Exploring Large Language Models for Multi-Modal Out-of-Distribution Detection
Not on the map. Why?
Abstract (arXiv)
Out-of-distribution (OOD) detection is essential for reliable and trustworthy machine learning. Recent multi-modal OOD detection leverages textual information from in-distribution (ID) class names for visual OOD detection, yet it currently neglects the rich contextual information of ID classes. Large language models (LLMs) encode a wealth of world knowledge and can be prompted to generate descriptive features for each class. Indiscriminately using such knowledge causes catastrophic damage to OOD detection due to LLMs' hallucinations, as is observed by our analysis. In this paper, we propose to apply world knowledge to enhance OOD detection performance through selective generation from LLMs. Specifically, we introduce a consistency-based uncertainty calibration method to estimate the confidence score of each generation. We further extract visual objects from each image to fully capitalize on the aforementioned world knowledge. Extensive experiments demonstrate that our method consistently outperforms the state-of-the-art.
Key facts. Cited by 12 papers in the map (data through 7 Oct 2026). Builds on 46 of its 64 references that are in the map. First posted on arXiv 12 Oct 2023.
Builds on
46 of 64 references.
References · 46
- 2017Attention Is All You Needcited by 59,932
- 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
- 2020Language Models are Few-Shot Learnerscited by 32,199
- 2022Training language models to follow instructions with human feedbackcited by 15,712
- 2022Self-Consistency Improves Chain of Thought Reasoning in Language Modelscited by 5,601
- 2016A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networkscited by 2,362
- 2013Describing Textures in the Wildcited by 2,110
- 2022Survey of Hallucination in Natural Language Generationcited by 2,056
- 2019Language Models as Knowledge Bases?cited by 1,922
- 2022Language Models (Mostly) Know What They Knowcited by 1,783
- 2016Concrete Problems in AI Safetycited by 1,582
- 2018A Simple Unified Framework for Detecting Out-of-Distribution Samples and Adversarial Attackscited by 1,472
34 more
- 2022Competition-Level Code Generation with AlphaCodecited by 1,427
- 2020Energy-based Out-of-distribution Detectioncited by 1,296
- 2023Teaching Large Language Models to Self-Debugcited by 963
- 2021Generalized Out-of-Distribution Detection: A Surveycited by 796
- 2023Semantic Uncertainty: Linguistic Invariances for Uncertainty Estimation in Natural Language Generationcited by 795
- 2022Out-of-Distribution Detection with Deep Nearest Neighborscited by 596
- 2022InternImage: Exploring Large-Scale Vision Foundation Models with Deformable Convolutionscited by 479
- 2020Generalized ODIN: Detecting Out-of-distribution Image without Learning from Out-of-distribution Datacited by 408
- 2020LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understandingcited by 406
- 2023Check Your Facts and Try Again: Improving Large Language Models with External Knowledge and Automated Feedbackcited by 351
- 2022Visual Classification via Description from Large Language Modelscited by 331
- 2021An Empirical Study of GPT-3 for Few-Shot Knowledge-Based VQAcited by 298
- 2021Exploring the Limits of Out-of-Distribution Detectioncited by 283
- 2022DETRs with Collaborative Hybrid Assignments Trainingcited by 273
- 2022Prompting GPT-3 To Be Reliablecited by 258
- 2021MOS: Towards Scaling Out-of-distribution Detection for Large Semantic Spacecited by 233
- 2014Towards Open World Recognitioncited by 226
- 2022Delving into Out-of-Distribution Detection with Vision-Language Representationscited by 215
- 2022Factuality Enhanced Language Models for Open-Ended Text Generationcited by 200
- 2022Out-of-Distribution Detection and Selective Generation for Conditional Language Modelscited by 193
- 2020Contrastive Training for Improved Out-of-Distribution Detectioncited by 186
- 2021Zero-Shot Out-of-Distribution Detection Based on the Pre-trained Model CLIPcited by 151
- 2021WenLan: Bridging Vision and Language by Large-Scale Multi-Modal Pre-Trainingcited by 84
- 2021Contrastive Out-of-Distribution Detection for Pretrained Transformerscited by 70
- 2021On the Impact of Spurious Correlation for Out-of-distribution Detectioncited by 68
- 2023YOLOv6 v3.0: A Full-Scale Reloadingcited by 58
- 2023See, Think, Confirm: Interactive Prompting Between Vision and Language Models for Knowledge-based Visual Reasoningcited by 49
- 2021Out-of-Scope Intent Detection with Self-Supervision and Discriminative Trainingcited by 42
- 2022BigDetection: A Large-scale Benchmark for Improved Object Detector Pre-trainingcited by 33
- 2023A Survey on Out-of-Distribution Detection in NLPcited by 31
- 2021Enhancing the Generalization for Intent Classification and Out-of-Domain Detection in SLUcited by 22
- 2022Estimating Soft Labels for Out-of-Domain Intent Detectioncited by 11
- 2023Long-Tailed Question Answering in an Open Worldcited by 7
- 2023Domain Incremental Lifelong Learning in an Open Worldcited by 5
Built on by
Citing papers per year: 2023: 1; 2024: 3; 2025: 6; 2026: 2.
12 papers.
- 2023Open-Set Image Tagging with Multi-Grained Text Supervisioncited by 67
- 2024Out-of-Distribution Detection: A Task-Oriented Survey of Recent Advancescited by 24
- 2025ANTS: Adaptive Negative Textual Space Shaping for OOD Detection via Test-Time MLLM Understanding and Reasoningcited by 10
- 2024SeTAR: Out-of-Distribution Detection with Selective Low-Rank Approximationcited by 9
- 2025SimLabel: Consistency-Guided OOD Detection with Pretrained Vision-Language Modelscited by 5
- 2025Comparing Uncertainty Measurement and Mitigation Methods for Large Language Models: A Systematic Reviewcited by 3
- 2025Spurious-Aware Prototype Refinement for Reliable Out-of-Distribution Detectioncited by 2
- 2025Analysis of Image-and-Text Uncertainty Propagation in Multimodal Large Language Models with Cardiac MR-Based Applicationscited by 1
- 2024'No' Matters: Out-of-Distribution Detection in Multimodality Long Dialoguecited by 0
- 2025STEMTOX: From Collaborative Tags to Fine-Grained Toxic Meme Detection via Entropy-Guided Multi-Task Learningcited by 0
- 2026Multi-Agent Cooperative Learning for Robust Vision-Language Alignment under OOD Conceptscited by 0
- 2026Bridging the Semantic Chasm: Synergistic Conceptual Anchoring for Generalized Few-Shot and Zero-Shot OOD Perceptioncited by 0
- 2026Bridging the Semantic Chasm: Synergistic Conceptual Anchoring for Generalized Few-Shot and Zero-Shot OOD Perception30 Jan 2026
- 2026Multi-Agent Cooperative Learning for Robust Vision-Language Alignment under OOD Concepts11 Jan 2026
- 2025ANTS: Adaptive Negative Textual Space Shaping for OOD Detection via Test-Time MLLM Understanding and Reasoning4 Sep 2025
- 2025STEMTOX: From Collaborative Tags to Fine-Grained Toxic Meme Detection via Entropy-Guided Multi-Task Learning6 Aug 2025
- 2025Analysis of Image-and-Text Uncertainty Propagation in Multimodal Large Language Models with Cardiac MR-Based Applications17 Jul 2025
- 2025Spurious-Aware Prototype Refinement for Reliable Out-of-Distribution Detection30 Jun 2025
- 2025Comparing Uncertainty Measurement and Mitigation Methods for Large Language Models: A Systematic Review25 Apr 2025
- 2025SimLabel: Consistency-Guided OOD Detection with Pretrained Vision-Language Models20 Jan 2025
- 2024'No' Matters: Out-of-Distribution Detection in Multimodality Long Dialogue31 Oct 2024
- 2024Out-of-Distribution Detection: A Task-Oriented Survey of Recent Advances18 Sep 2024
- 2024SeTAR: Out-of-Distribution Detection with Selective Low-Rank Approximation18 Jun 2024
- 2023Open-Set Image Tagging with Multi-Grained Text Supervision23 Oct 2023