Technical Advisor, Preferred Networks, Inc.
2 papers at NeurIPS 2025
We present a scheme of learning condition transfer achieving pairwise optimal transport, from a generic dataset of from {(x, c)} where a sample corresponding to a specific c is unique
Moving beyond $L_p$ geometric structure, we propose novel efficient Orlicz-Sobolev approaches (i.e., Orlicz-EPT, and Orlicz-Sobolev transport) for measures on a graph, possibly having different total mass.
