AP-OOD: Attention Pooling for Out-of-Distribution Detection
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866918324758118400 |
|---|---|
| author | Hofmann, Claus Huber, Christian Lehner, Bernhard Klotz, Daniel Hochreiter, Sepp Zellinger, Werner |
| author_facet | Hofmann, Claus Huber, Christian Lehner, Bernhard Klotz, Daniel Hochreiter, Sepp Zellinger, Werner |
| contents | Out-of-distribution (OOD) detection, which maps high-dimensional data into a scalar OOD score, is critical for the reliable deployment of machine learning models. A key challenge in recent research is how to effectively leverage and aggregate token embeddings from language models to obtain the OOD score. In this work, we propose AP-OOD, a novel OOD detection method for natural language that goes beyond simple average-based aggregation by exploiting token-level information. AP-OOD is a semi-supervised approach that flexibly interpolates between unsupervised and supervised settings, enabling the use of limited auxiliary outlier data. Empirically, AP-OOD sets a new state of the art in OOD detection for text: in the unsupervised setting, it reduces the FPR95 (false positive rate at 95% true positives) from 27.84% to 4.67% on XSUM summarization, and from 77.08% to 70.37% on WMT15 En-Fr translation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_06031 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | AP-OOD: Attention Pooling for Out-of-Distribution Detection Hofmann, Claus Huber, Christian Lehner, Bernhard Klotz, Daniel Hochreiter, Sepp Zellinger, Werner Machine Learning Out-of-distribution (OOD) detection, which maps high-dimensional data into a scalar OOD score, is critical for the reliable deployment of machine learning models. A key challenge in recent research is how to effectively leverage and aggregate token embeddings from language models to obtain the OOD score. In this work, we propose AP-OOD, a novel OOD detection method for natural language that goes beyond simple average-based aggregation by exploiting token-level information. AP-OOD is a semi-supervised approach that flexibly interpolates between unsupervised and supervised settings, enabling the use of limited auxiliary outlier data. Empirically, AP-OOD sets a new state of the art in OOD detection for text: in the unsupervised setting, it reduces the FPR95 (false positive rate at 95% true positives) from 27.84% to 4.67% on XSUM summarization, and from 77.08% to 70.37% on WMT15 En-Fr translation. |
| title | AP-OOD: Attention Pooling for Out-of-Distribution Detection |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2602.06031 |