Low-Complexity Semantic Packet Aggregation for Token Communication via Lookahead Search

2025-06-24

AI Summary

To address severe semantic distortion in token communication (TC) for generative AI under intermittent channels—caused by contextual dependencies—this paper proposes a semantics-aware token grouping optimization method. The core innovation introduces the Residual Semantic Score (RSS) as a differentiable proxy metric for semantic preservation and designs SemPA-Look, a linear-complexity look-ahead aggregation framework. SemPA-Look jointly employs bandwidth-constrained look-ahead search, without-replacement intra-packet sampling, and with-replacement inter-packet sampling to achieve efficient grouping. Evaluated on remote AIGC tasks using MS-COCO, the method achieves average token similarity (ATS) and LPIPS performance comparable to exhaustive search, while reducing computational complexity by 40× versus exhaustive search and 10× versus genetic algorithms. This significantly enhances semantic fidelity and deployment feasibility of AIGC over intermittent channels.

Technology Category

• Natural Language Processing: Generation
• Search and Optimization: Learning to Search
• Multiagent Systems: Agent Communication

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Tokens are fundamental processing units of generative AI (GenAI) and large language models (LLMs), and token communication (TC) is essential for enabling remote AI-generate content (AIGC) and wireless LLM applications. Unlike traditional bits, each of which is independently treated, the semantics of each token depends on its surrounding context tokens. This inter-token dependency makes TC vulnerable to outage channels, where the loss of a single token can significantly distort the original message semantics. Motivated by this, this paper focuses on optimizing token packetization to maximize the average token similarity (ATS) between the original and received token messages under outage channels. Due to inter-token dependency, this token grouping problem is combinatorial, with complexity growing exponentially with message length. To address this, we propose a novel framework of semantic packet aggregation with lookahead search (SemPA-Look), built on two core ideas. First, it introduces the residual semantic score (RSS) as a token-level surrogate for the message-level ATS, allowing robust semantic preservation even when a certain token packet is lost. Second, instead of full search, SemPA-Look applies a lookahead search-inspired algorithm that samples intra-packet token candidates without replacement (fixed depth), conditioned on inter-packet token candidates sampled with replacement (fixed width), thereby achieving linear complexity. Experiments on a remote AIGC task with the MS-COCO dataset (text captioned images) demonstrate that SemPA-Look achieves high ATS and LPIPS scores comparable to exhaustive search, while reducing computational complexity by up to 40$ imes$. Compared to other linear-complexity algorithms such as the genetic algorithm (GA), SemPA-Look achieves 10$ imes$ lower complexity, demonstrating its practicality for remote AIGC and other TC applications.

Problem

• Optimize token packetization to maximize received message similarity
• Reduce combinatorial complexity in token grouping for outage channels
• Preserve semantics efficiently in token communication for AI applications

Innovation

• Residual semantic score for robust token preservation
• Lookahead search with fixed depth and width
• Linear complexity semantic packet aggregation