Arya Honraopatil

Hi! I am a third year Ph.D. student at University of Maryland, Baltimore County advised by Dr. Lara J. Martin.

My research focuses on how language models can stay coherent, emotionally aware, and cooperative over long, open-ended interactions. My goal is to build AI agents that help foster meaningful, collaborative partnerships between humans and machines.

Long-horizon coherence: Language models forget rules, hallucinate and have trouble tracking events in long interactions. I’m interested in methods that help language models have a consistent memory across long, evolving interactions.

Emotional awareness: Understanding what someone says is different from understanding how they feel about it. I work on helping LLMs overcome this emotional blindness by recognizing and responding with emotional context.

Cooperative multi-agent systems: Working toward shared goals with humans requires understanding a partner’s intent and reasoning about how to act on it. I design multi-agent frameworks using reinforcement learning so that AI agents can coordinate with humans and act as creative partners.

I am broadly interested in research that helps people use AI in achieving their goals. If you are interested in similar problems or would like to talk about coffee, cats, or books, feel free to reach out!