Have a UROP opening you would like to submit?
Please fill out the form.
Artie, an Artificial Heat Transfer Analyst: Natural Language Processing
2: Mechanical Engineering
Tony Patera: email@example.com
We are developing an "Artificial Heat Transfer Student" which could earn an A in an undergraduate heat transfer subject. The software will accept problem set-like statements and provide solutions as expected of actual students. Our approach divides the software into two parts: The frontend applies natural language and image processing techniques to translate problem statements into heat transfer specifications. The frontend also summarizes the solution, complete with derivations and justifications, in a technical report and associated graphics. The backend operates on the heat transfer specifications to develop appropriate mathematical models and associated approximate solutions, with particular attention to the most challenging limit — back-of-the-envelope estimates justified through appropriate simplifications of the geometry and physics. In this particular UROP project we will develop a crucial part of the frontend: natural language processing and machine learning software to translate non-prescriptive problem statements into appropriate specifications — type of problem to be solved; characterizations of the object and its environment; engineering-relevant outputs — for the backend. Ultimately design decisions may also be incorporated.
Extensive Matlab experience; some coursework in AI, natural language processing, or machine learning.