|Full Title:||Workshop: Can Cognitive Scientists Help Computers Recognize Irony?|
|Short Title:||Irony at CogSci 2014|
|Location:||Quebec City, Canada|
|Start Date:||23-Jul-2014 - 23-Jul-2014|
|Meeting Email:||click here to access email|
Irony is an important rhetorical device that takes many forms. The successful ironist effectively communicates something other than (and often opposite to) what he or she has literally said. Historically, the ironic voice has been studied by researchers in philosophy, language, social cognition and cognitive science. More recently, the problem of automatically detecting irony has garnered attention from computer scientists working in machine learning (ML) and natural language processing (NLP).
But classifying utterances as ironic has proven uniquely difficult. The standard ML approach to text classification is the 'bag-of-words' approach. With a sufficient amount of manually categorized examples (i.e., training data), such models can be extremely successful in a variety of classification tasks, e.g., spam filtering. But irony detection has proven to be much harder. Our view is that cognitive scientists may have much to offer computer science researchers interested in this problem.
Capitalizing on the co-location of CogSci with AAAI, this workshop thus aims to bring cognitive and computer scientists together to explore novel models for irony detection. In particular, we believe that developing representations of speakers and contexts and building models that factor these representations into judgments of utterances may drastically improve automated irony detection.
|Linguistic Subfield:||Cognitive Science; Ling & Literature; Philosophy of Language; Pragmatics; Psycholinguistics|
|Calls and Conferences main page|