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

2025-10-23

AI Summary

To address the tripartite challenges of privacy leakage, high cloud latency, and low edge-side accuracy in autism spectrum disorder (ASD) intelligent diagnosis within school settings, this paper proposes C3EKD, a cloud-edge collaborative hierarchical inference framework. C3EKD employs an edge-side confidence-based sample filtering mechanism to dynamically offload low-confidence samples to the cloud; integrates temperature-scaled soft labels and reverse knowledge distillation to enable efficient cloud-to-edge knowledge transfer; and introduces a global aggregation loss function to support cross-domain knowledge fusion and model generalization across decentralized multi-school datasets. Evaluated on two public ASD facial image datasets, the system achieves 87.4% classification accuracy—significantly outperforming both pure-cloud and pure-edge baselines—while preserving data privacy and ensuring low-latency inference. Results demonstrate C3EKD’s feasibility for large-scale deployment and its superiority in balancing accuracy, efficiency, and privacy.

Technology Category

• Machine Learning: Learning on the Edge & Model Compression
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization
• Computer Vision: Multi-modal Vision

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Cloud, edge and content delivery systems for the Web
• Responsible Web: Data and user privacy-enhancing technologies for the Web
• Security and Privacy: Data transparency and provenance

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.

Problem

• Addresses privacy and latency in cloud-based ASD diagnosis
• Improves edge inference accuracy for autism spectrum disorder
• Enables scalable deployment without centralizing sensitive data

Innovation

• Edge-cloud collaborative framework for ASD diagnosis
• Selective cloud upload of low-confidence samples
• Knowledge distillation with temperature-scaled soft labels