Parallel and Flexible Sampling from Autoregressive Models via Langevin Dynamics
arXiv:2105.08164
Abstract
This paper introduces an alternative approach to sampling from autoregressive models. Autoregressive models are typically sampled sequentially, according to the transition dynamics defined by the model. Instead, we propose a sampling procedure that initializes a sequence with white noise and follows a Markov chain defined by Langevin dynamics on the global log-likelihood of the sequence. This approach parallelizes the sampling process and generalizes to conditional sampling. Using an autoregressive model as a Bayesian prior, we can steer the output of a generative model using a conditional likelihood or constraints. We apply these techniques to autoregressive models in the visual and audio domains, with competitive results for audio source separation, super-resolution, and inpainting.
16 pages, 7 figures, ICML 2021
References in corpus (10)
- Language Models are Few-Shot Learners
- MelGAN: Generative Adversarial Networks for Conditional Waveform Synthesis
- Generating Long Sequences with Sparse Transformers
- SampleRNN: An Unconditional End-to-End Neural Audio Generation Model
- Compressed Sensing using Generative Models
- Music Source Separation in the Waveform Domain
- Jukebox: A Generative Model for Music
- One Network to Solve Them All --- Solving Linear Inverse Problems using Deep Projection Models
- WaveFlow: A Compact Flow-based Model for Raw Audio
- Problems using deep generative models for probabilistic audio source separation