[2309.02022] Early Exiting Predictive Coding Neural Networks for Edge AI

Computer Science > Machine Learning arXiv:2309.02022 (cs) [Submitted on 5 Sep 2023 (v1), last revised 31 Mar 2026 (this version, v2)] Title:Early Exiting Predictive Coding Neural Networks for Edge AI Authors:Alaa Zniber, Mounir Ghogho, Ouassim Karrakchou, Mehdi Zakroum View a PDF of the paper titled Early Exiting Predictive Coding Neural Networks for Edge AI, by Alaa Zniber and 3 other authors View PDF HTML (experimental) Abstract:The Internet of Things is transforming various fields, with sensors increasingly embedded in wearables, smart buildings, and connected equipment. While deep learning enables valuable insights from IoT data, conventional models are too computationally demanding for resource-limited edge devices. Moreover, privacy concerns and real-time processing needs make local computation a necessity over cloud-based solutions. Inspired by the brain's energy efficiency, we propose a shallow bidirectional predictive coding network with early exiting, dynamically halting computations once a performance threshold is met. This reduces the memory footprint and computational overhead while maintaining high accuracy. We validate our approach using the CIFAR-10 dataset. Our model achieves performance comparable to deep networks with significantly fewer parameters and lower computational complexity, demonstrating the potential of biologically inspired architectures for efficient edge AI. Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computer Visio...