Token Communication in the Era of Large Models: An Information Bottleneck-Based Approach

TL;DR

This paper introduces UniToCom, a unified token communication framework utilizing an information bottleneck principle to enhance efficiency and stability in multimodal large language models over wireless channels.

Contribution

It proposes a novel GenIB-based tokenization method and a causal Transformer-based MLLM receiver, unifying token processing and improving communication in multimodal settings.

Findings

GenIB improves token efficiency and generation quality.

UniToCom outperforms baselines under dynamic channel conditions.

The approach enhances multimodal understanding and scalability.

Abstract

This letter proposes UniToCom, a unified token communication paradigm that treats tokens as the fundamental units for both processing and wireless transmission. Specifically, to enable efficient token representations, we propose a generative information bottleneck (GenIB) principle, which facilitates the learning of tokens that preserve essential information while supporting reliable generation across multiple modalities. By doing this, GenIB-based tokenization is conducive to improving the communication efficiency and reducing computational complexity. Additionally, we develop $\sigma$-GenIB to address the challenges of variance collapse in autoregressive modeling, maintaining representational diversity and stability. Moreover, we employ a causal Transformer-based multimodal large language model (MLLM) at the receiver to unify the processing of both discrete and continuous tokens under the next-token prediction paradigm. Simulation results validate the effectiveness and superiority of the proposed UniToCom compared to baselines under dynamic channel conditions. By integrating token processing with MLLMs, UniToCom enables scalable and generalizable communication in favor of multimodal understanding and generation, providing a potential solution for next-generation intelligent communications.