Colloquium Details
Generalizing Beyond the Training Distribution through Compositional Generation
Speaker: Yilun Du, MIT
Location: 60 Fifth Avenue Room C15
Date: March 28, 2024, 2 p.m.
Host: Anirudh Sivaraman & Jinyang Li
Synopsis:
Generative AI has led to stunning successes in recent years but is fundamentally limited by the amount of data available. This is especially limiting in the embodied setting – where an agent must solve new tasks in new environments. In this talk, I’ll introduce the idea of compositional generative modeling, which enables generalization beyond the training data by building complex generative models from smaller constituents. I’ll first introduce the idea of energy-based models and illustrate how they enable compositional generative modeling. I’ll then illustrate how such compositional models enable us to synthesize complex plans for unseen tasks at inference time. Finally, I'll show how such compositionality can be applied to multiple foundation models trained on various forms of Internet data, enabling us to construct decision-making systems that can hierarchically plan and solve long-horizon problems in a zero-shot manner.
Speaker Bio:
Yilun Du is a final year PhD student at MIT CSAIL advised by Leslie Kaelbling, Tomas Lozano-Perez and Joshua Tenenbaum. His research spans the fields of machine learning, artificial intelligence, computer vision and robotics, with a focus on using generative models to construct intelligent embodied agents. He is supported by the NSF Graduate Research Fellowship and was previously a research fellow at OpenAI, a visiting researcher at FAIR and a student researcher at Google Deepmind.
Notes:
In-person attendance only available to those with active NYU ID cards.