https://nico-espinosadice.github.io/projects/Recent content in Projects on Nicolas Espinosa-DiceHugo -- gohugo.ioen-ushttps://nico-espinosadice.github.io/projects/evor/Mon, 01 Jan 0001 00:00:00 +0000https://nico-espinosadice.github.io/projects/evor/Expressive Value Learning for Scalable Offline Reinforcement Learning Nicolas Espinosa-Dice1, Kianté Brantley2, Wen Sun1 1Cornell University, 2Harvard University Paper | Code | Thread Abstract RL has yet to be fully leveraged in robotics, principally due to its lack of scalability. Offline RL offers a promising avenue by training agents on large, diverse datasets, avoiding the costly real-world interactions of online RL. Scaling offline RL to increasingly complex datasets requires expressive generative models such as diffusion and flow matching.https://nico-espinosadice.github.io/projects/sorl/Mon, 01 Jan 0001 00:00:00 +0000https://nico-espinosadice.github.io/projects/sorl/Scaling Offline RL via Efficient and Expressive Shortcut Models Nicolas Espinosa-Dice1, Yiyi Zhang1, Yiding Chen1, Bradley Guo1, Owen Oertell1, Gokul Swamy2, Kianté Brantley3, Wen Sun1 1Cornell University, 2Carnegie Mellon University, 3Harvard University Paper | Code | Thread | Blog Post Abstract Diffusion and flow models have emerged as powerful generative approaches capable of modeling diverse and multimodal behavior. However, applying these models to offline reinforcement learning (RL) remains challenging due to the iterative nature of their noise sampling processes, making policy optimization difficult.