Likelihood Training of Schrödinger Bridge using Forward-Backward SDEs Theory
arXiv:2110.11291
Abstract
Schrödinger Bridge (SB) is an entropy-regularized optimal transport problem that has received increasing attention in deep generative modeling for its mathematical flexibility compared to the Scored-based Generative Model (SGM). However, it remains unclear whether the optimization principle of SB relates to the modern training of deep generative models, which often rely on constructing log-likelihood objectives.This raises questions on the suitability of SB models as a principled alternative for generative applications. In this work, we present a novel computational framework for likelihood training of SB models grounded on Forward-Backward Stochastic Differential Equations Theory - a mathematical methodology appeared in stochastic optimal control that transforms the optimality condition of SB into a set of SDEs. Crucially, these SDEs can be used to construct the likelihood objectives for SB that, surprisingly, generalizes the ones for SGM as special cases. This leads to a new optimization principle that inherits the same SB optimality yet without losing applications of modern generative training techniques, and we show that the resulting training algorithm achieves comparable results on generating realistic images on MNIST, CelebA, and CIFAR10. Our code is available at https://github.com/ghliu/SB-FBSDE.
arXiv commentfix appendix net arh error
References in corpus (34)
- Decoupled Weight Decay Regularization
- Denoising Diffusion Probabilistic Models
- GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium
- Diffusion Models Beat GANs on Image Synthesis
- Solving high-dimensional partial differential equations using deep learning
- Score-Based Generative Modeling through Stochastic Differential Equations
- Generative Modeling by Estimating Gradients of the Data Distribution
- Training Generative Adversarial Networks with Limited Data
- Searching for Activation Functions
- HiFi-GAN: Generative Adversarial Networks for Efficient and High Fidelity Speech Synthesis
- Improved Denoising Diffusion Probabilistic Models
- NVAE: A Deep Hierarchical Variational Autoencoder
- Variational Diffusion Models
- Improved Techniques for Training Score-Based Generative Models
- FFJORD: Free-form Continuous Dynamics for Scalable Reversible Generative Models
- DiffWave: A Versatile Diffusion Model for Audio Synthesis
- BIVA: A Very Deep Hierarchy of Latent Variables for Generative Modeling
- Learning Optimal Control via Forward and Backward Stochastic Differential Equations
- Sliced Score Matching: A Scalable Approach to Density and Score Estimation
- Diffusion Schrödinger Bridge with Applications to Score-Based Generative Modeling
- WaveFlow: A Compact Flow-based Model for Raw Audio
- On Fast Sampling of Diffusion Probabilistic Models
- An Optimal Control Approach to Deep Learning and Applications to Discrete-Weight Neural Networks
- Maximum Likelihood Training of Score-Based Diffusion Models
- Very Deep VAEs Generalize Autoregressive Models and Can Outperform Them on Images
- A Variational Perspective on Diffusion-Based Generative Models and Score Matching
- Augmented Normalizing Flows: Bridging the Gap Between Generative Flows and Latent Variable Models
- VFlow: More Expressive Generative Flows with Variational Data Augmentation
- Gotta Go Fast When Generating Data with Score-Based Models
- Score-based Generative Modeling in Latent Space
- Styleformer: Transformer based Generative Adversarial Networks with Style Vector
- Deep Generative Learning via Schrödinger Bridge
- DDPNOpt: Differential Dynamic Programming Neural Optimizer
- Interpreting diffusion score matching using normalizing flow