New sensor grouping method improves federated learning in internet of things
Pod-wise Multimodal Federated Learning for Heterogeneous IoT
Networking and Internet Architecture
Summary
Federated learning lets many devices learn together without sharing their raw data. The problem is that IoT devices have different sensors, and they don’t always have the same sensors active, so mixing all their data blindly can cause confusion. The authors created a new method that groups sensors on devices into 'pods' and only merges information from devices that share the same sensors, while keeping some parts common to all. This approach led to better recognition of activities across various datasets and reduced communication between devices.
What this means in practice
- •For iot system designers: Implement sensor-specific model components on IoT devices to improve federated learning accuracy despite differing sensor setups.
- •For mobile device manufacturers: Use sensor-grouped federated learning models to optimize activity recognition features while reducing communication costs across devices with varied sensors.$Commercial implications: Enables better and more efficient health or fitness tracking products by leveraging federated learning tailored to device sensor sets.
Authors
Joy Okolo, Zihao Ding, Jun Huang
Abstract
Federated learning (FL) is well suited to Internet of Things (IoT) sensing. However, devices carry different sensors and come online in periodic patterns, and as such, the sensors among the participants change from round to round. Under naive averaging, devices without a sensor overwrite its parameters, and a single-sensor device never learns how its sensor relates to the others. Existing methods either redesign local training for mismatched sensors (assuming always-available devices) or select clients without regard to their sensors. We propose Tier-aware Intra-pod Decoupled Encoders (TIDE). Specifically, each sensor has its own encoder and auxiliary head, whose parameters are merged only within the "pod" (i.e., the devices that hold the sensor), while the fusion layer and the task head remain shared by all devices. On three multimodal activity recognition datasets against eight baselines, TIDE reaches 78.06% on ActionSense, 7.63 points above the next method, is the most accurate on all three datasets at half its own communication, and at equal communication leads on ActionSense and is ahead within seed variation on the other two.