I am a Ph.D. student advised by Dr. Bharat Bhargava at the Department of Computer Science at Purdue University. My current research focuses on developing reinforcement learning techniques to build agents capable of detecting and adapting to novel situations(unseen during training) in multi-agent environments.
Previously, I earned my Masters Degree in Computer Science from Johns Hopkins University where I worked on Causal Inference with Dr. Ilya Shpitser affiliated with the Malone Center for Engineering in Healthcare.
Before that, I worked at Cummins Inc. and Tata Consultancy Services Ltd. as a Software Engineer for 5 years. I received my Bachelors in Technology in Instrumentation and Control Engineering from Bharati Vidyapeth College of Engineering, New Delhi, India.
My main areas of interest are Reinforcement Learning, Machine Learning, Natural Language Processing, and Causal Inference. Broadly, I'm interested in Artificial Intelligence, Cognitive Science, and Philosophy of Science.
Bonjour, Trevor, et al. "Decision Making in Monopoly using a Hybrid Deep Reinforcement Learning Approach." accepted in IEEE Transactions on Emerging Topics in Computational Intelligence (2022)
Bonjour, Trevor, Aggarwal, Vaneet, and Bharat Bhargava “Information Theoretic Approach to Detect Collusion in Multi-Agent Games”, accepted in AAAI Symposium (2022)
A software package to convert ASL gestures to text/audio. The project aims to create a product for people with speech impairments to communicate with people who may not know ASL. We have implemented gesture recognition and conversion for the 26 alphabets for the English language. Currently, we are working on adding simple phrases.
Performed topic modeling for online clickstream data using Automatic Differentiation Variational Inference to bundle forum discussions by topics for online learning platforms.
Used deep neural networks to learn on BPTI protein trajectories for multiple users leveraging commonalities.
Used various machine learning techniques like SVM, Naïve Bayes, k-fold Cross-Validation to classify and analyze the results. Found some interesting results, especially using PCA to show similarities in composers.