ANGELO – Assigning Grammatical Gender to Loanwords
How do speakers of languages with grammatical gender know what gender to assign to borrowed words? I built a supervised analogical model to predict the gender of English loanwords borrowed into Hindi-Urdu. You can take a closer look at the model and its predictions here.
The World Possible project is dedicated to providing open-source educational content to schools that lack internet connectivity. This is done through a device called RACHEL(Remote Area Community Hotspot for Education and Learning), which is a cached web server containing open-sourced content like Khan Academy videos, Wikipedia pages, and books from the Gutenberg corpus.
The only hindrance is that this device currently costs $500 USD. This semester, as part of the World Librarians Initiative at UMass Amherst, I helped recreate the RACHEL server using a Raspberry-Pi which brought the cost of the device down to $80 USD. The documentation for creating your own RACHEL-Pi at home can be found here.
I studied word association in non-literal Hindi data by creating a corpus of Bollywood song lyrics and performing association tests on the resulting data. You can take a closer look at the analysis here.
At UMass, I am a research assistant at the Cognitive Science of Language Lab. Here, I perform eye-tracking to help conduct sentence-processing experiments. I recently designed and piloted an experiment of my own. You can read more about it here.
Previously, I worked at the Language, Intersensory Perception and Speech Lab with Professor Alexandra Jesse where I assisted with the implementation of speech-processing experiments