Source-Target Similarity Modelings for Multi-Source Transfer Gaussian Process Regression
Pengfei Wei, Ramon Sagarna, Yiping Ke, Yew-Soon Ong, Chi-Keong Goh
Abstract
A key challenge in multi-source transfer learning is to capture the diverse inter-domain similarities. In this paper, we study different approaches based on Gaussian process models to solve the multi-source transfer regression problem. Precisely, we first investigate the feasibility and performance of a family of transfer covariance functions that represent the pairwise similarity of each source and the target domain. We theoretically show that using such a transfer covariance function for general Gaussian process modelling can only capture the same similarity coefficient for all the sources, and thus may result in unsatisfactory transfer performance. This leads us to propose
BibTeX
@InProceedings{pmlr-v70-wei17a,
title = {Source-Target Similarity Modelings for Multi-Source Transfer {G}aussian Process Regression},
author = {Pengfei Wei and Ramon Sagarna and Yiping Ke and Yew-Soon Ong and Chi-Keong Goh},
booktitle = {Proceedings of the 34th International Conference on Machine Learning},
pages = {3722--3731},
year = {2017},
editor = {Precup, Doina and Teh, Yee Whye},
volume = {70},
series = {Proceedings of Machine Learning Research},
month = {06--11 Aug},
publisher = {PMLR},
pdf = {http://proceedings.mlr.press/v70/wei17a/wei17a.pdf},
url = {https://proceedings.mlr.press/v70/wei17a.html},
abstract = {A key challenge in multi-source transfer learning is to capture the diverse inter-domain similarities. In this paper, we study different approaches based on Gaussian process models to solve the multi-source transfer regression problem. Precisely, we first investigate the feasibility and performance of a family of transfer covariance functions that represent the pairwise similarity of each source and the target domain. We theoretically show that using such a transfer covariance function for general Gaussian process modelling can only capture the same similarity coefficient for all the sources, and thus may result in unsatisfactory transfer performance. This leads us to propose