The Language-Endowed Intelligent Agents (LEIA) Lab

Co-Directors: Marjorie McShane & Sergei Nirenburg

Our lab investigates aspects of human-like behavior in artificial intelligent agents by developing computational cognitive models. The goal is to build agents that not only serve as useful collaborators but can also explain their behavior, which is key to building trust in AI.

Building human-like, explainable language-endowed intelligent agents (LEIAs) is a long-term objective for which good strategy is essential. Our methodologies include:

• focusing on meaning, a hard problem that mainstream AI abandoned decades ago
• working within a knowledge-centric cognitive architecture that takes responsibility for the content, not only the design, of intelligent agents
• building microtheories (computational approaches to operationalizing cognitive capabilities) that are useful in the near term and can be enhanced over time
• developing generalized agent capabilities rather than demo-specific plugs
• addressing agent capabilities holistically, rather than pursuing narrowly-defined subtasks whose simplifying assumptions undercut any realistic utility
• applying subsets of agent capabilities to actual domains and systems

Our lab's work is primarily funded by the Office of Naval Research.

Select books

McShane, M., Nirenburg, S., and English, J. (In preparation). AI You Can Trust: Building hybrid AI through cognitive modeling plus machine learning. MIT Press. McShane, M., and Nirenburg, S. (2021). Linguistics for the Age of AI. The MIT Press. Available open access here Nirenburg, S. (ed.). 2009. Language Engineering for Lesser-Studied Languages. Amsterdam: IOS Press. McShane, Marjorie. 2005. A Theory of Ellipsis. Oxford University Press. Nirenburg, S. and V. Raskin. 2004. Ontological Semantics. Cambridge, MA: MIT Press. Nirenburg, S., H. Somers and Y. Wilks (eds.), 2002. Readings in Machine Translation. Cambridge, MA: MIT Press. Nirenburg, S. (ed.) Progress in Machine Translation. Amsterdam: IOS Press, 1993. Nirenburg, S. (ed.) Machine Translation: Theoretical and Methodological Issues. The ACL Series Studies in Natural Language, Cambridge University Press, 1987.

For more publications see Marge's publications page & Sergei at Google Scholar