Abstract:Out-of-distribution (OOD) detection is essential for model trustworthiness which aims to sensitively identify semantic OOD samples and robustly generalize for covariate-shifted OOD samples. However, we discover that the superior OOD detection performance of state-of-the-art methods is achieved by secretly sacrificing the OOD generalization ability. Specifically, the classification accuracy of these models could deteriorate dramatically when they encounter even minor noise. This phenomenon contradicts the goal of model trustworthiness and severely restricts their applicability in real-world scenarios. What is the hidden reason behind such a limitation? In this work, we theoretically demystify the ``\textit{sensitive-robust}'' dilemma that lies in many existing OOD detection methods. Consequently, a theory-inspired algorithm is induced to overcome such a dilemma. By decoupling the uncertainty learning objective from a Bayesian perspective, the conflict between OOD detection and OOD generalization is naturally harmonized and a dual-optimal performance could be expected. Empirical studies show that our method achieves superior performance on standard benchmarks. To our best knowledge, this work is the first principled OOD detection method that achieves state-of-the-art OOD detection performance without compromising OOD generalization ability. Our code is available at \href{<a class="link-external link-https" href="https://github.com/QingyangZhang/DUL" rel="external noopener nofollow">this https URL</a>}{<a class="link-external link-https" href="https://github.com/QingyangZhang/DUL" rel="external noopener nofollow">this https URL</a>}.

Rethinking Test-time Likelihood: The Likelihood Path Principle and Its Application to OOD Detection

A Geometric Explanation of the Likelihood OOD Detection Paradox

Resultant: Incremental Effectiveness on Likelihood for Unsupervised Out-of-Distribution Detection

Combining Model Checking and Testing with an Application to Reliability Prediction and Distribution

Your Classifier Can Be Secretly a Likelihood-Based OOD Detector

ODIM: Outlier Detection via Likelihood of Under-Fitted Generative Models

Test-Time Linear Out-of-Distribution Detection

The Best of Both Worlds: On the Dilemma of Out-of-distribution Detection

A Simple Test-Time Method for Out-of-Distribution Detection

Rethinking Out-of-Distribution Detection for Reinforcement Learning: Advancing Methods for Evaluation and Detection

DICE: Detecting In-distribution Contamination in LLM's Fine-tuning Phase for Math Reasoning

Opportunistic Detection Rules: Finite and Asymptotic Analysis.

Your Finetuned Large Language Model is Already a Powerful Out-of-distribution Detector

FLatS: Principled Out-of-Distribution Detection with Feature-Based Likelihood Ratio Score

The Open-World Lottery Ticket Hypothesis for OOD Intent Classification

A View on Out-of-Distribution Identification from a Statistical Testing Theory Perspective

Comprehensive OOD Detection Improvements

Rethinking Out-of-Distribution Detection on Imbalanced Data Distribution

OAL: Enhancing OOD Detection Using Latent Diffusion

Dissecting Out-of-Distribution Detection and Open-Set Recognition: A Critical Analysis of Methods and Benchmarks

Rethinking Out-of-Distribution Detection From a Human-Centric Perspective