AP-OOD: Attention Pooling for Out-of-Distribution Detection

Claus Hofmann; Christian Huber; Bernhard Lehner; Daniel Klotz; Sepp Hochreiter; Werner Zellinger

AP-OOD: Attention Pooling for Out-of-Distribution Detection

Claus Hofmann
, Christian Huber
, Bernhard Lehner
, Daniel Klotz
, Sepp Hochreiter
, Werner Zellinger

Research output: Chapter in Book/Report/Conference proceeding › Conference proceedings › peer-review

Abstract

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 27.77% to 5.91% on XSUM summarization, and from 75.19% to 68.13% on WMT15 En–Fr translation.

Original language	English
Title of host publication	EurIPS 2025 Workshop on Metacognition in Generative AI
Number of pages	25
Publication status	Published - 2025

Fields of science

101019 Stochastics
102003 Image processing
103029 Statistical physics
101018 Statistics
101017 Game theory
102001 Artificial intelligence
202017 Embedded systems
101016 Optimisation
101015 Operations research
101014 Numerical mathematics
101029 Mathematical statistics
101028 Mathematical modelling
101026 Time series analysis
101024 Probability theory
102032 Computational intelligence
102004 Bioinformatics
102013 Human-computer interaction
101027 Dynamical systems
305907 Medical statistics
101004 Biomathematics
305905 Medical informatics
101031 Approximation theory
102033 Data mining
102 Computer Sciences
305901 Computer-aided diagnosis and therapy
102019 Machine learning
106007 Biostatistics
102018 Artificial neural networks
106005 Bioinformatics
202037 Signal processing
202036 Sensor systems
202035 Robotics

JKU Focus areas

Digital Transformation

Access to Document

https://openreview.net/forum?id=swqwc6h1P7

Cite this

@inproceedings{f3f6964628334be29bf1aa88dec9ce1e,

title = "AP-OOD: Attention Pooling for Out-of-Distribution Detection",

abstract = "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 27.77\% to 5.91\% on XSUM summarization, and from 75.19\% to 68.13\% on WMT15 En–Fr translation.",

author = "Claus Hofmann and Christian Huber and Bernhard Lehner and Daniel Klotz and Sepp Hochreiter and Werner Zellinger",

year = "2025",

language = "English",

booktitle = "EurIPS 2025 Workshop on Metacognition in Generative AI",

}

TY - GEN

T1 - AP-OOD: Attention Pooling for Out-of-Distribution Detection

AU - Hofmann, Claus

AU - Huber, Christian

AU - Lehner, Bernhard

AU - Klotz, Daniel

AU - Hochreiter, Sepp

AU - Zellinger, Werner

PY - 2025

Y1 - 2025

N2 - 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 27.77% to 5.91% on XSUM summarization, and from 75.19% to 68.13% on WMT15 En–Fr translation.

AB - 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 27.77% to 5.91% on XSUM summarization, and from 75.19% to 68.13% on WMT15 En–Fr translation.

M3 - Conference proceedings

BT - EurIPS 2025 Workshop on Metacognition in Generative AI

ER -