Journal
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
Volume -, Issue -, Pages -Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TNNLS.2022.3209085
Keywords
Uncertainty; Data models; Training; Task analysis; Costs; Training data; Supervised learning; Active learning (AL); data bias; deep learning; feature fusion; feature matching; neural network; uncertainty
Categories
Funding
- National Natural Science Foundation of China [62272298, 61872241, 62077037]
- Shanghai Municipal Science and Technology Major Project [2021SHZDZX0102]
- National Key Research and Development Program of China [2019YFB1703600]
- Hong Kong Polytechnic University [P0030419, P0042740, P0035358]
Ask authors/readers for more resources
This article proposes a feature-matching-based uncertainty method to alleviate the data bias issue by resampling selected uncertainty data. To meet the requirement of no additional costs, the authors specially designed an unsupervised fusion feature matching (UFFM), and redesigned classic uncertainty methods for more complex visual tasks. Experimental results demonstrate that the proposed method outperforms other similar techniques.
Active learning (AL) aims to sample the most valuable data for model improvement from the unlabeled pool. Traditional works, especially uncertainty-based methods, are prone to suffer from a data bias issue, which means that selected data cannot cover the entire unlabeled pool well. Although there have been lots of literature works focusing on this issue recently, they mainly benefit from the huge additional training costs and the artificially designed complex loss. The latter causes these methods to be redesigned when facing new models or tasks, which is very time-consuming and laborious. This article proposes a feature-matching-based uncertainty that resamples selected uncertainty data by feature matching, thus removing similar data to alleviate the data bias issue. To ensure that our proposed method does not introduce a lot of additional costs, we specially design a unsupervised fusion feature matching (UFFM), which does not require any training in our novel AL framework. Besides, we also redesign several classic uncertainty methods to be applied to more complex visual tasks. We conduct rigorous experiments on lots of standard benchmark datasets to validate our work. The experimental results show that our UFFM is better than the similar unsupervised feature matching technologies, and our proposed uncertainty calculation method outperforms random sampling, classic uncertainty approaches, and recent state-of-the-art (SOTA) uncertainty approaches.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available