Dominic Bryce Simon

me.png

Feel free to reach out for any collaborations or job offers.

Doctoral Candidate at the University of Florida researching lightweight methods to edit Large Language Model knowledge.

Personal Profile

I was born and raised in Orlando, Florida. I went to Lake Highland Preparatory School for middle and high school and graduated in May 2017. I started my undergraduate degree at the University of Central Florida in August 2017 and graduated in December 2021. I went straight into my Doctorate, starting at UCF in January 2022. I earned enough credits through my Doctorate to also recieve my Master’s in May 2024. I came with my advisor, Dr. Rickard Ewetz, to the University of Florida in August 2024, and am finishing out my degree there now.

In my free time, I like to play video games, drink craft beer, pretend to understand botany, and pretend to learn other languages. I also REALLY like to eat. If any of those things resonate with you, feel free to reach out.

Research Profile

Recent years have seen an extreme proliferation of Large Language Models into both business and personal life. The usage of these models relies on the extensive “knowledge” they gain during training. However, this “knowledge” can become out of date due to the natural passage of time. The brute force solution to update models is then to update the training data to reflect the new reality and retrain the model. Unfortunately, training is extremely computationally and financially expensive, warranting research into methods that can insert new knowledge into small numbers of model paramaters.

My research spans this area of Knowledge Editing, as it is called. My contributions to this field include retrieved-augmented generation-based editors, paratermeter modification-based editors, and a dataset. I have also previously worked on adversarial robustness of computer vision models. Through personal efforts and collaborations with colleagues, I have papers published in the top AI conferences ICML and IJCAI. All research areas I have published in include:

  • Adversarial Robustness of Computer Vision Models (specifically, defenses)
  • Knowledge Editing of Large Language Models
  • Explainable AI (attributions for Computer Vision models)
  • Neuro-Symbolic AI

If any of my work is of interest or if you would like to collaborate, please feel free to reach out.