ICML Conference (2026)
Sung Moon Ko, Jaewan Lee, Sumin Lee, Soorin Yim, Sehui Han
Abstract
Geometrical interpretations of machine learning models offer insightful perspectives into their underlying mathematical structures. In this work, we introduce a novel approach that leverages differential geometry, particularly concepts from Riemannian geometry, to integrate multiple models into a unified transfer learning framework. By aligning the Ricci curvature of latent space of individual models, we construct an interrelated architecture, namely Geometric Embedding Alignment via Curvature Matching in Transfer Learning (GEAR), which ensures comprehensive geometric representation across datapoints. This framework enables the effective aggregation of knowledge from diverse sources, thereby improving performance on target tasks. We evaluate our model on 23 molecular dataset pairs sourced from various domains and demonstrate significant performance gains over existing benchmarks under both random (14.5%) and scaffold data (7.7%) splits.