Deep Unsupervised Learning using Nonequilibrium Thermodynamics

  title={Deep Unsupervised Learning using Nonequilibrium Thermodynamics},
  author={Jascha Sohl-Dickstein and Eric A. Weiss and Niru Maheswaranathan and Surya Ganguli},
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… CONTINUE READING
Highly Cited
This paper has 73 citations. REVIEW CITATIONS
Recent Discussions
This paper has been referenced on Twitter 89 times over the past 90 days. VIEW TWEETS
54 Citations
51 References
Similar Papers


Publications citing this paper.

73 Citations

Citations per Year
Semantic Scholar estimates that this publication has 73 citations based on the available data.

See our FAQ for additional information.

Similar Papers

Loading similar papers…