About me
I'm an undergraduate student pursuing degrees in Computer Science and Mathematics at Brown University, where I was fortunately advised by Professor Amy Greenwald and Chen Sun. I have interests for a variety of topics, currently I'm devoted to research in games, learning, and multi-agent interactions.
During my time at Brown, I received an Undergraduate Teaching and Research Awards, and I was ranked 2nd place in the International Automated Negotiation Agents Competition's Supply Chain Management League. Besides school and research, I have a strong passion for teaching as well. I appreciate the opportunity given by Professor Ritambbara Singh to be a part of the teaching assistant staff for Deep Learning.
Other than Computer Science, I'm also interested in theatre works and movies. My favorite play is 'Angels in America' and I've been writing since high school.
News
Summer 2024, I'm employed by Professor Amy Greenwald to work on the intersection of Stackelberg games and reinforcement learning algorithms. I'm gaining experience in implicit function differentiation, bi-level optimization, and I'm building a reinforcement learning baseline repo for future comparison.
Selected Projects
Experimental testing and analysis of aerodynamic interactions between quadrotors
Over the course of the semester, I supported a graduate student in gathering data from flight experiments involving quadrotors traversing each other. These experiments included both static setups, where drones were tethered as they traversed each other, and dynamic, free-flight experiments conducted in the laboratory. I had the opportunity to learn about quadrotor hardware through hands-on experiments and integrated it with external localization systems for indoor flight. Working collaboratively, I contributed to the development of these experiments and played a role in setting up and testing drones in close-proximity flight.
TOLD-ZERO
TDMPC-2 is a model-based reinforcement learning algorithm on continuous action control. In this work we present TOLD-ZERO, a generalization of the TDMPC paper that specializes in discrete action tasks and uses Monte-Carlo Tree Search as a local trajectory optimization method. We also try to argue the role of planning in model-based RL in both continuous and discrete action tasks, specifically by benchmarking algorithms with or without planning on the LightZero Environment.
Matching Pennis(An Negotiation Agent Submitted to SCML 2024)
The MatchingPennies Agent employs concurrent negotiation and a heuristic function to identify the most profitable subsets of current offers in a setting of multi-agent negotiation provided by negmas. The ability to balance between fulfilling the desired quantities and negoatiating for the best price gives its name. The agent won the 2nd place in the SCML2024 one-shot track.
Pipelining Natural Language to Multi-Robot Task Allocation
This initiative aims to simplify commanding a team of robots through a logical intermediary. By translating natural language commands into linear temporal logic(LTL), we've enhanced the success rate of task execution. The approach supports complex commands and collaborative tasks between human and teams of robots.
