Problem
Research questions and friction points this paper is trying to address.
cross-city transfer
region correspondence
optimal transport
multi-source transfer
label scarcity
SCOT: Multi-Source Cross-City Transfer with Optimal-Transport Soft-Correspondence Objective
🤖 AI Summary
This work addresses the challenge of label scarcity in cross-city transfer learning caused by incompatible regional partitions and the absence of true correspondences. To tackle this, the authors propose an explicit soft correspondence modeling approach based on entropy-regularized optimal transport (OT). The method employs the Sinkhorn algorithm to learn interpretable soft matchings between regions of source and target cities and introduces target-guided prototype hubs as shared semantic anchors. Multi-source representation alignment is achieved through OT-weighted contrastive learning combined with cycle-consistent reconstruction regularization. Evaluated on multiple real-world urban datasets, the proposed framework significantly improves both transfer accuracy and robustness while offering an interpretable assessment of region alignment quality.
📝 Abstract
Cross-city transfer improves prediction in label-scarce cities by leveraging labeled data from other cities, but it becomes challenging when cities adopt incompatible partitions and no ground-truth region correspondences exist. Existing approaches either rely on heuristic region matching, which is often sensitive to anchor choices, or perform distribution-level alignment that leaves correspondences implicit and can be unstable under strong heterogeneity. We propose SCOT, a cross-city representation learning framework that learns explicit soft correspondences between unequal region sets via Sinkhorn-based entropic optimal transport. SCOT further sharpens transferable structure with an OT-weighted contrastive objective and stabilizes optimization through a cycle-style reconstruction regularizer. For multi-source transfer, SCOT aligns each source and the target to a shared prototype hub using balanced entropic transport guided by a target-induced prototype prior. Across real-world cities and tasks, SCOT consistently improves transfer accuracy and robustness, while the learned transport couplings and hub assignments provide interpretable diagnostics of alignment quality.
Innovation
Methods, ideas, or system contributions that make the work stand out.
optimal transport
cross-city transfer
soft correspondence
contrastive learning
multi-source adaptation
🔎 Similar Papers
No similar papers found.
