A Confidence-Constrained Cloud-Edge Collaborative Framework for Autism Spectrum Disorder Diagnosis

TL;DR

This work tackles the latency and privacy challenges of cloud-centric autism screening in school settings by introducing the Confidence-Constrained Cloud-Edge Knowledge Distillation (C3EKD) framework. C3EKD runs the majority of inference at edge servers and uploads only low-confidence cases to the cloud, where a larger model provides temperature-scaled soft labels that are distilled back to edge models. The edge updates are guided by a global loss aggregated across multiple schools, enabling robust generalization across heterogeneous populations. Experiments on two ASD facial-image datasets demonstrate near cloud-level accuracy (87.4%) with substantially reduced latency (≈7.1% less than pure cloud), indicating practical potential for privacy-preserving, real-time ASD diagnosis in schools.

Abstract

Autism Spectrum Disorder (ASD) diagnosis systems in school environments increasingly relies on IoT-enabled cameras, yet pure cloud processing raises privacy and latency concerns while pure edge inference suffers from limited accuracy. We propose Confidence-Constrained Cloud-Edge Knowledge Distillation (C3EKD), a hierarchical framework that performs most inference at the edge and selectively uploads only low-confidence samples to the cloud. The cloud produces temperature-scaled soft labels and distils them back to edge models via a global loss aggregated across participating schools, improving generalization without centralizing raw data. On two public ASD facial-image datasets, the proposed framework achieves a superior accuracy of 87.4\%, demonstrating its potential for scalable deployment in real-world applications.

Figures (2)

• Figure 1: The C3EKD framework for cloud-edge collaborative ASD diagnosis.
• Figure 2: Relative accuracy (rAcc) progression over 60 communication rounds demonstrating edge model convergence toward cloud-level performance.