analiza aiqbench
This paper benchmarks VAE, GAN, and diffusion-based anomaly detectors in federated predictive maintenance across five clients and two industrial time-series datasets. Its main finding is a tradeoff: full federation helps GANs but they stay less robust than VAE and DDPM variants, while sharing only decoder parameters lets diffusion models match full federation at a much smaller communication cost. Results are limited to two benchmarks and a cross-silo setup, so practitioners should treat them as guidance for their own IoT deployments rather than general rules.
On the Tradeoffs of On-Device Generative Models in Federated Predictive Maintenance Systems
arXiv:2605.07860v2 Announce Type: replace-cross Abstract: Federated Learning (FL) has emerged as a promising paradigm for preserving client data ownership and control over distributed Internet of Things (IoT) environments. While discriminative models dominate most FL use cases, recent advances in generative models - such as Variational Autoencoders (VAE), Generative Adversarial Networks (GAN), and Diffusion Models (DM) - offer new opportunities for unsupervised anomaly detection in time series analysis, with relevant applications in predictive maintenance (PdM) in critical industrial infrastructures. In this work, we present a comprehensive analysis of VAEs, GANs, and DMs in the context of federated PdM. We analyze their performance and communication overhead under both full and partial federation setups, where only subsets of model components are shared. Building on split-parameter personalized FL, this paper provides an empirical instantiation of shared/private parameter partitioning for unsupervised generative Time-Series Anomaly Detection (TSAD), mapping the split to analysis and synthesis modules across the generative architectures. Our experiments over five clients and two industrial time-series benchmarks reveal distinct trade-offs in model utility, stability, and communication efficiency in heterogeneous cross-silo FL. For the evaluated GAN-based configurations, full federation improves final anomaly-detection utility relative to independent local training, although these models remain less robust than the VAE- and DDPM-based alternatives. For DMs, decoder sharing matches full federation while substantially reducing the communication payload.
Abstract:Federated Learning (FL) has emerged as a promising paradigm for preserving client data ownership and control over distributed Internet of Things (IoT) environments. While discriminative models dominate most FL use cases, recent advances in generative models - such as Variational Autoencoders (VAE), Generative Adversarial Networks (GAN), and Diffusion Models (DM) - offer new opportunities for unsupervised anomaly detection in time series analysis, with relevant applications in predictive maintenance (PdM) in critical industrial infrastructures. In this work, we present a comprehensive analysis of VAEs, GANs, and DMs in the context of federated PdM. We analyze their performance and communication overhead under both full and partial federation setups, where only subsets of model components are shared. Building on split-parameter personalized FL, this paper provides an empirical instantiation of shared/private parameter partitioning for unsupervised generative Time-Series Anomaly Detection (TSAD), mapping the split to analysis and synthesis modules across the generative architectures. Our experiments over five clients and two industrial time-series benchmarks reveal distinct trade-offs in model utility, stability, and communication efficiency in heterogeneous cross-silo FL. For the evaluated GAN-based configurations, full federation improves final anomaly-detection utility relative to independent local training, although these models remain less robust than the VAE- and DDPM-based alternatives. For DMs, decoder sharing matches full federation while substantially reducing the communication payload.
Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
arXiv:2605.07860 [cs.LG]
(or arXiv:2605.07860v2 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2605.07860 Focus to learn more arXiv-issued DOI via DataCite
https://doi.org/10.1109/JIOT.2026.3739066 Focus to learn more DOI(s) linking to related resources
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 318 cuvinte
Subiecte conexe
Breaking the Homogeneity Assumption: Specialized Multi-Generator Adversarial Learning for Rare Failure Detection in Predictive Maintenance
arXiv · 23 iul. 2026
How Trackunit turns construction data into decisions with AI
Databricks · 28 aug. 2026
Orchestrating Wheeled Mobile Robots Online without Conflicts
arXiv · 22 sept. 2026
Predicting Wind Turbine Power Using Machine Learning and Weather Forecasts
arXiv · 9 sept. 2026