Natural Language Processing
"Think before you speak": Language Generation with Planning
Guarding Against Spurious Correlations in Natural Language Understanding
Challenges and Opportunities in Evaluating Progress in NLP
Abstract: The past few years have seen remarkable advances in NLP, as evidenced both by continued and rapid gains on benchmark tasks, as well as by the increasing prominence of real NLP systems in the wild. In assessing such progress, however, it is important to ask not only what system achieves the best performance, but how it achieves that level of performance, how much we can trust the evaluation, and what the consequences of deploying such a system might be.
Xiaozhong Liu (School of Informatics, Computing, and Engineering; Indiana University, Bloomington)
Natural Language Processing and Text Mining with User Behavior Data
Cache Transition Systems for Semantic Parsing
Abstract: We describe a transition system that generalizes standard transition-based dependency parsing techniques to generate a graph rather than a tree. Our system includes a cache with fixed size m, and we characterize the relationship between the parameter m and the class of graphs that can be produced through the graph-theoretic concept of tree decomposition. We train a sequence-to-sequence neural model based on this system to parse text into Abstract Meaning Representation (AMR).
Assisting Interpreters with Technical Terms
Abstract: To assist simultaneous interpreters with their work, one needs to find a balance between providing useful information and minimizing distraction. This talk will discuss a simple method for detecting and translating technical terms and a crowdsourced experiment to determine whether the method strikes the right balance for interpreter assistance--without actually using interpreters.
Lunch served by 12:15.
Pages
