Hometown: Seattle, WA
Degree: PhD, 4th year
Program: Natural Language and Computer Reasoning
Research Group:Len Schubert
What have you enjoyed most about your time at UR?
At URCS I’ve been able to pursue bold, risky research projects in collaboration with professors that have provided the tools and support to succeed. The tight-knit URCS community fosters collaboration with other professors and students. I’ve even found it easy to have inter-departmental collaboration with the linguistics department. The departmental graduate student community is vibrant and engaging and so I’ve served in the events committee and am now serving as the graduate student representative. URCS also has a very strong undergraduate program which leads to rewarding teaching experiences as a TA and opportunities to work with and encourage undergraduate researchers.
I've really enjoyed the strong and diverse NLP research going on at Rochester and that I can go up to any of the professors with questions. I can find someone to get feedback on my work whether it relates to symbolic NLP approaches or more recently popular statistical NLP approaches.
How did you choose your research area? Did you already have an idea when you came about what you wanted to do, and did that change after you started your studies here?
I fell into my area of research while looking for research projects to work on as an undergrad. I knew that I wanted to do something related to AI and my first research project was on information extraction and knowledge base completion from text. Having enjoyed that, I decided to go into NLP in grad school. I had already decided on this topic upon coming to Rochester and haven't changed my mind in the macro-scale, though the exact research project I'm working on is a bit different from what I had imagined I would do originally.
I am working with Len Schubert on mapping natural language sentences to formal representations of meaning to enable computer reasoning in a generalized manner. The set-theoretic meaning representations I work with have foundations in logic, which provides inferential guarantees, and linguistics, which ensure a correspondence between the formal meaning and human intuitions. A reliable mapping of this sort would enable inferences over language without specialized task-specific machinery while making subtle distinctions in meaning which are not captured in current state-of-the-art language understanding systems.
Do you have future plans?
I haven't completely decided on my future plans. I really like academia, but I'd like to be able to choose where I live, which is difficult with the academic job market. There are some industry labs doing exciting work so I'm not opposed to going to one of those labs as well. Generally, I'm trying to keep my options open.