Joint Optimization of Model Partitioning and Resource Allocation for Anti-Jamming Collaborative Inference Systems

TL;DR

This paper proposes a joint optimization framework for model partitioning and resource allocation to enhance the robustness and performance of collaborative DNN inference systems against malicious jamming attacks.

Contribution

It introduces a novel anti-jamming collaborative inference system with an optimization algorithm that jointly allocates resources and partitions models to maximize delay and accuracy.

Findings

The proposed scheme outperforms baseline methods in RDA metrics.

Joint optimization improves inference robustness under jamming.

Efficient algorithms solve complex mixed-integer nonlinear problems.

Abstract

With the increasing computational demands of deep neural network (DNN) inference on resource-constrained devices, DNN partitioning-based device-edge collaborative inference has emerged as a promising paradigm. However, the transmission of intermediate feature data is vulnerable to malicious jamming, which significantly degrades the overall inference performance. To counter this threat, this letter focuses on an anti-jamming collaborative inference system in the presence of a malicious jammer. In this system, a DNN model is partitioned into two distinct segments, which are executed by wireless devices and edge servers, respectively. We first analyze the effects of jamming and DNN partitioning on inference accuracy via data regression. Based on this, our objective is to maximize the system's revenue of delay and accuracy (RDA) under inference accuracy and computing resource constraints by jointly optimizing computation resource allocation, devices' transmit power, and DNN partitioning. To address the mixed-integer nonlinear programming problem, we propose an efficient alternating optimization-based algorithm, which decomposes the problem into three subproblems that are solved via Karush-Kuhn-Tucker conditions, convex optimization methods, and a quantum genetic algorithm, respectively. Extensive simulations demonstrate that our proposed scheme outperforms baselines in terms of RDA.