NorMatch: Matching Normalizing Flows with Discriminative Classifiers for Semi-Supervised Learning
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2022
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866916127142051840 |
|---|---|
| author | Deng, Zhongying Ke, Rihuan Schonlieb, Carola-Bibiane Aviles-Rivero, Angelica I |
| author_facet | Deng, Zhongying Ke, Rihuan Schonlieb, Carola-Bibiane Aviles-Rivero, Angelica I |
| contents | Semi-Supervised Learning (SSL) aims to learn a model using a tiny labeled set and massive amounts of unlabeled data. To better exploit the unlabeled data the latest SSL methods use pseudo-labels predicted from a single discriminative classifier. However, the generated pseudo-labels are inevitably linked to inherent confirmation bias and noise which greatly affects the model performance. In this work we introduce a new framework for SSL named NorMatch. Firstly, we introduce a new uncertainty estimation scheme based on normalizing flows, as an auxiliary classifier, to enforce highly certain pseudo-labels yielding a boost of the discriminative classifiers. Secondly, we introduce a threshold-free sample weighting strategy to exploit better both high and low confidence pseudo-labels. Furthermore, we utilize normalizing flows to model, in an unsupervised fashion, the distribution of unlabeled data. This modelling assumption can further improve the performance of generative classifiers via unlabeled data, and thus, implicitly contributing to training a better discriminative classifier. We demonstrate, through numerical and visual results, that NorMatch achieves state-of-the-art performance on several datasets. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2211_09593 |
| institution | arXiv |
| publishDate | 2022 |
| record_format | arxiv |
| spellingShingle | NorMatch: Matching Normalizing Flows with Discriminative Classifiers for Semi-Supervised Learning Deng, Zhongying Ke, Rihuan Schonlieb, Carola-Bibiane Aviles-Rivero, Angelica I Computer Vision and Pattern Recognition Semi-Supervised Learning (SSL) aims to learn a model using a tiny labeled set and massive amounts of unlabeled data. To better exploit the unlabeled data the latest SSL methods use pseudo-labels predicted from a single discriminative classifier. However, the generated pseudo-labels are inevitably linked to inherent confirmation bias and noise which greatly affects the model performance. In this work we introduce a new framework for SSL named NorMatch. Firstly, we introduce a new uncertainty estimation scheme based on normalizing flows, as an auxiliary classifier, to enforce highly certain pseudo-labels yielding a boost of the discriminative classifiers. Secondly, we introduce a threshold-free sample weighting strategy to exploit better both high and low confidence pseudo-labels. Furthermore, we utilize normalizing flows to model, in an unsupervised fashion, the distribution of unlabeled data. This modelling assumption can further improve the performance of generative classifiers via unlabeled data, and thus, implicitly contributing to training a better discriminative classifier. We demonstrate, through numerical and visual results, that NorMatch achieves state-of-the-art performance on several datasets. |
| title | NorMatch: Matching Normalizing Flows with Discriminative Classifiers for Semi-Supervised Learning |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2211.09593 |