Joint Channel and Semantic-aware Grouping for Effective Collaborative Edge Inference

2025-10-02

AI Summary

To address performance degradation in wireless edge collaborative inference caused by data corruption, this paper proposes a channel- and semantic-aware dynamic grouping mechanism. The method embeds channel state information (CSI) into attention weights—replacing conventional single-dimension grouping based solely on semantics or channel conditions—and enables selective feature exchange under joint constraints of semantic similarity and channel quality via a key-value matching mechanism. Its key innovation lies in the first deep integration of physical-layer CSI into the semantic collaborative inference pipeline, overcoming the robustness limitations of decoupled modeling under adverse channel conditions. Experiments on image classification tasks demonstrate that the proposed approach achieves an average accuracy improvement of 12.7%–23.4% over local inference, purely semantic grouping, and purely channel-based grouping, significantly enhancing inference robustness and generalization under corrupted data.

Technology Category

• Machine Learning: Learning on the Edge & Model Compression
• Computer Vision: Multi-modal Vision
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Security and Privacy: Large-scale security measurements

Abstract

We focus on collaborative edge inference over wireless, which enables multiple devices to cooperate to improve inference performance in the presence of corrupted data. Exploiting a key-query mechanism for selective information exchange (or, group formation for collaboration), we recall the effect of wireless channel impairments in feature communication. We argue and show that a disjoint approach, which only considers either the semantic relevance or channel state between devices, performs poorly, especially in harsh propagation conditions. Based on these findings, we propose a joint approach that takes into account semantic information relevance and channel states when grouping devices for collaboration, by making the general attention weights dependent of the channel information. Numerical simulations show the superiority of the joint approach against local inference on corrupted data, as well as compared to collaborative inference with disjoint decisions that either consider application or physical layer parameters when forming groups.

Problem

• Improving edge inference performance via device collaboration
• Addressing semantic relevance and channel impairments jointly
• Enhancing group formation through channel-aware attention mechanisms

Innovation

• Jointly considers semantic relevance and channel states
• Makes attention weights dependent on channel information
• Groups devices for collaborative edge inference effectively