8 papers
QF3: Fast Flow RL with Filtered Q-Gradients
Chung Min Kim, Brent Yi, David McAllister +7
Flow policies have become a standard policy class for learning robot behaviors from demonstrations, but reinforcement learning is still critical for improving pre-trained flow poli…
Scalable Behavior Cloning with Open Data, Training, and Evaluation
Arthur Allshire, Himanshu Gaurav Singh, Ritvik Singh +15
We introduce ABC, a fully open-source stack for manipulation with behavior cloning. At its core is ABC-130K: the largest open-source teleoperation dataset to date, featuring 3,500…
LoV3D: Grounding Cognitive Prognosis Reasoning in Longitudinal 3D Brain MRI via Regional Volume Assessments
Zhaoyang Jiang, Zhizhong Fu, David McAllister +2
Longitudinal brain MRI is essential for characterizing the progression of neurological diseases such as Alzheimer's disease assessment. However, current deep-learning tools fragmen…
Finite Difference Flow Optimization for RL Post-Training of Text-to-Image Models
David McAllister, Miika Aittala, Tero Karras +4
Reinforcement learning (RL) has become a standard technique for post-training diffusion-based image synthesis models, as it enables learning from reward signals to explicitly impro…
Flow Matching Policy Gradients
David McAllister, Songwei Ge, Brent Yi +5
Flow-based generative models, including diffusion models, excel at modeling continuous distributions in high-dimensional spaces. In this work, we introduce Flow Policy Optimization…
Visual Imitation Enables Contextual Humanoid Control
Arthur Allshire, Hongsuk Choi, Junyi Zhang +7
How can we teach humanoids to climb staircases and sit on chairs using the surrounding environment context? Arguably, the simplest way is to just show them-casually capture a human…
Decentralized Diffusion Models
David McAllister, Matthew Tancik, Jiaming Song +1
Large-scale AI model training divides work across thousands of GPUs, then synchronizes gradients across them at each step. This incurs a significant network burden that only centra…
Rethinking Score Distillation as a Bridge Between Image Distributions
David McAllister, Songwei Ge, Jia-Bin Huang +4
Score distillation sampling (SDS) has proven to be an important tool, enabling the use of large-scale diffusion priors for tasks operating in data-poor domains. Unfortunately, SDS…