Evaluating Federated Learning for Cross-Country Mood Inference from Smartphone Sensing Data

TL;DR

This paper tackles mood inference from smartphone sensing in a cross-country setting, addressing privacy, data heterogeneity, and non-IID distributions. It introduces FedFAP, a feature-aware personalized federated learning framework that separates shared representations learned across countries from client-specific local representations, enabling effective learning despite heterogeneous sensor modalities. Across six countries and 26 sensor modalities, FedFAP achieves an AUROC of 0.744 and generally outperforms centralized models and existing personalized FL baselines, with a 1D-CNN encoder providing the best performance among architectures. The work highlights the importance of population-aware personalization and privacy-preserving learning for scalable mood-aware mobile sensing, offering design guidance for deploying cross-country mood inference systems in real-world, privacy-conscious contexts.

Abstract

Mood instability is a key behavioral indicator of mental health, yet traditional assessments rely on infrequent and retrospective reports that fail to capture its continuous nature. Smartphone-based mobile sensing enables passive, in-the-wild mood inference from everyday behaviors; however, deploying such systems at scale remains challenging due to privacy constraints, uneven sensing availability, and substantial variability in behavioral patterns. In this work, we study mood inference using smartphone sensing data in a cross-country federated learning setting, where each country participates as an independent client while retaining local data. We introduce FedFAP, a feature-aware personalized federated framework designed to accommodate heterogeneous sensing modalities across regions. Evaluations across geographically and culturally diverse populations show that FedFAP achieves an AUROC of 0.744, outperforming both centralized approaches and existing personalized federated baselines. Beyond inference, our results offer design insights for mood-aware systems, demonstrating how population-aware personalization and privacy-preserving learning can enable scalable and mood-aware mobile sensing technologies.

Figures (6)

• Figure 1: Distribution of samples across five label classes.
• Figure 2: Distribution of samples across three label classes.
• Figure 3: Distribution of feature missingness across countries, in which the marked areas denote the number of features belonging to different missingness ranges.
• Figure 4: Proposed architecture comprising of a global server and N clients.
• Figure 5: AUROC scores across different Federated averaging methods at 10 epochs and 50 rounds.