RockNet: Distributed Learning on Ultra-Low-Power Devices

2025-10-15

AI Summary

To address privacy leakage, high latency, and severe resource constraints in distributed machine learning training on ultra-low-power microcontroller units (MCUs), this paper proposes the first distributed framework enabling pretraining-free, end-to-end training of efficient neural networks on MCU clusters. The framework integrates a lightweight distributed optimization algorithm with a customized wireless multi-hop communication protocol to enable collaborative parallel training. Key innovations include gradient sparsification, local computation offloading, and communication-computation overlap—collectively minimizing per-node memory footprint, energy consumption, and latency. Evaluated on a 20-node real hardware testbed, our approach achieves a 2× improvement in classification accuracy over state-of-the-art methods, while reducing memory usage, end-to-end latency, and energy consumption by up to 90%.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Search and Optimization: Distributed Search
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Security and Privacy: Security and privacy of machine learning and AI applications

Abstract

As Machine Learning (ML) becomes integral to Cyber-Physical Systems (CPS), there is growing interest in shifting training from traditional cloud-based to on-device processing (TinyML), for example, due to privacy and latency concerns. However, CPS often comprise ultra-low-power microcontrollers, whose limited compute resources make training challenging. This paper presents RockNet, a new TinyML method tailored for ultra-low-power hardware that achieves state-of-the-art accuracy in timeseries classification, such as fault or malware detection, without requiring offline pretraining. By leveraging that CPS consist of multiple devices, we design a distributed learning method that integrates ML and wireless communication. RockNet leverages all devices for distributed training of specialized compute efficient classifiers that need minimal communication overhead for parallelization. Combined with tailored and efficient wireless multi-hop communication protocols, our approach overcomes the communication bottleneck that often occurs in distributed learning. Hardware experiments on a testbed with 20 ultra-low-power devices demonstrate RockNet's effectiveness. It successfully learns timeseries classification tasks from scratch, surpassing the accuracy of the latest approach for neural network microcontroller training by up to 2x. RockNet's distributed ML architecture reduces memory, latency and energy consumption per device by up to 90 % when scaling from one central device to 20 devices. Our results show that a tight integration of distributed ML, distributed computing, and communication enables, for the first time, training on ultra-low-power hardware with state-of-the-art accuracy.

Problem

• Enabling distributed machine learning on ultra-low-power microcontroller devices
• Overcoming communication bottlenecks in distributed learning for CPS
• Achieving state-of-the-art timeseries classification accuracy without pretraining

Innovation

• Distributed learning method integrates ML with wireless communication
• Specialized efficient classifiers minimize communication overhead for parallelization
• Tailored wireless protocols overcome distributed learning communication bottlenecks