<html><head><meta http-equiv="Content-Type" content="text/html charset=utf-8"><meta http-equiv="Content-Type" content="text/html charset=utf-8"><meta http-equiv="Content-Type" content="text/html charset=utf-8"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;" class=""><div class="">This is an announcement of Bowen Wang’s MS Presentation.</div><div class=""><br class=""></div><div class=""><br class=""></div><div class="">===========================</div><div class="">Department of Computer Science</div><div class="">The University of Chicago</div><div class=""><br class=""></div><div class="">Date: Thursday, July 12, 2018</div><div class=""><br class=""></div><div class="">Time: 10:00am</div><div class=""><br class=""></div><div class="">Location: Ryerson 255</div><div class=""><br class=""></div><div class="">Bx/MS Candidate: Bowen Wang</div><div class=""><br class=""></div><div class="">MS Advisor: Ravi Chugh</div><div class=""><br class=""></div><div class="">MS Paper Title: Predicting Haskell Type Signatures From Names</div><div class=""><br class=""></div><div class="">Abstract: </div><div class=""><br class=""></div><div class="">Neural Program Synthesis has been a fast-growing field with many exciting advances such as translating natural language into code. However, those applications often suffer from the lack of high-quality data, as it is fairly difficult to obtain code annotated with natural language that transcribes its functionality precisely. Therefore, we want to understand what information we can extract from source code to facilitate program synthesis without explicit supervision. To ground our task more concretely, we study the problem of predicting type signatures given identifier names. We focus on Haskell, a strongly typed functional language, and collect data from Haskell repositories on GitHub. We use two different approaches: unstructured pre- diction, which is based on the sequence-to-sequence model, and structured prediction, which models the tree structure of the type signatures directly. Surprisingly, even though the struc- tured prediction model is better at predicting the structures of the type signatures (61.66% structural accuracy), the unstructured prediction model achieves much better signature accuracy (23.71%). </div><div class=""><br class=""></div><div class="">A draft copy of Bowen’s MS paper will be available soon.</div><div class=""><br class=""></div><div apple-content-edited="true" class="">
<span class="Apple-style-span" style="border-collapse: separate; border-spacing: 0px;"><div style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;" class=""><div style="border-collapse: separate; color: rgb(0, 0, 0); font-variant: normal; letter-spacing: normal; line-height: normal; text-transform: none; white-space: normal; word-spacing: 0px; border-spacing: 0px; -webkit-text-stroke-width: 0px; margin: 0px;" class=""><font face="Garamond" size="2" class=""><div style="border-collapse: separate; border-spacing: 0px; margin: 0px;" class="">Karin Czaplewski</div><div style="border-collapse: separate; border-spacing: 0px; margin: 0px;" class="">Student & Faculty Services Specialist</div><div style="border-collapse: separate; border-spacing: 0px; margin: 0px;" class="">Masters Program in Computer Science</div><div style="border-collapse: separate; border-spacing: 0px; margin: 0px;" class="">The University of Chicago</div><div style="border-collapse: separate; border-spacing: 0px; margin: 0px;" class="">5555 S. Ellis Avenue - Room 305</div><div style="border-collapse: separate; border-spacing: 0px; margin: 0px;" class="">(773) 834-8587</div><div style="border-collapse: separate; border-spacing: 0px; margin: 0px;" class="">Email: <a href="mailto:karin@cs.uchicago.edu" class="">karin@cs.uchicago.edu</a></div><p style="margin: 0px;" class=""></p><div style="border-collapse: separate; border-spacing: 0px;" class=""><font class="Apple-style-span" color="#002efa"><u class=""><a href="http://csmasters.uchicago.edu/" class="">http://csmasters.uchicago.edu</a></u></font></div><div style="border-collapse: separate; border-spacing: 0px;" class=""><font class="Apple-style-span" color="#002efa"><u class=""><a href="http://facebook.com/UChicagoMastersCS" class="">Facebook.com/UChicagoMastersCS</a></u></font></div><div style="border-collapse: separate; border-spacing: 0px;" class=""><font class="Apple-style-span" color="#002efa"><u class=""><a href="http://twitter.com/UChicagoMPCS" class="">Twitter.com/UChicagoMPCS</a></u></font></div></font></div></div></span>
</div>
<br class=""><div class=""><br class=""></div><div class=""></div></body></html>