About CoLa Lab

The Computation of Language Laboratory investigates the complex system we call human language, focusing on information acquisition and extraction.

The overview in 2020, targeted at prospective graduate students

Under the direction of Dr. Lisa S. Pearl, members of CoLa Lab attempt to answer questions such as these:

  • How do children acquire the information about language that they do from the data that they have?
    This includes questions about the learning strategies they use, the learning biases they have, the kinds of knowledge representations that are easier (or harder) to learn, and the underlying factors that result in the learning trajectory we observe.

    You can see presentations describing how to use computational modeling to understand acquisition here and here.

  • How do people acquire the non-linguistic information about the world that they do from the language data that they have? Can machines be made to do it as well?
    This includes questions about authorship & identity, message tone, and transmission of mental states through language. It sometimes goes under the general heading of "computational sociolinguistics".

The main technique of investigation we use is empirically-grounded computational modeling, drawing on constraints from realistic examples of human language and what we know about how humans process language information. Complementary techniques include psycholinguistic methodologies to assess knowledge in children and adults, and human computation methodologies for gathering realistic samples of language use and interpretation.

You can see me talking about the kind of research we do in the lab in an interview here and a video here.