I’m interested in describing the mechanisms behind indirect request production and comprehension. Probably my biggest question in this area is: if we were to build a computational model of indirect request comprehension and intention inference, what factors would such a model need to include to reflect human behavior?
Some more specific questions I’m looking at:
- What factors constrain the form and conventionality of an indirect request?
- How does the conventionality of an indirect request affect its comprehension by the listener?
- To what extent do people build mental models of their interlocutors during conversation, and how do they use these models to aid in inferring the speaker’s intentions?
The relationship between words and their meaning is mostly arbitrary, but there are systematic or statistical trends in form-meaning pairings. How can we build better tools to uncover these trends automatically? And what do these pockets of systematicity tell us about language evolution and language learning?
Trust and Human-Robot Interaction
As smart technology becomes more pervasive, it’s important that we develop the right levels of trust in our machines; we should trust them to do what they can do, but we shouldn’t assume that they can do something they can’t. I hope to address this problem by answering a couple of questions:
- What factors (behaviors, interactions, etc.) of a machine affect the degree of trust we place in it, and the expertise judgments we make about it? What can we learn about judgments of human expertise?
- Can we design machines to efficiently calibrate the appropriate level of trust in their users?
Natural Language Understanding
My previous work at the International Computer Science Institute involved building a general framework for natural language understanding – producing action from language. I am still interested in the problem of building systems that can understand natural language. In the future, I hope to approach the problem from the perspectives of language acquisition (e.g. with construction grammar) and inference (see “indirect requests” above).