Diversity-enhancing Generative Network for Few-shot Hypothesis Adaptation

TL;DR

This paper introduces DEG-Net, a generative network that enhances diversity in unlabeled data generation for few-shot hypothesis adaptation, leading to improved classifier performance in target domains.

Contribution

The paper proposes a novel diversity-enhancing generative network using HSIC to generate diverse unlabeled data, improving FHA performance over existing methods.

Findings

DEG-Net outperforms existing FHA baselines.

Diverse generated data improves adaptation effectiveness.

HSIC effectively promotes data diversity.

Abstract

Generating unlabeled data has been recently shown to help address the few-shot hypothesis adaptation (FHA) problem, where we aim to train a classifier for the target domain with a few labeled target-domain data and a well-trained source-domain classifier (i.e., a source hypothesis), for the additional information of the highly-compatible unlabeled data. However, the generated data of the existing methods are extremely similar or even the same. The strong dependency among the generated data will lead the learning to fail. In this paper, we propose a diversity-enhancing generative network (DEG-Net) for the FHA problem, which can generate diverse unlabeled data with the help of a kernel independence measure: the Hilbert-Schmidt independence criterion (HSIC). Specifically, DEG-Net will generate data via minimizing the HSIC value (i.e., maximizing the independence) among the semantic features of the generated data. By DEG-Net, the generated unlabeled data are more diverse and more effective for addressing the FHA problem. Experimental results show that the DEG-Net outperforms existing FHA baselines and further verifies that generating diverse data plays a vital role in addressing the FHA problem