Summary of Amortizing Intractable Inference in Diffusion Models For Vision, Language, and Control, by Siddarth Venkatraman et al.
Amortizing intractable inference in diffusion models for vision, language, and control
by Siddarth Venkatraman, Moksh Jain, Luca Scimeca, Minsu Kim, Marcin Sendera, Mohsin Hasan, Luke Rowe, Sarthak Mittal, Pablo Lemos, Emmanuel Bengio, Alexandre Adam, Jarrid Rector-Brooks, Yoshua Bengio, Glen Berseth, Nikolay Malkin
First submitted to arxiv on: 31 May 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Computer Vision and Pattern Recognition (cs.CV)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper studies amortized sampling of the posterior over data in diffusion models used as priors in downstream tasks. It proposes a data-free learning objective called relative trajectory balance for training a diffusion model that samples from this posterior, which is an open problem in existing methods. The approach uses deep reinforcement learning techniques to improve mode coverage and has broad potential applications in vision, language, and multimodal data. In addition, the paper applies relative trajectory balance to continuous control with a score-based behavior prior, achieving state-of-the-art results on offline reinforcement learning benchmarks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper explores how diffusion models can be used as priors in other tasks, but there’s an issue with calculating the posterior distribution of this prior. The researchers propose a new way to learn a model that can sample from this posterior without needing any data. This approach uses ideas from deep reinforcement learning and could be useful for many applications, including generating images or filling in missing text. |
Keywords
» Artificial intelligence » Diffusion » Diffusion model » Reinforcement learning
