Ibraheem Muhammad Moosa

LanguageX Lab, Department of Computer Science and Engineering, Penn State University, State College, PA

ibraheem_moosa_pro_pic.webp

Westgate

University Park, PA 16802

Hi, I am Moosa. I am a 1st year Computer Science PhD student at Penn State University, supervised Dr. Wenpeng Yin.

My current research focus is on understanding In-context Learning in language models, including how to use Task Instructions to achieve better results. I am also interested in trainign dynamics of language models and how that knowledge can be used to achieve better continual learning in these models.

Previously, I was working on multilingual language models. Specifically I studied how to improve tokenization using transliteration to train better multilingual language models. Before that I worked at Samsung R&D Institute Bangladesh where I worked on Samsung Internet and MyFiles.

Some of the stuff that I enjoy other than my work:

  1. Learning about history
  2. Reading books
  3. Watching movies
  4. Trying out different cuisines and
  5. Running/hiking.

The best way to reach out to me is via email.

selected publications

  1. Does Transliteration Help Multilingual Language Modeling?
    Ibraheem Muhammad Moosa, Mahmud Elahi Akhter, and Ashfia Binte Habib
    2022