<div dir="ltr"><div><span style="font-family:arial,sans-serif;font-size:13px"><b>CL+NLP Lunch </b>(</span><a href="http://www.cs.cmu.edu/~nlp-lunch/" target="_blank" style="font-family:arial,sans-serif;font-size:13px">http://www.cs.cmu.edu/~nlp-lunch/</a><span style="font-family:arial,sans-serif;font-size:13px">)</span><br style="font-family:arial,sans-serif;font-size:13px">
<b style="font-family:arial,sans-serif;font-size:13px">Speaker</b><span style="font-family:arial,sans-serif;font-size:13px">: </span><span style="font-family:arial,sans-serif;font-size:13px">André Martins</span><span style="font-family:arial,sans-serif;font-size:13px">, </span><span style="font-family:arial,sans-serif;font-size:13px">Priberam Labs</span><br style="font-family:arial,sans-serif;font-size:13px">
<b style="font-family:arial,sans-serif;font-size:13px">Date</b><span style="font-family:arial,sans-serif;font-size:13px">: Monday, May 20, 2013</span><br style="font-family:arial,sans-serif;font-size:13px"><b style="font-family:arial,sans-serif;font-size:13px">Time</b><span style="font-family:arial,sans-serif;font-size:13px">: 12:00 noon</span><br style="font-family:arial,sans-serif;font-size:13px">
<b style="font-family:arial,sans-serif;font-size:13px">Venue</b><span style="font-family:arial,sans-serif;font-size:13px">: GHC </span><span style="font-family:arial,sans-serif;font-size:13px">4405</span><span style="font-family:arial,sans-serif;font-size:13px"><br>
</span></div><span style="font-family:arial,sans-serif;font-size:13px"><div><span style="font-family:arial,sans-serif;font-size:13px"><br></span></div><b>Title</b>: </span><div><span style="font-family:arial,sans-serif;font-size:13px">Fast and Robust Compressive Summarization with Dual Decomposition and Multi-Task Learning</span><br style="font-family:arial,sans-serif;font-size:13px">
<br style="font-family:arial,sans-serif;font-size:13px"><span style="font-family:arial,sans-serif;font-size:13px"><b>Abstract</b>: </span><div><span style="font-family:arial,sans-serif;font-size:13px">We present a dual decomposition framework for multi-document summarization, using a model that jointly extracts and compresses sentences. Compared with previous work based on integer linear programming, our approach does not require external solvers, is significantly faster, and is modular in the three qualities a summary should have: conciseness, informativeness, and grammaticality. In addition, we propose a multi-task learning framework to take advantage of existing data for extractive summarization and sentence compression. Experiments in the TAC-2008 dataset yield the highest published ROUGE scores to date, with runtimes that rival those of extractive summarizers. This work was done jointly with Miguel Almeida at Priberam Labs, and will appear soon at ACL 2013.</span><br style="font-family:arial,sans-serif;font-size:13px">
<br style="font-family:arial,sans-serif;font-size:13px"><span style="font-family:arial,sans-serif;font-size:13px"><b>Biography</b>: </span></div><div><span style="font-family:arial,sans-serif;font-size:13px">André Martins is a research scientist at Priberam Labs, in Lisbon, Portugal. He received his dual-degree PhD in Language Technologies in 2012 from Carnegie Mellon University and Instituto Superior Técnico. His PhD dissertation was awarded Honorable Mention in CMU's SCS Dissertation Award competition. Martins' research interests include natural language processing, machine learning, structured prediction, sparse modeling, and optimization. He received a best paper award at the ACL 2009 conference and the Portuguese IBM 2011 Scientific Prize.</span><br>
</div></div></div>