<div dir="ltr"><div class="gmail_quote"><div dir="ltr">I hope everyone is having a quiet and productive summer!  Manoj Prabhakaran from UIUC will be visiting us this week, to give a seminar about a unified theory of cryptographic agents.  Seminar will be 10am on Wed July 9, with lunch following.  <br>



<br>Hope to see many of you there,<br><div><br>Sharon<p> <span>BUsec</span> Calendar:  <a href="http://www.bu.edu/cs/busec/" target="_blank">http://www.bu.edu/cs/<span>busec</span>/</a><br>
 <span>BUsec</span> Mailing list: <a href="http://cs-mailman.bu.edu/mailman/listinfo/busec" target="_blank">http://cs-mailman.bu.edu/mailman/listinfo/<span>busec</span></a><br>

 How to get to BU from MIT: The CT2 bus or MIT&#39;s &quot;Boston Daytime Shuttle&quot; <a href="http://web.mit.edu/facilities/transportation/shuttles/daytime_boston.html" target="_blank">http://web.mit.edu/facilities/transportation/shuttles/daytime_boston.html</a></p>



<br><div class="gmail_quote"><div bgcolor="#FFFFFF" text="#000000"><div><div dir="ltr"><div>
          <div>Title: Towards a Unified Theory of Cryptographic Agents</div><div>Speaker: Manoj Prabhakaran, UIUC.<br></div><div>July 9, 2014, 10-11:30 am<br></div><div>MCS137, at 111 Cummington St, Boston MA <br></div><div>



<br>
          </div>
          <div>Abstract:</div>
          <div><br>
          </div>
          <div>In recent years there has been a fantastic boom of
            increasingly</div>
          <div>sophisticated &quot;cryptographic objects&quot; --- Identity-Based
            Encryption,</div>
          <div>Fully-Homomorphic Encryption, Functional Encryption,
            various forms of</div>
          <div>obfuscation, Witness Encryption, Property-Preserving
            Encryption, etc.</div>
          <div>As these constructions have grown in number, variety,
            complexity and</div>
          <div>inter-connectedness, the relationships among them have
            become</div>
          <div>increasingly confusing.  </div>
          <div><br>
          </div>
          <div>We provide a new framework of {\em cryptographic agents}
            that unifies various cryptographic objects and security definitions,
            similar to how the Universal Composition framework unifies various
            multi-party</div>
          <div>computation tasks like commitment, coin-tossing and
            zero-knowledge</div>
          <div>proofs. Various primitives are modeled by different
            &quot;schemas&quot;</div>
          <div>(analogous to functionalities in the UC framework).</div>
          <div><br>
          </div>
          <div>We use a new {\em indistinguishability preserving}
            (INDPRE) definition</div>
          <div>of security, that often side-steps the impossibility
            results</div>
          <div>associated with simulation-based definitions, and yet
            admits a</div>
          <div>composition theorem. Also, when appropriately restricted,
            our</div>
          <div>definitions often yield existing definitions of the
            corresponding</div>
          <div>primitives.</div>
          <div><br>
          </div>
          <div>Interestingly, our framework can also be used to model
            abstractions</div>
          <div>like the generic group model and the random oracle model,
            letting one</div>
          <div>translate a general class of constructions in these
            heuristic models</div>
          <div>to constructions based on {\em standard model
            assumptions}.  </div>
          <div><br>
          </div>
          <div>We illustrate the new framework by presenting new
            constructions of</div>
          <div>functional encryption (FE) schemes, with and without
            function-hiding,</div>
          <div>in terms of schematic reductions to the obfuscation
            schema or the</div>
          <div>bilinear generic group schema. When combined with
            candidate schemes for these schemas using our composition theorem, these
            constructions yield concrete FE schemes.<br><br><div>(Joint work with Shweta Agrawal and Shashank Agrawal)</div>
          <br><br></div></div></div></div></div></div></div></div>
</div><br><br clear="all"><br>-- <br>Sharon Goldberg<br>Computer Science, Boston University<br><a href="http://www.cs.bu.edu/~goldbe" target="_blank">http://www.cs.bu.edu/~goldbe</a>
</div>