For fans of Computational Linguistics / Natural Language Processing: ---------- Forwarded message ---------- From: Harrington, Jacob Walter Date: Tue, Oct 11, 2016 at 1:55 PM Subject: [Cs-ugrads] IVC Seminar, 1pm - 2pm on 10/12, MCS 148 To: "cs-grads@cs.bu.edu" , "cs-ugrads@cs.bu.edu" Humour, Argument, and Semantics: Learning for Complex Natural Language Tasks Professor Anna Rumshisky, University of Massachusetts Lowell Wednesday, October 12 at 1pm in MCS 148 ABSTRACT In the past few years, deep learning approaches that rely on learned, rather than engineered feature representations have shown excellent results in many natural language prediction and classification tasks, from machine translation to sentiment analysis. Despite the burgeoning success of models that use attention and memory mechanisms to handle problems that require deeper reasoning, many complex natural language tasks still remain out of reach. In this talk, I will discuss our work on several such tasks, including computing semantic similarity of text, argumentation mining, and humour detection. Assessing semantic similarity of text is the holy grail of language understanding and is at the core of many tough NLP problems -- such as entailment recognition or paraphrase detection. I will examine the challenges inherent in this task, present our current results, and discuss possible solutions. Argumentation mining involves identifying relations (such as support or attack) between the claims made in text, and I will discuss our results that apply attention-based pointer networks to this problem. Finally, humour detection requires a combination of multiple divergent aspects of language understanding. I will discuss the effects of integrating different information and outline future directions. BIO Anna Rumshisky is an Assistant Professor of Computer Science at the University of Massachusetts Lowell, where she heads the Text Machine Lab for NLP. Her primary research area is natural language processing, with applications in computational lexical semantics, temporal reasoning, and clinical informatics, as well as digital humanities and social science. She received her PhD from Brandeis University and completed postdoctoral training at MIT CSAIL. Her work has been funded by National Institutes of Health, Army Research Office, and other government and industry grants. _______________________________________________ Cs-ugrads mailing list Cs-ugrads@cs.bu.edu http://cs-mailman.bu.edu/mailman/listinfo/cs-ugrads