Tight Rates in Supervised Outlier Transfer Learning
Mohammadreza M. Kalan, Samory Kpotufe
Abstract
A critical barrier to learning an accurate decision rule for outlier detection is the scarcity of outlier data. As such, practitioners often turn to the use of similar but imperfect outlier data from which they might transfer information to the target outlier detection task. Despite the recent empirical success of transfer learning approaches in outlier detection, a fundamental understanding of when and how knowledge can be transferred from a source to a target outlier detection task remains elusive. In this work, we adopt the traditional framework of Neyman-Pearson classification -- which formalizes supervised outlier detection -- with the added assumption that one has access to some related but imperfect outlier data. Our main results are as follows: We first determine the information-theoretic limits of the problem under a measure of discrepancy that extends some existing notions from traditional balanced classification; interestingly, unlike in balanced classification, seemingly very dissimilar sources can provide much information about a target, thus resulting in fast transfer. We then show that, in principle, these information-theoretic limits are achievable by adaptive procedures, i.e., procedures with no a priori information on the discrepancy between source and target outlier distributions.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
1 top-tier paper has cited it since. Know what they found before your reviewers do.
Lune can draw on
the full paper · the top-tier papers that cite it · related work
Free to start. No credit card required.
Lineage
- Neyman-Pearson Classification under Both Null and Alternative Distributions Shift cites Minimax Lower Bounds for Transfer Learning with Linear and One-hidden Layer Neural Networks
- Near-Optimal Linear Regression under Distribution Shift cites Minimax Lower Bounds for Transfer Learning with Linear and One-hidden Layer Neural Networks
Cited by top-tier papers1
Builds on2
- Near-Optimal Linear Regression under Distribution ShiftQi Lei, Wei Hu, Jason D. LeeICML 2021 · 45 citations
- Minimax Lower Bounds for Transfer Learning with Linear and One-hidden Layer Neural NetworksSeyed Mohammadreza Mousavi Kalan, Zalan Fabian, Salman Avestimehr, Mahdi SoltanolkotabiNeurIPS 2020 · 37 citations
Related papers
- Transfer Learning for Anomaly Detection through Localized and Unsupervised Instance SelectionVincent Vercruyssen, Wannes Meert, Jesse DavisAAAI 2020 · 35 citations
- Automatic Unsupervised Outlier Model SelectionYue Zhao, Ryan A. Rossi, Leman AkogluNeurIPS 2021 · 104 citations
- Mutual Transfer Learning for Massive DataChing-Wei Cheng, Xingye Qiao, Guang ChengICML 2020 · 1 citation
- Automatic Unsupervised Ensemble Outlier Model SelectionHong-Phuc Phan, Tuan-Anh Vu, Tung Kieu, Sơn Hà Xuân et al.ICML 2026
- Exploiting Mixed Unlabeled Data for Detecting Samples of Seen and Unseen Out-of-Distribution ClassesYi-Xuan Sun, Wei WangAAAI 2022 · 5 citations
