Have a UROP opening you would like to submit?
Please fill out the form.
Computational Modeling of Generic Language
9: Brain and Cognitive Sciences
MH Tessler: firstname.lastname@example.org
Generics are statements such as “tigers are striped”, “a duck lays eggs”, and “ticks carry Lyme disease”. Generics express generalizations, and have funny properties that make them difficult to define in mathematical terms: Only 50% of ducks lay eggs (i.e., the female ducks) , and yet “Ducks lay eggs is true”; 50% of ducks are female and yet “Ducks are female” sounds like a strange sentence. “Ticks carry Lyme disease” is true, even though only a small fraction of ticks carry the disease. What makes a generic sentence true or false? In this UROP (summer, can be extended to the fall), you learn about modeling techniques in computational cognitive science and probabilistic language understanding, and be trained to understand a state-of-the-art computational model of generic language and probe it for its capacity to understand different generic statements. You will try to understand the conditions under which the model produces a human-like intuition by taking examples from formal linguistics and testing them in the model. Priority will be given to those who have taken 9.66/6.804, who have familiar with a probabilistic programming language, who are comfortable using R / RMarkdown / the tidyverse, experience with GitHub, knowledge of semantics and pragmatics and/or web programming (js / html / css).
- Responsible, independent, and highly attentive to detail. - Available to work 30-40 hours per week during the summer