mlsanityData to 14 Oct 2026

Improving State-of-the-Art in One-Class Classification by Leveraging Unlabeled Data

arXiv 2203.07206 ↗First posted 14 Mar 2022cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

When dealing with binary classification of data with only one labeled class data scientists employ two main approaches, namely One-Class (OC) classification and Positive Unlabeled (PU) learning. The former only learns from labeled positive data, whereas the latter also utilizes unlabeled data to improve the overall performance. Since PU learning utilizes more data, we might be prone to think that when unlabeled data is available, the go-to algorithms should always come from the PU group. However, we find that this is not always the case if unlabeled data is unreliable, i.e. contains limited or biased latent negative data. We perform an extensive experimental study of a wide list of state-of-the-art OC and PU algorithms in various scenarios as far as unlabeled data reliability is concerned. Furthermore, we propose PU modifications of state-of-the-art OC algorithms that are robust to unreliable unlabeled data, as well as a guideline to similarly modify other OC algorithms. Our main practical recommendation is to use state-of-the-art PU algorithms when unlabeled data is reliable and to use the proposed modifications of state-of-the-art OC algorithms otherwise. Additionally, we outline procedures to distinguish the cases of reliable and unreliable unlabeled data using statistical tests.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 19 of its 66 references that are in the map. First posted on arXiv 14 Mar 2022.

Builds on

19 of 66 references.

References · 19

  1. 2018Deep Anomaly Detection with Outlier Exposurecited by 942
  2. 2020CSI: Novelty Detection via Contrastive Learning on Distributionally Shifted Instancescited by 441
  3. 2020A Unifying Review of Deep and Shallow Anomaly Detectioncited by 425
  4. 2019Deep Learning for Anomaly Detection: A Surveycited by 358
  5. 2019Deep Semi-Supervised Anomaly Detectioncited by 291
  6. 2017Positive-Unlabeled Learning with Non-Negative Risk Estimatorcited by 219
  7. 2018Learning from positive and unlabeled data: a surveycited by 219
  8. 2020DROCC: Deep Robust One-Class Classificationcited by 121
  9. 2021Understanding Failures in Out-of-Distribution Detection with Deep Generative Modelscited by 89
  10. 2018Anomaly Detection using One-Class Neural Networkscited by 82
  11. 2020Rethinking Assumptions in Deep Anomaly Detectioncited by 45
  12. 2016Theoretical Comparisons of Positive-Unlabeled Learning against Positive-Negative Learningcited by 41
7 more
  1. 2019One-Class Convolutional Neural Networkcited by 37
  2. 2020Learning from Positive and Unlabeled Data with Arbitrary Positive Shiftcited by 23
  3. 2016DNA-inspired online behavioral modeling and its application to spambot detectioncited by 16
  4. 2016Nonparametric semi-supervised learning of class proportionscited by 13
  5. 2018Positive-Unlabeled Classification under Class Prior Shift and Asymmetric Errorcited by 7
  6. 2021Positive-Unlabeled Classification under Class-Prior Shift: A Prior-invariant Approach Based on Density Ratio Estimationcited by 7
  7. 2019Revisiting Sample Selection Approach to Positive-Unlabeled Learning: Turning Unlabeled Data into Positive rather than Negativecited by 6

Built on by

Not cited yet.