U-MASK: User-adaptive Spatio-Temporal Masking for Personalized Mobile AI Applications

TL;DR

U-MASK is proposed, a user-adaptive spatio-temporal masking method that allocates evidence budgets based on user reliability and task sensitivity and model mobile behavior as a partially observed spatio-temporal tensor and unify short-term adaptation, long-horizon forecasting, and cold-start recommendation as a conditional completion problem.

Abstract

Personalized mobile artificial intelligence applications are widely deployed, yet they are expected to infer user behavior from sparse and irregular histories under a continuously evolving spatio-temporal context. This setting induces a fundamental tension among three requirements, i.e., immediacy to adapt to recent behavior, stability to resist transient noise, and generalization to support long-horizon prediction and cold-start users. Most existing approaches satisfy at most two of these requirements, resulting in an inherent impossibility triangle in data-scarce, non-stationary personalization. To address this challenge, we model mobile behavior as a partially observed spatio-temporal tensor and unify short-term adaptation, long-horizon forecasting, and cold-start recommendation as a conditional completion problem, where a user- and task-specific mask specifies which coordinates are treated as evidence. We propose U-MASK, a user-adaptive spatio-temporal masking method that allocates evidence budgets based on user reliability and task sensitivity. To enable mask generation under sparse observations, U-MASK learns a compact, task-agnostic user representation from app and location histories via U-SCOPE, which serves as the sole semantic conditioning signal. A shared diffusion transformer then performs mask-guided generative completion while preserving observed evidence, so personalization and task differentiation are governed entirely by the mask and the user representation. Experiments on real-world mobile datasets demonstrate consistent improvements over state-of-the-art methods across short-term prediction, long-horizon forecasting, and cold-start settings, with the largest gains under severe data sparsity. The code and dataset will be available at https://github.com/NICE-HKU/U-MASK.

Figures (9)

• Figure 1: The impossibility triangle of personalized mobile AI app usage behavior modeling. Each edge represents a shared modeling assumption that enables the two adjacent objectives while structurally excluding the third.
• Figure 2: Unified framework for personalized mobile AI behavior modeling.
• Figure 3: Overview of the U-SCOPE framework. Given sparse mobile application usage data, U-SCOPE infers semantic user representations that encode application preferences, location tendencies, and temporal regularities, enabling downstream task-oriented inference under data scarcity.
• Figure 4: The figure displays multi-dimensional user behavior characteristic comparisons across 7 datasets. Each subplot employs a polar coordinate system with six axes representing the six indicators: Location Diversity, Traffic Usage, App Diversity, Peak Hour Traffic Ratio, Residence Locations, and Active app Count. The colored area size indicates the overall activity level of user behavior. The legend in the lower right corner uses different colors to distinguish each dataset.
• Figure 5: Comparison of app usage distributions (a) Difference heatmap (Generated -- Original) with significant deviations ($\lvert \mathrm{diff} \rvert > 0.05$) circled. (b) Point-wise comparison showing strong correlation ($R^{2} = 0.9063$, $p < 0.001$) between original and generated data, indicating preservation of app--location relationships.