Summary of Deep Positive-unlabeled Anomaly Detection For Contaminated Unlabeled Data, by Hiroshi Takahashi et al.
Deep Positive-Unlabeled Anomaly Detection for Contaminated Unlabeled Data
by Hiroshi Takahashi, Tomoharu Iwata, Atsutoshi Kumagai, Yuuki Yamanaka
First submitted to arxiv on: 29 May 2024
Categories
- Main: Machine Learning (stat.ML)
- Secondary: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The proposed deep positive-unlabeled anomaly detection framework integrates positive-unlabeled learning with deep anomaly detection models, such as autoencoders and deep support vector data descriptions. This framework enables the approximation of anomaly scores for normal data using unlabeled data and labeled anomaly data. The approach trains anomaly detectors by minimizing anomaly scores for normal data while maximizing those for labeled anomaly data. Experimental results on various datasets demonstrate that this method achieves better detection performance than existing approaches. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper proposes a new way to detect anomalies in data, which is important because it helps us identify unusual patterns or outliers in large datasets. Currently, most methods assume that most data is normal and only use labeled data (where we know what’s an anomaly) to train the detector. However, this can be tricky because the unlabeled data might actually contain some anomalies too. The proposed method uses both labeled and unlabeled data together to create a better detector. It does this by learning how to score the likelihood of each piece of data being normal or an anomaly. This approach performs better than existing methods on various datasets. |
Keywords
» Artificial intelligence » Anomaly detection » Likelihood
