Geometrically Aligned Transfer Encoder for Inductive Transfer in Regression Tasks
This work addresses a gap in transfer learning for regression tasks, particularly in domains like molecular graphs, though it appears incremental as it adapts geometric concepts to a new problem type.
The paper tackles the lack of transfer learning methods for regression tasks by proposing a novel technique based on differential geometry, demonstrating that it outperforms conventional methods and shows stable behavior on molecular graph datasets.
Transfer learning is a crucial technique for handling a small amount of data that is potentially related to other abundant data. However, most of the existing methods are focused on classification tasks using images and language datasets. Therefore, in order to expand the transfer learning scheme to regression tasks, we propose a novel transfer technique based on differential geometry, namely the Geometrically Aligned Transfer Encoder (GATE). In this method, we interpret the latent vectors from the model to exist on a Riemannian curved manifold. We find a proper diffeomorphism between pairs of tasks to ensure that every arbitrary point maps to a locally flat coordinate in the overlapping region, allowing the transfer of knowledge from the source to the target data. This also serves as an effective regularizer for the model to behave in extrapolation regions. In this article, we demonstrate that GATE outperforms conventional methods and exhibits stable behavior in both the latent space and extrapolation regions for various molecular graph datasets.