NLP & Speech
Modeling Reportable Events as Turning Points in Narrative | Sarcastic or Not: Word-Embeddings to Predict the Literal or Sarcastic Meaning of Words | A Sentiment Lexicon for Standard Arabic | Density-Driven Cross-Lingual Transfer of Dependency Parsers
"Turn-taking cues in task-oriented dialogue" is named by International Speech and Communications Association as best paper published in Computer Speech and Language, 2011–2015.
Using AI techniques, Kathleen McKeown and Julia Hirschberg are creating an automated, universal sentiment detection system for new languages, including low-resource languages where translators are few.
About
Natural Language Processing and Spoken Language Processing involves computational approaches to the analysis and generation of text and speech. At Columbia these include text and speech summarization, question answering, machine translation, syntax and parsing, language generation, spoken dialogue systems, semantic representation and analysis, and the study of emotional and deceptive speech, in English, Arabic, and Mandarin, inter alia.
Spoken Language Processing Group 
Natural Language Processing Group 
Add to GMail