Adaptive learning improves urban data fusion for smarter city insights
Rethinking Multi-View Fusion for Urban Representation Learning
Machine LearningArtificial Intelligence
Summary
Urban areas generate many different types of data, like people’s movement, places to visit, and land use. Existing methods try to combine these data types evenly but don't work well because some data matters more in some places than others. The authors show that current ways cause unstable learning and biased results. They created PRISM, a new system that adjusts how it mixes data depending on each area and controls training to avoid conflicts. This approach leads to more balanced and stable urban data representations.
What this means in practice
- •For city planners: Enhance predictions about urban phenomena by tailoring data fusion to specific neighborhoods for more accurate service demand or crime forecasting.
- •For urban analytics teams: Improve stability and balance in models that integrate multiple data types, reducing bias in urban insights for diversified city regions.
Authors
Sean Bin Yang, Zongyi Xu, Jilin Hu, Ying Sun, Kristian Torp, Bin Yang, Torben Bach Pedersen, Christian S. Jensen
Abstract
Urban region representation learning encodes heterogeneous urban signals, such as mobility flows, points of interest, and land-use patterns, into unified embeddings for downstream tasks, including check-in prediction, crime forecasting, and service demand estimation. Existing approaches rely on globally shared fusion and joint optimization, implicitly assuming uniform modality importance and naturally balanced multi-objective training. However, urban data is inherently heterogeneous. Our empirical analysis reveals two structural issues: (i) modality utility varies markedly across regions, causing static fusion to overemphasize signals from certain modalities; and (ii) heterogeneous objectives exhibit persistent gradient imbalance, causing representational bias and unstable training. We present PRISM, a self-adjusting framework for adaptive multi-view learning. PRISM adapts sparse MoE routing to learn region-specific view allocations, while a PID (Proportional Integral Derivative)-inspired regulation mechanism mitigates gradient conflicts during optimization. Experiments on three large-scale urban datasets show that PRISM achieves competitive performance compared with strong multi-view baselines, while providing more stable optimization behavior and more balanced region representations.