arXiv (Cornell University)· 2015· Preprint
Deep Unsupervised Learning using Nonequilibrium Thermodynamics
- 1,421citations
- 2015year
Short summary
A novel deep generative model learns complex data distributions by reversing a diffusion process, inspired by non-equilibrium thermodynamics.
AI-generated from the title and abstract; the full text is not read.
Key points
- Develops a deep generative model inspired by non-equilibrium statistical physics.
- Uses a forward diffusion process to destroy data structure and a learned reverse process to restore it.
- Achieves flexibility and tractability in modeling complex data distributions.
- Enables rapid learning, sampling, and evaluation in deep models with thousands of layers.
- Allows computation of conditional and posterior probabilities.
AI-generated from the title and abstract; the full text is not read.
Abstract
A central problem in machine learning involves modeling complex data-sets using highly flexible families of probability distributions in which learning, sampling, inference, and evaluation are still analytically or computationally tractable. Here, we develop an approach that simultaneously achieves both flexibility and tractability. The essential idea, inspired by non-equilibrium statistical physics, is to systematically and slowly destroy structure in a data distribution through an iterative forward diffusion process. We then learn a reverse diffusion process that restores structure in data, yielding a highly flexible and tractable generative model of the data. This approach allows us to rapidly learn, sample from, and evaluate probabilities in deep generative models with thousands of layers or time steps, as well as to compute conditional and posterior probabilities under the learned model. We additionally release an open source reference implementation of the algorithm.
The authors' abstract, as published at the source. arXiv (Cornell University), 2015 · DOI ↗
Continue with a free account
Ask the paper: 3 free questions a day about this paper; save it, get its citation, new summaries every day for your field. Takeaways are Premium.
Continue free on the webSign in with Google or Apple; no card needed. You come back to this paper.
On your phone:
Field: Statistical and Nonlinear Physics
Statistical and Nonlinear PhysicsPhysics and Astronomy