CSL PhD Thesis Award 2025 Matthew Chang

Advances in LLM technology have created many opportunities for expressive and interactive robot systems. With the right infrastructure, users should be able to provide natural language instructions to robots and collaborate with robots as naturally as they would with another human. In this talk, Matthew Chang presents our latest work towards realizing these goals on an LLM-backed robotic platform, capable of performing long-horizon tasks in home environments specified in natural language, with human collaborators.