Efficient Mismatch-Tolerant Coding for Model-Driven Compression

A central insight in lossless data compression is the close connection between probabilistic next- symbol prediction and efficient sequence compression, whereby predictive models can be combined with classical coding techniques to achieve strong compression performance. Applying this approach with powerful modern learned models, such as LLMs, has been shown to achieve markedly better compression than traditional techniques across a wide range of domains. However, significant practical challenges remain, including model non-determinism, in which a model produces different predictions on different machines despite identical parameters and inputs; such mismatches between the encoder and decoder can lead to complete decoding failure. Probability Matching Interval Coding (PMATIC) was recently introduced as a drop-in framework for mismatch-robust coding and shown to enable reliable compression and decompression in the presence of bounded prediction mismatch (Adler & Tang, 2026). In this work, we present a generalization of PMATIC that allows the incorpo- ration of tight theoretical results into the design and more flexible parameter optimization, resulting in substantial improvements in compression efficiency and robustness.

Data compression is important for making file sizes smaller for efficient storage and transmission. The predictive power of LLMs (large language models) can be used to help compress data to very small sizes. However, current implementations of data compression with LLMs do not always work when the compress and decompress steps are performed on two different machines, because tiny differences in computation can cause the decompression to fail to recover the original data. We improved upon an algorithm called PMATIC, which can be executed at the compress and decompress machines to combat the issues resulting from some amount of computational differences for data compression. Our improvements are possible because we look deeper into the mathematics of how certain numerical settings for PMATIC affect the compression result and compute nearly-optimal choices for these numerical settings given the data which needs to be compressed. Our new algorithm is able to achieve small compressed file sizes while tolerating more computational differences between machines. This allows data compression using LLMs to work across different machines in an easy and practical way, which can lead to an overall reduction in storage memory sizes in the future.