Latent Discretization for Continuous-time Sequence Compression

TL;DR

This paper introduces a continuous-time sequence compression method that decouples data from its temporal discretization, enabling higher compression rates and flexible decoding, validated across video and motion capture data.

Contribution

It proposes a novel continuous-time discretization approach for sequence compression that improves efficiency and adaptability over traditional methods.

Findings

Achieves higher compression rates on video and motion capture data.

Allows decoding at different time intervals.

Reduces computational complexity compared to existing models.

Abstract

Neural compression offers a domain-agnostic approach to creating codecs for lossy or lossless compression via deep generative models. For sequence compression, however, most deep sequence models have costs that scale with the sequence length rather than the sequence complexity. In this work, we instead treat data sequences as observations from an underlying continuous-time process and learn how to efficiently discretize while retaining information about the full sequence. As a consequence of decoupling sequential information from its temporal discretization, our approach allows for greater compression rates and smaller computational complexity. Moreover, the continuous-time approach naturally allows us to decode at different time intervals. We empirically verify our approach on multiple domains involving compression of video and motion capture sequences, showing that our approaches can automatically achieve reductions in bit rates by learning how to discretize.