Linguistic Cognition Laboratory

The Linguistic Cognition Laboratory focuses on research at the intersection of machine learning and computational linguistics. A key aim is identifying, extracting, and applying non-referential properties of text, such as rhetorical organization, metaphor, and style. We are particularly interested in examining how machine learning techniques can help to capture large-scale linguistic properties of texts that capture levels of meaning above those available at the word or clause level.

Much of our research is on Computational Stylistics, in which we devise computational methods for capturing aspects of linguistic style variation, such as that between different individuals or groups of people. In this work, we explore a variety of different potential stylistic features, such as word use, syntactic patterns, and semantic nuances, for more accurate and interpretable authorship attribution. A key theoretical and practical open problem we are working on is authorship attribution across different genres and languages. We also explore application to forensic linguistics in legal contexts.

Lab Director

Dr. Argamon smiling in front of window

Shlomo Engelson Argamon, PhD

Associate Provost for Artificial IntelligenceProfessor of Computer Science

Datasets

Collection of 19,320 blogs (681,288 posts) used in the following research:

  • Shlomo Argamon, Moshe Koppel, James W. Pennebaker, and Jonathan Schler. Mining the Blogosphere: Age, gender and the varieties of self-expression First Monday12(9), September 2007.

Appraisal lexicon, used in the following research:

  • Kenneth Bloom and Shlomo Argamon. Unsupervised Extraction of Appraisal Expressions. In Atefeh Farzindar and Vlado Kešelj, eds. Lecture Notes in Computer Science, Volume 6085/2010, pp. 290-294, Springer, 2010.
  • Kenneth Bloom, Navendu Garg, and Shlomo Argamon. Extracting appraisal expressions. In Human Language Technologies: Conference of the North American Association for Computational Linguistics (NAACL-HLT), Rochester, New York, April, 2007.
  • Casey Whitelaw, Navendu Garg, and Shlomo Argamon. Using appraisal groups for sentiment analysis. In Conference on Information and Knowledge Management, Bremen, Germany, November 2005