Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces Convex Distance Operator Transport (CDOT), a novel convex optimal transport framework that effectively aligns distributions across heterogeneous domains while preserving both feature correspondence and intrinsic geometric structures. By employing an operator-based regularization, CDOT enhances robustness against local geometric variations and establishes a valid pseudometric for attributed compact metric-measure spaces. The authors also elucidate the relationship between CDOT and Gromov-Wasserstein transport, highlighting the geometric non-convexity of the latter and demonstrating superior performance of CDOT through experiments on various datasets.
CDOT not only preserves geometric structures across domains but also outperforms existing methods in robustness and reliability on diverse datasets.
We introduce Convex Distance Operator Transport (CDOT), the first convex optimal transport framework that aligns distributions across heterogeneous domains by jointly preserving feature correspondence and intrinsic geometric structure. Specifically, CDOT employs an operator-based regularization that aligns aggregated distance structures by introducing distance and conditional expectation operators. Consequently, the proposed regularization improves the robustness to local geometric variations. We further prove that the resulting CDOT discrepancy is a valid pseudometric on the space of attributed compact metric-measure spaces. In addition, we characterize the relationship between CDOT and Gromov--Wasserstein (GW) through a new notion of dispersion gap, formally elucidating the geometric source of non-convexity in GW compared to the convexity of CDOT. In the finite-sample regime, we derive a non-asymptotic risk bound decomposed into optimization and statistical errors, establishing risk consistency under a globally convergent Frank--Wolfe algorithm. Experiments on synthetic point clouds, brain connectomes, and graph classification benchmarks demonstrate better performance over existing methods, with stable and reliable behavior in practice.