<html>
  <head>

    <meta http-equiv="content-type" content="text/html; charset=utf-8">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    <div dir="ltr">
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"
        id="gmail-docs-internal-guid-27f3a02f-d801-f3f4-c53a-fad0dbbb235e"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Please
          forward to all interested parties/mailing lists.</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
                              </span><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">   
        </span><span style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">New
          York Area Theory Day</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
                           </span><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">   
        </span><span style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Organized
          by: IBM/NYU/Columbia</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
                              </span><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">   
        </span><span style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Friday,
          December 1, 2017</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">The
          Theory Day will be held at Courant Institute of Mathematical
          Sciences,</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">New
          York University, 251 Mercer Street, Auditorium 109, New York.</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Directions:
          <a href="http://www.cims.nyu.edu/direct.html">http://www.cims.nyu.edu/direct.html</a></span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Apart
          from invited distinguished speakers, we also have short talk</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">sessions
          in which junior researchers will present some of their
          research.</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
                                       </span><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">   
        </span><span style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">PROGRAM</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
          9:30 - 10:00 Coffee and bagels</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">10:00
          - 10:55 Prof. Christos Papadimitriou</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
                   </span><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">   
        </span><span style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
           A Computer Scientist Thinks about the Brain</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">10:55
          - 11:15 Coffee break</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">11:15
          - 12:05 Short talks:</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
                       Shay Solomon, Ilya Razenshteyn, Sepideh Mahabadi</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">12:05
          - 2:00 Lunch break</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
          2:00 - 2:55 Dr. Ilan Komargodski</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
                   </span><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">   
        </span><span style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
           White-Box vs. Black-Box Complexity of Search Problems:</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
                      </span><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">   
        </span><span style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Ramsey
          and Graph Property Testing</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
          2:55 - 3:15 Short talk: Sandro Coretti</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">   
        </span><span style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
        </span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
          3:15 - 3:35 Coffee break</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
          3:35 - 3:55 Short talk: Greg Bodwin</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
          3:55 - 4:50 Prof. Virginia Vassilevska Williams</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
                  </span><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">   
        </span><span style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
          Fine-Grained Complexity of Problems in P</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
          5:00 - 7:00 Follow-up social</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">For
          directions, please see <a
            href="http://www.cims.nyu.edu/direct.html">http://www.cims.nyu.edu/direct.html</a></span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">To
          subscribe to our mailing list, follow instructions at</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
           <a href="http://www.cs.nyu.edu/mailman/listinfo/theory-ny">http://www.cs.nyu.edu/mailman/listinfo/theory-ny</a></span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Organizers:
          Alex Andoni (<a href="mailto:andoni@cs.columbia.edu">andoni@cs.columbia.edu</a>);</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
                 </span><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">   
        </span><span style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Yevgeniy
          Dodis (<a href="mailto:dodis@cs.nyu.edu">dodis@cs.nyu.edu</a>);</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
                 </span><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">   
        </span><span style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Krzysztof
          Onak (<a href="mailto:konak@us.ibm.com">konak@us.ibm.com</a>)</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">================================================================================</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
                                    </span><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">   
        </span><span style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">LONG
          TALKS                                  </span><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">   
        </span><span style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
        </span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">================================================================================</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Christos
          H. Papadimitriou (Columbia University)</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Title:
          A Computer Scientist Thinks about the Brain</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Abstract:
          Understanding the ways in which the Brain gives rise to the
          Mind</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">(memory,
          behavior, cognition, intelligence, language) is the most
          challenging</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">problem
          facing science today. As the answer seems likely to be at
          least partly</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">computational,
          computer scientists should work on this problem. Using a</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">simplified
          model of Hebbian plasticity and employing random graph theory
          and</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Bernoulli
          approximations we reproduce a form of association between
          memories</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">recently
          observed experimentally in humans. We discuss several
          implications and</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">extensions.
          (Joint work with W. Maass and S. Vempala).</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">================================================================================</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Ilan
          Komargodski (Cornell Tech)</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Title:
          White-Box vs. Black-Box Complexity of Search Problems: Ramsey
          and Graph</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
            </span><span style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">   
        </span><span style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Property
          Testing</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Abstract:
          Ramsey theory assures us that in any graph there is a clique
          or</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">independent
          set of a certain size, roughly logarithmic in the graph size.
          But</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">how
          difficult is it to find the clique or independent set? This
          problem is in</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">TFNP,
          the class of search problems with guaranteed solutions.  If
          the graph is</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">given
          explicitly, then it is possible to do so while examining a
          linear number</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">of
          edges. If the graph is given by a black-box, where to figure
          out whether a</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">certain
          edge exists the box should be queried, then a large number of
          queries</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">must
          be issued.</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">1)
          What if one is given a program or circuit ("white-box") for
          computing the</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
            existence of an edge. Does the search problem remain hard?</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">2)
          Can we generically translate all TFNP black-box hardness into
          white-box</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
            hardness?</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">3)
          Does the problem remain hard if the black-box instance is
          small?</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">We
          will answer all of these questions and discuss related
          questions in the</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">setting
          of property testing.</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Joint
          work with Moni Naor and Eylon Yogev.</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">================================================================================</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Virginia
          Vassilevska Williams (MIT)</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Title:
          Fine-Grained Complexity of Problems in P</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Abstract:
          A central goal of algorithmic research is to determine how
          fast</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">computational
          problems can be solved in the worst case. Theorems from
          complexity</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">theory
          state that there are problems that, on inputs of size n, can
          be solved in</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">t(n)
          time but not in t(n)^{1-eps} time for eps&gt;0. The main
          challenge is to</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">determine
          where in this hierarchy various natural and important problems
          lie.</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Throughout
          the years, many ingenious algorithmic techniques have been
          developed</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">and
          applied to obtain blazingly fast algorithms for many problems.
          Nevertheless,</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">for
          many other central problems, the best known running times are
          essentially</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">those
          of their classical algorithms from the 1950s and 1960s.</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Unconditional
          lower bounds seem very difficult to obtain, and so practically
          all</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">known
          time lower bounds are conditional. For years, the main tool
          for proving</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">hardness
          of computational problems have been NP-hardness reductions,
          basing</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">hardness
          on P\neq NP. However, when we care about the exact running
          time (as</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">opposed
          to merely polynomial vs non-polynomial), NP-hardness is not
          applicable,</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">especially
          if the problem is already solvable in polynomial time. In
          recent</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">years,
          a new theory has been developed, based on ``fine-grained
          reductions''</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">that
          focus on exact running times. Mimicking NP-hardness, the
          approach is to (1)</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">select
          a key problem X that is conjectured to require essentially
          t(n) time for</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">some
          t, and (2) reduce X in a fine-grained way to many important
          problems. This</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">approach
          has led to the discovery of many meaningful relationships
          between</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">problems,
          and even sometimes to equivalence classes.</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">The
          main key problems used to base hardness on have been: the 3SUM
          problem, the</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">CNF-SAT
          problem (based on the Strong Exponential TIme Hypothesis
          (SETH)) and the</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">All
          Pairs Shortest Paths Problem. Research on SETH-based lower
          bounds has</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">flourished
          in particular in recent years showing that the classical
          algorithms</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">are
          optimal for problems such as Approximate Diameter, Edit
          Distance, Frechet</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Distance,
          Longest Common Subsequence etc.</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">In
          this talk I will give an overview of the current progress in
          this area of</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">study,
          and will highlight some exciting new developments.</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">================================================================================</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
                                    </span><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">   
        </span><span style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">SHORT
          TALKS                                  </span><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">   
        </span><span style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">
        </span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">================================================================================</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Shay
          Solomon (IBM Research)</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Title:
          Fully Dynamic Maximal Matching</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Abstract:
          Graph matching is one of the most well-studied problems in</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">combinatorial
          optimization, with a reach plethora of applications.
          Nevertheless,</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">until
          recently little was known about this problem in real-life
          dynamic</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">networks,
          which aim to model the constantly changing physical world.</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Focusing
          on the maximal matching problem, there is a trivial linear
          time</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">algorithm
          for computing such a matching in static networks, which
          naturally</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">translates
          into an algorithm with a constant amortized cost in the
          incremental</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">case.
          In this talk I will discuss some of the challenges that we
          overcame to</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">obtain
          a constant amortized cost algorithm in the fully dynamic case.</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">================================================================================</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Ilya
          Razenshteyn (Columbia)</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Title:
          Practical Data-Dependent Metric Compression with Provable
          Guarantees</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Abstract:
          How well can one compress a dataset of points from a
          high-dimensional</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">space
          while preserving pairwise distances? Indyk and Wagner have
          recently</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">obtained
          almost optimal bounds for this problem, but their construction
          (based</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">on
          hierarchical clustering) is not practical. In this talk, I
          will show a new</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">practical,
          quadtree-based compression scheme, whose provable performance</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">essentially
          matches that of the result of Indyk and Wagner.</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">In
          addition to the theoretical results, we will see experimental
          comparison of</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">the
          new scheme and Product Quantization (PQ)--one of the most
          popular heuristics</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">for
          distance-preserving compression--on several datasets. Unlike
          PQ and other</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">heuristics
          that rely on the clusterability of the dataset, the new
          algorithm</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">ends
          up being more robust.</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">The
          talk is based on a joint work with Piotr Indyk and Tal Wagner.</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">================================================================================</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Sepideh
          Mahabadi (Columbia)</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Title:
          Set Cover in Sub-linear Time</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Abstract:
          We study the classic set cover problem from the perspective of</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">sub-linear
          algorithms. Given access to a collection of $m$ sets over $n$</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">elements
          in the query model, we show that sub-linear algorithms derived
          from</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">existing
          techniques have almost tight query complexities.</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">On
          one hand, first we show that an adaptation of streaming
          algorithms to the</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">sub-linear
          query model returns an $\alpha$-approximate cover using
          $\tilde</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">O(m(n/k)^{1/(\alpha-1)}
          + nk)$ queries to the input, where $k$ denotes the value</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">of
          a minimum set cover. We then complement this upper bound by
          proving that for</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">lower
          values of $k$, the required number of queries is $\tilde</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Omega(m(n/k)^{1/(2\alpha)})$,
          even for estimating the optimal cover size.</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Moreover,
          we prove that even checking whether a given collection of sets
          covers</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">all
          the elements would require $\Omega(nk)$ queries. These two
          lower bounds</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">provide
          strong evidence that the upper bound is almost tight for
          certain values</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">of
          the parameter $k$. On the other hand, we show that this bound
          is not optimal</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">for
          larger values of the parameter $k$, as there exists a</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">$(1+\eps)$-approximation
          algorithm with $\tilde O(mn/k\eps^2)$ queries. We show</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">that
          this bound is essentially tight for sufficiently small
          constant $\eps$, by</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">establishing
          a lower bound of $\tilde Omega(mn/k)$ query complexity.</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">This
          is a joint work with Piotr Indyk, Ronitt Rubinfeld, Ali
          Vakilian, and Anak</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Yodpinyanee.</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">================================================================================</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Sandro
          Coretti (New York University)</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Title:
          Random Oracles and Non-Uniformity</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Abstract:
          Hash functions are ubiquitous in cryptography. They are widely
          used in</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">practice
          to build one-way functions (OWFs), collision-resistant hash
          functions</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">(CRHFs),
          pseudorandom functions/generators (PRFs/PRGs), message
          authentication</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">codes
          (MACs), etc. Moreover, they are often used together with other</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">computational
          assumptions to prove the security of higher-level
          applications,</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">such
          as Fiat-Shamir heuristics for signature schemes,
          full-domain-hash</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">signatures,
          or OAEP encryption, among many others.</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">The
          security of such schemes is commonly analyzed in the
          random-oracle model</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">(ROM),
          in which the hash function is modeled as a truly random
          function that can</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">be
          queried by an attacker at a bounded number of points. The
          traditional ROM,</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">however,
          does not account for the fact that a non-uniform attacker may
          obtain</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">arbitrary
          advice about the hash function H before attacking a
          construction. As a</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">result,
          bounds derived in the ROM do not accurately reflect security
          against</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">such
          attackers.</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">To
          remedy this situation, Unruh (CRYPTO ’07) considered the
          auxiliary-input ROM</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">(AI-ROM),
          in which an attacker may obtain (a bounded amount of)
          arbitrary</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">leakage
          about the random oracle before attacking a cryptographic
          scheme. In this</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">talk
          we discuss significant improvements to Unruh's presampling
          technique,</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">which,
          intuitively, allows a leaky random oracle to be replaced by
          one that is</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">adversarially
          fixed on a subset of the coordinates but uniformly random</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">otherwise.
          We also show how this improved presampling leads to better and
          easily</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">derivable
          AI-ROM security bounds for a number of cryptographic
          applications. We</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">conclude
          by noting that in several cases (such as OWFs, PRGs, and
          CRHFs) there</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">are
          still significant gaps between the derived bounds and the best
          attacks,</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">leaving
          interesting open problems.</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">================================================================================</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Greg
          Bodwin (MIT)</span></p>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Title:
          Compressing Graphs While Preserving Their Distances</span></p>
      <br>
      <p dir="ltr"
        style="line-height:1.38;margin-top:0pt;margin-bottom:0pt"><span
          style="font-size:10pt;font-family:&quot;Courier
New&quot;;color:rgb(0,0,0);background-color:transparent;font-weight:400;font-style:normal;font-variant:normal;text-decoration:none;vertical-align:baseline">Abstract:
          TBA</span></p>
      <br>
    </div>
  </body>
</html>