background-position:center center; background-repeat:no-repeat no-repeat">
<table border="0" cellpadding="0" cellspacing="0" width="100%" class="x_x_x_x_x_x_mcnTextBlock" style="border-collapse:collapse; min-width:100%">
<tbody class="x_x_x_x_x_x_mcnTextBlockOuter">
<tr class="">
<td valign="top" class="x_x_x_x_x_x_mcnTextBlockInner" style="padding-top:9px">
<table align="left" border="0" cellpadding="0" cellspacing="0" width="100%" class="x_x_x_x_x_x_mcnTextContentContainer" style="border-collapse:collapse; max-width:100%; min-width:100%">
<tbody class="">
<tr class="">
<td valign="top" class="x_x_x_x_x_x_mcnTextContent" style="word-break:break-word; color:rgb(32,32,32); font-family:Helvetica; font-size:16px; line-height:24px; text-align:left; padding:0px 18px 9px">
<h1 class="" style="display:block; margin:0px; padding:0px; color:rgb(32,32,32); font-family:Helvetica; font-size:26px; font-style:normal; font-weight:bold; line-height:32.5px; letter-spacing:normal; text-align:center">
Upcoming IDEAL Workshop: <br class="">
Workshop: Computational vs Statistical Tradeoffs in Network Inference</h1>
<p class="" style="margin-top: 0px; margin-bottom: 0px;margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin:10px 0px; padding:0px; color:rgb(32,32,32); font-family:Helvetica; font-size:16px; line-height:24px; text-align:left">
The Spring 2020 Special Quarter on Inference and Data Science on Networks is underway. Courses and workshops are all being offered in virtual format. The final workshop of the quarter is on<span class="x_x_x_x_x_x_Apple-converted-space"> </span><a href="https://urldefense.com/v3/__https://northwestern.us20.list-manage.com/track/click?u=3fc8e0df393510ea0a5b018e1&id=bee5522cbc&e=cc6e98141b__;!!Dq0X2DkFhyF93HkjWTBQKhk!GSO9DfJNZHBoPACsaggNOr5ToawtFBv83CebCEW0btJjUphvIYZNNiUW33HA37TRK9n6fQ$" target="_blank" class="" style="color:rgb(0,124,137); font-weight:normal; text-decoration:underline">Computational
vs Statistical Tradeoffs in Network Inference</a><span class="x_x_x_x_x_x_Apple-converted-space"> </span>takes place on June 29 with short talks from 11am-3:15pm Central and a panel discussion from 3:25-4:00 pm. Participants can register to join on Zoom,
or stream the video on Panopto (details below). We are looking forward to seeing everyone there!<br class="">
</p>
<h4 class="" style="display:block; margin:0px; padding:0px; color:rgb(32,32,32); font-family:Helvetica; font-size:18px; font-style:normal; font-weight:bold; line-height:22.5px; letter-spacing:normal; text-align:left">
<strong class="">About the Series</strong></h4>
<p class="" style="margin-top: 0px; margin-bottom: 0px;margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin:10px 0px; padding:0px; color:rgb(32,32,32); font-family:Helvetica; font-size:16px; line-height:24px; text-align:left">
The IDEAL workshop series brings in four experts on topics related to the foundations of data science to present their perspective and research on a common theme. Chicago area researchers with an interest in the foundations of data science. The technical program
is in the morning and includes coffee and lunch. The afternoon of the workshop will allow for continued discussion between attendees and the speakers.</p>
<p class="" style="margin-top: 0px; margin-bottom: 0px;margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin:10px 0px; padding:0px; color:rgb(32,32,32); font-family:Helvetica; font-size:16px; line-height:24px; text-align:left">
Part of the <a href="https://urldefense.com/v3/__https://northwestern.us20.list-manage.com/track/click?u=3fc8e0df393510ea0a5b018e1&id=6cafbb5baf&e=cc6e98141b__;!!Dq0X2DkFhyF93HkjWTBQKhk!GSO9DfJNZHBoPACsaggNOr5ToawtFBv83CebCEW0btJjUphvIYZNNiUW33HA37SiU4yvKA$" class="" style="color:rgb(0,124,137); font-weight:normal; text-decoration:underline">Special
Quarter on Inference and Data Science on Networks</a>.</p>
<h4 class="" style="display:block; margin:0px; padding:0px; color:rgb(32,32,32); font-family:Helvetica; font-size:18px; font-style:normal; font-weight:bold; line-height:22.5px; letter-spacing:normal; text-align:left">
<br class="">
<strong class="">Synopsis</strong></h4>
<p class="" style="margin-top: 0px; margin-bottom: 0px;margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin:10px 0px; padding:0px; color:rgb(32,32,32); font-family:Helvetica; font-size:16px; line-height:24px; text-align:left">
Network models have been used as a tool to understand the role of interconnections between entities, by diverse communities such as sociology, biology, meteorology, economics, and computer science, to name a few. Moreover emerging technological developments
allow collecting data on increasingly larger networks. This leads to both computational and statistical challenges when inferring or learning the structure of such networks. This workshop will cover some of the advances in the last decade on understanding
trade-offs between statistical and computational efficiency for many inference problems on large networks. The workshop speakers are Andrea Montanari, Ankur Moitra, and Liza Levina.</p>
<h4 class="" style="display:block; margin:0px; padding:0px; color:rgb(32,32,32); font-family:Helvetica; font-size:18px; font-style:normal; font-weight:bold; line-height:22.5px; letter-spacing:normal; text-align:left">
<br class="">
<strong class="">Logistics</strong></h4>
<ul class="">
<li class=""><strong class="">Date: </strong>Monday, June 29, 2020.</li><li class=""><strong class="">Location:</strong> Zoom participation (register below). Panopto<span class="x_x_x_x_x_x_Apple-converted-space"> </span><a href="https://urldefense.com/v3/__https://northwestern.us20.list-manage.com/track/click?u=3fc8e0df393510ea0a5b018e1&id=a18e77c3e3&e=cc6e98141b__;!!Dq0X2DkFhyF93HkjWTBQKhk!GSO9DfJNZHBoPACsaggNOr5ToawtFBv83CebCEW0btJjUphvIYZNNiUW33HA37Q7zJHvBQ$" target="_blank" class="" style="color:rgb(0,124,137); font-weight:normal; text-decoration:underline">streaming</a>.</li><li class=""><strong class="">Registration:</strong> <a href="https://urldefense.com/v3/__https://northwestern.us20.list-manage.com/track/click?u=3fc8e0df393510ea0a5b018e1&id=b0a7d30d2b&e=cc6e98141b__;!!Dq0X2DkFhyF93HkjWTBQKhk!GSO9DfJNZHBoPACsaggNOr5ToawtFBv83CebCEW0btJjUphvIYZNNiUW33HA37TDJn27SA$" target="_blank" class="" style="color:rgb(0,124,137); font-weight:normal; text-decoration:underline">Registration
form</a>. Registered participants will get a Zoom link to the workshop by email.</li></ul>
<h4 class="" style="display:block; margin:0px; padding:0px; color:rgb(32,32,32); font-family:Helvetica; font-size:18px; font-style:normal; font-weight:bold; line-height:22.5px; letter-spacing:normal; text-align:left">
<br class="">
<strong class="">Schedule</strong></h4>
<ul class="">
<li class=""><strong class="">11:00-11:05</strong>: Opening Remarks</li><li class=""><strong class="">11:05-11:50</strong>: Andrea Montanari, (Stanford University, EE)</li><li class=""><strong class="">11:50-1:45</strong>: Lunch Break</li><li class=""><strong class="">1:45-2:30</strong>: Ankur Moitra, (MIT, Mathematics)<br class="">
Learning with Massart Noise</li><li class=""><strong class="">2:30-3:15</strong>: Liza Levina, (University of Michigan, Statistics)<br class="">
Hierarchical community detection by recursive partitioning</li><li class=""><strong class="">3:25-4:00</strong>: Panel Discussion with the Speakers</li></ul>
<p class="" style="margin-top: 0px; margin-bottom: 0px;margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin:10px 0px; padding:0px; color:rgb(32,32,32); font-family:Helvetica; font-size:16px; line-height:24px; text-align:left">
Please use this<span class="x_x_x_x_x_x_Apple-converted-space"> </span><a href="https://urldefense.com/v3/__https://northwestern.us20.list-manage.com/track/click?u=3fc8e0df393510ea0a5b018e1&id=95b76cf8aa&e=cc6e98141b__;!!Dq0X2DkFhyF93HkjWTBQKhk!GSO9DfJNZHBoPACsaggNOr5ToawtFBv83CebCEW0btJjUphvIYZNNiUW33HA37Q5Rvnx2Q$" target="_blank" class="" style="color:rgb(0,124,137); font-weight:normal; text-decoration:underline">Google
form</a> to pose questions for the Panel and the speakers.<br class="">
</p>
<h4 class="" style="display:block; margin:0px; padding:0px; color:rgb(32,32,32); font-family:Helvetica; font-size:18px; font-style:normal; font-weight:bold; line-height:22.5px; letter-spacing:normal; text-align:left">
<strong class="">Titles and Abstracts</strong><br class="">
</h4>
<p class="" style="margin-top: 0px; margin-bottom: 0px;margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin:10px 0px; padding:0px; color:rgb(32,32,32); font-family:Helvetica; font-size:16px; line-height:24px; text-align:left">
<strong class="">Speaker:</strong> Andrea Montanari, Stanford<br class="">
<strong class="">Title:</strong> <em class="">TBA.</em><br class="">
<strong class="">Abstract:</strong><br class="">
TBA.</p>
<hr class="">
<p class="" style="margin-top: 0px; margin-bottom: 0px;margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin:10px 0px; padding:0px; color:rgb(32,32,32); font-family:Helvetica; font-size:16px; line-height:24px; text-align:left">
<strong class="">Speaker:</strong> Ankur Moitra, MIT<br class="">
<strong class="">Title:</strong> Learning with Massart Noise<br class="">
<strong class="">Abstract: </strong>In supervised learning, there is often a tension between being robust against increasingly more powerful adversaries and computational complexity. We will be interested in the Massart noise model, where one interpretation
is that an adversary can arbitrarily control the labels on a random subset of the data. This seems to be a comfortable middle ground where algorithms cannot overtune to the noise distribution, but it is nevertheless possible to give algorithms with strong
provable guarantees.</p>
<p class="" style="margin-top: 0px; margin-bottom: 0px;margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin:10px 0px; padding:0px; color:rgb(32,32,32); font-family:Helvetica; font-size:16px; line-height:24px; text-align:left">
We give a new game theoretic framework for learning under Massart noise, and our algorithms are derived from low regret online algorithms for playing relaxations of this game. Along the way we give an algorithm for properly learning halfspaces under Massart
noise, improving upon the recent work of Diakonikolas, Goulekakis and Tzamos that gives an improper learning algorithm, as well as extensions to generalized linear models. Finally we evaluate our algorithms empirically and find that they exhibit some appealing
fairness properties, perhaps as a byproduct of their robustness guarantees.</p>
<br class="">
This is based on joint work with Sitan Chen, Frederic Koehler and Morris Yau.<br class="">
<br class="">
<hr class="">
<p class="" style="margin-top: 0px; margin-bottom: 0px;margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin-top:0px; margin-bottom:0px; margin:10px 0px; padding:0px; color:rgb(32,32,32); font-family:Helvetica; font-size:16px; line-height:24px; text-align:left">
<br class="">
<strong class="">Speaker:</strong> Liza Levina, Michigan<br class="">
<strong class="">Title:</strong> Hierarchical community detection by recursive partitioning<br class="">
<strong class="">Abstract: </strong>Community detection in networks has been extensively studied in the form of finding a single partition into a “correct” number of communities. In large networks, however, a multi-scale hierarchy of communities is much more
realistic. We show that a hierarchical tree of communities, obviously more interpretable, is also potentially more accurate and more computationally efficient. We construct this tree with a simple top-down recursive algorithm, at each step splitting the nodes
into two communities with a non-iterative spectral algorithm, until a stopping rule suggests there are no more communities. The algorithm is model-free, extremely fast, and requires no tuning other than selecting a stopping rule. We propose a natural model
for this setting, a binary tree stochastic block model, and prove that the algorithm correctly recovers the entire community tree under relatively mild assumptions. As a by-product, we obtain explicit and intuitive results for fitting the stochastic block
model under model misspecification. We illustrate the algorithm on a statistics papers dataset constructing a highly interpretable tree of statistics research communities, and on a network based on gene co-occurrence in research papers on anemia. <br class="">
<br class="">
This is joint work with Tianxi Li, Lihua Lei, Sharmodeep Bhattacharyya, Koen van de Berge, Purnamrita Sarkar, and Peter Bickel.</p>
</td>
</tr>
</tbody>
</table>
</td>
</tr>
</tbody>
</table>
<table border="0" cellpadding="0" cellspacing="0" width="100%" class="x_x_x_x_x_x_mcnImageBlock" style="border-collapse:collapse; min-width:100%">
<tbody class="x_x_x_x_x_x_mcnImageBlockOuter">
<tr class="">
<td valign="top" class="x_x_x_x_x_x_mcnImageBlockInner" style="padding:9px">
<table align="left" width="100%" border="0" cellpadding="0" cellspacing="0" class="x_x_x_x_x_x_mcnImageContentContainer" style="border-collapse:collapse; min-width:100%">
<tbody class="">
<tr class="">
<td class="x_x_x_x_x_x_mcnImageContent" valign="top" style="padding:0px 9px; text-align:center">
<img align="center" alt="" width="562.75" class="x_x_x_x_x_x_mcnImage" style="border:0px; height:auto; outline:none; text-decoration:none; vertical-align:bottom; max-width:1200px; padding-bottom:0px; display:inline!important" src="https://mcusercontent.com/3fc8e0df393510ea0a5b018e1/images/bf59dc2c-9d1f-491e-9177-baf3fd2b8765.png" data-unique-identifier=""></td>
</tr>
</tbody>
</table>
</td>
</tr>
</tbody>
</table>
</td>
</tr>
<tr class="">
<td valign="top" id="x_x_x_x_x_x_templateFooter" class="" style="background-color:rgb(250,250,250); background-image:none; background-size:cover; border-top-width:0px; border-bottom-width:0px; padding-top:9px; padding-bottom:9px; background-position:center center; background-repeat:no-repeat no-repeat">
<table border="0" cellpadding="0" cellspacing="0" width="100%" class="x_x_x_x_x_x_mcnTextBlock" style="border-collapse:collapse; min-width:100%">
<tbody class="x_x_x_x_x_x_mcnTextBlockOuter">
<tr class="">
<td valign="top" class="x_x_x_x_x_x_mcnTextBlockInner" style="padding-top:9px">
<table align="left" border="0" cellpadding="0" cellspacing="0" width="100%" class="x_x_x_x_x_x_mcnTextContentContainer" style="border-collapse:collapse; max-width:100%; min-width:100%">
<tbody class="">
<tr class="">
<td valign="top" class="x_x_x_x_x_x_mcnTextContent" style="word-break:break-word; color:rgb(101,101,101); font-family:Helvetica; font-size:12px; line-height:18px; text-align:center; padding:0px 18px 9px">
<br class="">
<em class=""> <br class="">
Copyright © 2020 IDEAL, Northwestern University, All rights reserved.</em><br class="">
You are receiving this email because you opted in via our website.<br class="">
<br class="">
<strong class="">Our mailing address is:</strong><br class="">
<div class="x_x_x_x_x_x_vcard"><span class="x_x_x_x_x_x_org x_x_x_x_x_x_fn">IDEAL, Northwestern University</span>
<div class="x_x_x_x_x_x_adr">
<div class="x_x_x_x_x_x_street-address">2233 Tech Drive</div>
<span class="x_x_x_x_x_x_locality">Evanston</span>,<span class="x_x_x_x_x_x_Apple-converted-space"> </span><span class="x_x_x_x_x_x_region">IL</span><span class="x_x_x_x_x_x_Apple-converted-space"> </span>
<span class="x_x_x_x_x_x_postal-code">60208</span></div>
<br class="">
<a href="https://urldefense.com/v3/__https://northwestern.us20.list-manage.com/vcard?u=3fc8e0df393510ea0a5b018e1&id=c68ca2c43c__;!!Dq0X2DkFhyF93HkjWTBQKhk!GSO9DfJNZHBoPACsaggNOr5ToawtFBv83CebCEW0btJjUphvIYZNNiUW33HA37Ss8EqVUw$" class="x_x_x_x_x_x_hcard-download" style="color:rgb(101,101,101); font-weight:normal; text-decoration:underline">Add
us to your address book</a></div>
<br class="">
<br class="">
Want to change how you receive these emails?<br class="">
You can<span class="x_x_x_x_x_x_Apple-converted-space"> </span><a href="https://urldefense.com/v3/__https://northwestern.us20.list-manage.com/profile?u=3fc8e0df393510ea0a5b018e1&id=c68ca2c43c&e=cc6e98141b__;!!Dq0X2DkFhyF93HkjWTBQKhk!GSO9DfJNZHBoPACsaggNOr5ToawtFBv83CebCEW0btJjUphvIYZNNiUW33HA37QbbKfOag$" class="" style="color:rgb(101,101,101); font-weight:normal; text-decoration:underline">update
your preferences</a><span class="x_x_x_x_x_x_Apple-converted-space"> </span>or<span class="x_x_x_x_x_x_Apple-converted-space"> </span><a href="https://urldefense.com/v3/__https://northwestern.us20.list-manage.com/unsubscribe?u=3fc8e0df393510ea0a5b018e1&id=c68ca2c43c&e=cc6e98141b&c=c5a80af88d__;!!Dq0X2DkFhyF93HkjWTBQKhk!GSO9DfJNZHBoPACsaggNOr5ToawtFBv83CebCEW0btJjUphvIYZNNiUW33HA37Tg93DBCA$" class="" style="color:rgb(101,101,101); font-weight:normal; text-decoration:underline">unsubscribe
from this list</a>.<br class="">
<br class="">
<a href="https://urldefense.com/v3/__http://www.mailchimp.com/email-referral/?utm_source=freemium_newsletter&utm_medium=email&utm_campaign=referral_marketing&aid=3fc8e0df393510ea0a5b018e1&afl=1__;!!Dq0X2DkFhyF93HkjWTBQKhk!GSO9DfJNZHBoPACsaggNOr5ToawtFBv83CebCEW0btJjUphvIYZNNiUW33HA37TymaW5fg$" class="" style="color:rgb(101,101,101); font-weight:normal; text-decoration:underline"><img border="0" alt="Email Marketing Powered by Mailchimp" title="Mailchimp Email Marketing" width="139" height="54" class="" style="border:0px; height:auto!important; outline:none; text-decoration:none" src="https://cdn-images.mailchimp.com/monkey_rewards/MC_MonkeyReward_15.png" data-unique-identifier=""></a></td>
</tr>
</tbody>
</table>
</td>
</tr>
</tbody>
</table>
<table border="0" cellpadding="0" cellspacing="0" width="100%" class="x_x_x_x_x_x_mcnDividerBlock" style="border-collapse:collapse; table-layout:fixed!important; min-width:100%">
<tbody class="x_x_x_x_x_x_mcnDividerBlockOuter">
<tr class="">
<td class="x_x_x_x_x_x_mcnDividerBlockInner" style="min-width:100%; padding:10px 18px 25px">
<table class="x_x_x_x_x_x_mcnDividerContent" border="0" cellpadding="0" cellspacing="0" width="100%" style="border-collapse:collapse; min-width:100%; border-top-width:2px; border-top-style:solid; border-top-color:rgb(238,238,238)">
<tbody class="">
<tr class="">
<td class=""><span class=""></span></td>
</tr>
</tbody>
</table>
</td>
</tr>
</tbody>
</table>
</td>
</tr>
</tbody>
</table>
</td>
</tr>
</tbody>
</table>
</center>
<img height="1" width="1" class="" style="border:0px; height:auto; outline:none; text-decoration:none; font-family:Helvetica; font-size:12px; font-style:normal; font-variant-caps:normal; font-weight:normal; letter-spacing:normal; text-align:start; text-indent:0px; text-transform:none; white-space:normal; word-spacing:0px; background-color:rgb(250,250,250)" src="https://northwestern.us20.list-manage.com/track/open.php?u=3fc8e0df393510ea0a5b018e1&id=c5a80af88d&e=cc6e98141b" data-unique-identifier=""><span class="" style="font-family:Helvetica; font-size:12px; font-style:normal; font-variant-caps:normal; font-weight:normal; letter-spacing:normal; text-align:start; text-indent:0px; text-transform:none; white-space:normal; word-spacing:0px; background-color:rgb(250,250,250); text-decoration:none; float:none; display:inline!important"></span></div>
</div>
<br>
<br class="">
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div></blockquote></body></html>