Google Research

Beyond In-Place Corruption: Insertion and Deletion In Denoising Probabilistic Models

ICML Workshop on Invertible Neural Networks, Normalizing Flows, and Explicit Likelihood Models (2021)

Abstract

Denoising diffusion probabilistic models have shown impressive results for generation of sequences by iteratively corrupting each example and then learning to map corrupted versions back to the original. However, previous work has largely focused on in-place corruption, adding noise to each pixel or token individually while keeping their locations the same. In this work, we consider a broader class of corruption processes and denoising models over sequence data that can insert and delete elements, while still being efficient to train and sample from. We demonstrate that these models outperform standard in-place models on an arithmetic sequence task, and that when trained on the text8 dataset they can be used to fix spelling errors without any fine-tuning.

Research Areas

Learn more about how we do research

We maintain a portfolio of research projects, providing individuals and teams the freedom to emphasize specific types of work