Lucas Caccia
Senior Researcher
About
My current research centers on building agents that are modular and composable. I believe that modularity can enable decentralized, continual, and collaborative model development. One successful instantiation of this is model MoErging (opens in new tab), where different (independently trained) expert models are combined into an MoE-style architecture. To this end, I have done work both on the learning better experts (opens in new tab) and routing among available experts (opens in new tab).
Prior to that, I completed in November my Ph.D. at McGill and the Quebec Artificial Intelligence Institute (Mila), where I was advised by Joelle Pineau (opens in new tab). My PhD thesis (opens in new tab) focused on enabling efficient and robust Continual Learning in neural networks.