Have a UROP opening you would like to submit?
Please fill out the form.
Detecting Custom Features in Speech via Natural Language Processing
MAS: Media Arts and Sciences
May 5 2020
Joanne Leong, email@example.com
We would like to create a novel system and interface to support personalized and on-the-go language learning, in the realm of Human-Computer Interaction (HCI). We are interested in a system that can provide support in improving one’s speech. Help to program pipelines to detect custom features in speech (e.g. choice of words, sentence structure, prosody) in real-time. The student will focus primarily on designing and programming these pipelines, which may involve (1) building custom models and/or (2) leveraging existing NLP models and frameworks (e.g. BERT, NLTK) and/or (3) utilizing third-party services (e.g. Google Cloud, Azure, AWS) for real-time analysis. The resulting pipelines will be incorporated into a desktop-based experience and will be leveraged to conduct user-studies. This is an opportunity to collaborate in creating a novel experience, and in doing so bring your own creativity to the table to solve some unique challenges. We aim to publish the results of this project as a full conference-paper at the CHI 2021 conference.
The ideal candidate has prior experience in machine learning and natural language processing (NLP). Relevant knowledge to have: - Experience with third-party cloud computing services (Google Cloud, Azure, AWS) for speech processing, as well as state of the art Natural Language Processing (NLP) techniques (e.g. BERT, NLTK, etc.) - Programming (Python, Tensorflow)