On-Device Training of Machine Learning Models on Microcontrollers with Federated Learning
| Autor: | Nil Llisterri Giménez, Marc Monfort Grau, Roger Pueyo Centelles, Felix Freitag |
|---|---|
| Přispěvatelé: | Universitat Politècnica de Catalunya. Departament d'Arquitectura de Computadors, Universitat Politècnica de Catalunya. CNDS - Xarxes de Computadors i Sistemes Distribuïts |
| Jazyk: | angličtina |
| Rok vydání: | 2022 |
| Předmět: |
Informàtica::Intel·ligència artificial::Aprenentatge automàtic [Àrees temàtiques de la UPC]
Computer Networks and Communications Embedded systems Federated learning Ordinadors immersos Sistemes d' Embedded computer systems Keyword spotting Hardware and Architecture Control and Systems Engineering Signal Processing Machine learning Aprenentatge automàtic machine learning keyword spotting embedded systems federated learning Electrical and Electronic Engineering Informàtica::Arquitectura de computadors [Àrees temàtiques de la UPC] |
| Zdroj: | Electronics; Volume 11; Issue 4; Pages: 573 UPCommons. Portal del coneixement obert de la UPC Universitat Politècnica de Catalunya (UPC) |
| ISSN: | 2079-9292 |
| DOI: | 10.3390/electronics11040573 |
| Popis: | Recent progress in machine learning frameworks has made it possible to now perform inference with models using cheap, tiny microcontrollers. Training of machine learning models for these tiny devices, however, is typically done separately on powerful computers. This way, the training process has abundant CPU and memory resources to process large stored datasets. In this work, we explore a different approach: training the machine learning model directly on the microcontroller and extending the training process with federated learning. We implement this approach for a keyword spotting task. We conduct experiments with real devices to characterize the learning behavior and resource consumption for different hyperparameters and federated learning configurations. We observed that in the case of training locally with fewer data, more frequent federated learning rounds more quickly reduced the training loss but involved a cost of higher bandwidth usage and longer training time. Our results indicate that, depending on the specific application, there is a need to determine the trade-off between the requirements and the resource usage of the system. This work has received funding from the Spanish Government under contracts PID2019-106774RB-C21, PCI2019-111851-2 (LeadingEdge CHIST-ERA), PCI2019-111850-2 (DiPET CHIST-ERA). |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|