Lossless data compression by large models

TL;DR

This paper introduces LMCompress, a novel lossless data compression method leveraging large language models, significantly outperforming traditional algorithms across images, audio, video, and text data by exploiting models' understanding capabilities.

Contribution

The paper presents LMCompress, the first lossless compression approach that uses large models to achieve unprecedented compression ratios across multiple data types.

Findings

Doubles JPEG-XL, FLAC, H.264 compression ratios

Quadruples bz2 text compression ratio

Leverages models' understanding to improve compression efficiency

Abstract

Modern data compression methods are slowly reaching their limits after 80 years of research, millions of papers, and wide range of applications. Yet, the extravagant 6G communication speed requirement raises a major open question for revolutionary new ideas of data compression. We have previously shown all understanding or learning are compression, under reasonable assumptions. Large language models (LLMs) understand data better than ever before. Can they help us to compress data? The LLMs may be seen to approximate the uncomputable Solomonoff induction. Therefore, under this new uncomputable paradigm, we present LMCompress. LMCompress shatters all previous lossless compression algorithms, doubling the lossless compression ratios of JPEG-XL for images, FLAC for audios, and H.264 for videos, and quadrupling the compression ratio of bz2 for texts. The better a large model understands the data, the better LMCompress compresses.