AlphaZip: Neural Network-Enhanced Lossless Text Compression

TL;DR

A lossless text compression approach using a Large Language Model (LLM) that involves two key steps: first, prediction using a dense neural network architecture, such as a transformer block; second, compressing the predicted ranks with standard compression algorithms like Adaptive Huffman, LZ77, or Gzip.

Abstract

Data compression continues to evolve, with traditional information theory methods being widely used for compressing text, images, and videos. Recently, there has been growing interest in leveraging Generative AI for predictive compression techniques. This paper introduces a lossless text compression approach using a Large Language Model (LLM). The method involves two key steps: first, prediction using a dense neural network architecture, such as a transformer block; second, compressing the predicted ranks with standard compression algorithms like Adaptive Huffman, LZ77, or Gzip. Extensive analysis and benchmarking against conventional information-theoretic baselines demonstrate that neural compression offers improved performance.

Figures (5)

• Figure 1: Block diagram representing standard compression pipeline
• Figure 2: Block diagram representing the high level architecture of AlphaZip
• Figure 3: Distribution of ranks vs order of ranks for two different test cases.
• Figure 4: Compression Ratio vs Model Size
• Figure 5: Time taken vs Model Size