Corpus ID: 227209335

Score-Based Generative Modeling through Stochastic Differential Equations

@article{Song2020ScoreBasedGM,
  title={Score-Based Generative Modeling through Stochastic Differential Equations},
  author={Yang Song and Jascha Sohl-Dickstein and Diederik P. Kingma and A. Kumar and S. Ermon and Ben Poole},
  journal={ArXiv},
  year={2020},
  volume={abs/2011.13456}
}
Creating noise from data is easy; creating data from noise is generative modeling. We present a stochastic differential equation (SDE) that smoothly transforms a complex data distribution to a known prior distribution by slowly injecting noise, and a corresponding reverse-time SDE that transforms the prior distribution back into the data distribution by slowly removing the noise. Crucially, the reverse-time SDE depends only on the time-dependent gradient field (a.k.a., score) of the perturbed… Expand
15 Citations
On Maximum Likelihood Training of Score-Based Generative Models
  • Highly Influenced
  • PDF
Improved Denoising Diffusion Probabilistic Models
  • 4
  • Highly Influenced
  • PDF
Symbolic Music Generation with Diffusion Models
  • PDF
Argmax Flows and Multinomial Diffusion: Towards Non-Autoregressive Language Models
  • 1
  • PDF
Infinitely Deep Bayesian Neural Networks with Stochastic Differential Equations
  • PDF
Noise Estimation for Generative Diffusion Models
  • PDF
Stochastic Image Denoising by Sampling from the Posterior Distribution
  • 2
  • PDF
...
1
2
...

References

SHOWING 1-10 OF 59 REFERENCES
Generative Modeling by Estimating Gradients of the Data Distribution
  • 108
  • PDF
FFJORD: Free-form Continuous Dynamics for Scalable Reversible Generative Models
  • 271
  • Highly Influential
  • PDF
Deep Unsupervised Learning using Nonequilibrium Thermodynamics
  • 128
  • PDF
Variational Walkback: Learning a Transition Operator as a Stochastic Recurrent Net
  • 25
  • PDF
Permutation Invariant Graph Generation via Score-Based Generative Modeling
  • 6
  • PDF
Flow++: Improving Flow-Based Generative Models with Variational Dequantization and Architecture Design
  • 116
  • Highly Influential
  • PDF
Residual Flows for Invertible Generative Modeling
  • 96
  • Highly Influential
  • PDF
Generative Adversarial Nets
  • 21,533
  • PDF
Neural Ordinary Differential Equations
  • 938
  • Highly Influential
  • PDF
Efficient Learning of Generative Models via Finite-Difference Score Matching
  • 4
  • PDF
...
1
2
3
4
5
...