Geometrically Aligned Transfer Encoder for Inductive Transfer in Regression Tasks

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Ko, Sung Moon, Lee, Sumin, Jeong, Dae-Woong, Lim, Woohyung, Han, Sehui
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915340890406912
author Ko, Sung Moon
Lee, Sumin
Jeong, Dae-Woong
Lim, Woohyung
Han, Sehui
author_facet Ko, Sung Moon
Lee, Sumin
Jeong, Dae-Woong
Lim, Woohyung
Han, Sehui
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2310_06369
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Geometrically Aligned Transfer Encoder for Inductive Transfer in Regression Tasks
Ko, Sung Moon
Lee, Sumin
Jeong, Dae-Woong
Lim, Woohyung
Han, Sehui
Artificial Intelligence
Machine Learning
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.
title Geometrically Aligned Transfer Encoder for Inductive Transfer in Regression Tasks
topic Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2310.06369