Handling Out-of-Distribution Data: A Survey

In the field of Machine Learning (ML) and data-driven applications, one of the significant challenge is the change in data distribution between the training and deployment stages, commonly known as distribution shift. This paper outlines different mechanisms for handling two main types of distribution shifts: (i) Covariate shift: where the value of features or covariates change between train and test data, and (ii) Concept/Semantic-shift: where model experiences shift in the concept learned during training due to emergence of novel classes in the test phase. We sum up our contributions in three folds. First, we formalize distribution shifts, recite on how the conventional method fails to handle them adequately and urge for a model that can simultaneously perform better in all types of distribution shifts. Second, we discuss why handling distribution shifts is important and provide an extensive review of the methods and techniques that have been developed to detect, measure, and mitigate the effects of these shifts. Third, we discuss the current state of distribution shift handling mechanisms and propose future research directions in this area. Overall, we provide a retrospective synopsis of the literature in the distribution shift, focusing on OOD data that had been overlooked in the existing surveys.

• Publication year

  2025

• Venue

  IEEE Transactions on Knowledge and Data Engineering

• Publication date

  2025-07-25

• Fields of study

  Computer Science

• Identifiers
  DOI 10.1109/TKDE.2025.3592614arXiv 2507.21160
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Concept Drift in Large Language Models

  2025cited by this paper
• Selective Random Walk for Transfer Learning in Heterogeneous Label Spaces

  2024cited by this paper
• Enhancing Domain Adaptation through Prompt Gradient Alignment

  2024cited by this paper
• MCD: Defense Against Query-Based Black-Box Surrogate Attacks

  2024cited by this paper
• Out-of-Distribution Generalization With Causal Feature Separation

  2024cited by this paper
• Rethinking Multi-Domain Generalization with A General Learning Objective

  2024cited by this paper
• Margin-Bounded Confidence Scores for Out-of-Distribution Detection

  2024cited by this paper
• Hierarchical Attention Network for Open-Set Fine-Grained Image Recognition

  2024cited by this paper
• Generalize or Detect? Towards Robust Semantic Segmentation Under Multiple Distribution Shifts

  2024cited by this paper
• ICL: Iterative Continual Learning for Multi-domain Neural Machine Translation

  2024cited by this paper
• Soft Prompt Generation for Domain Generalization

  2024cited by this paper
• A data-driven approach for intrusion and anomaly detection using automated machine learning for the Internet of Things

  2023cited by this paper
• SIMPLE: Specialized Model-Sample Matching for Domain Generalization

  2023cited by this paper
• Astroformer: More Data Might not be all you need for Classification

  2023cited by this paper
• Decoupling MaxLogit for Out-of-Distribution Detection

  2023cited by this paper
• DeepLens: Interactive Out-of-distribution Data Detection in NLP Models

  2023cited by this paper
• Derandomized Novelty Detection with FDR Control via Conformal E-values

  2023cited by this paper
• Continual Semantic Segmentation with Automatic Memory Sample Selection

  2023cited by this paper
• Mind the Label Shift of Augmentation-based Graph OOD Generalization

  2023cited by this paper
• Pushing the Limits of Fewshot Anomaly Detection in Industry Vision: Graphcore

  2023cited by this paper
• Subspace Identification for Multi-Source Domain Adaptation

  2023cited by this paper
• Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection

  2023cited by this paper
• Bilateral Memory Consolidation for Continual Learning

  2023cited by this paper
• Feed Two Birds with One Scone: Exploiting Wild Data for Both Out-of-Distribution Generalization and Detection

  2023cited by this paper
• A Generic Learning Framework for Sequential Recommendation with Distribution Shifts

  2023cited by this paper
• Progressive Open Space Expansion for Open-Set Model Attribution

  2023cited by this paper
• OpenMix: Exploring Outlier Samples for Misclassification Detection

  2023cited by this paper
• PromptStyler: Prompt-driven Style Generation for Source-free Domain Generalization

  2023cited by this paper
• RLSbench: Domain Adaptation Under Relaxed Label Shift

  2023cited by this paper
• FREDOM: Fairness Domain Adaptation Approach to Semantic Scene Understanding

  2023cited by this paper
• Combining inherent knowledge of vision-language models with unsupervised domain adaptation through self-knowledge distillation

  2023cited by this paper
• Out-of-Distribution Generalization in Natural Language Processing: Past, Present, and Future

  2023cited by this paper
• GLIGEN: Open-Set Grounded Text-to-Image Generation

  2023cited by this paper
• Learning multiple gaussian prototypes for open-set recognition

  2023cited by this paper
• Rethinking Out-of-distribution (OOD) Detection: Masked Image Modeling is All You Need

  2023cited by this paper
• Progressive Prompts: Continual Learning for Language Models

  2023cited by this paper
• Meta OOD Learning For Continuously Adaptive OOD Detection

  2023cited by this paper
• Spatial-Temporal Federated Transfer Learning with multi-sensor data fusion for cooperative positioning

  2023cited by this paper
• Energy-based Out-of-Distribution Detection for Graph Neural Networks

  2023cited by this paper
• Comprehensive Assessment of the Performance of Deep Learning Classifiers Reveals a Surprising Lack of Robustness

  2023cited by this paper
• Deep discriminative transfer learning network for cross-machine fault diagnosis

  2023cited by this paper
• Learning Discriminative Feature Representation for Open Set Action Recognition

  2023cited by this paper
• Dual Alignment Unsupervised Domain Adaptation for Video-Text Retrieval

  2023cited by this paper
• On the Robustness of ChatGPT: An Adversarial and Out-of-distribution Perspective

  2023cited by this paper
• Large margin distribution multi-class supervised novelty detection

  2023cited by this paper
• Sharpness-Aware Gradient Matching for Domain Generalization

  2023cited by this paper
• MAP: Towards Balanced Generalization of IID and OOD through Model-Agnostic Adapters

  2023cited by this paper
• Diversity-Measurable Anomaly Detection

  2023cited by this paper
• Peer-to-Peer Federated Continual Learning for Naturalistic Driving Action Recognition

  2023cited by this paper
• Revisiting Out-of-distribution Robustness in NLP: Benchmark, Analysis, and LLMs Evaluations

  2023cited by this paper
• ADPL: Adaptive Dual Path Learning for Domain Adaptation of Semantic Segmentation

  2023cited by this paper
• State of the art: a review of sentiment analysis based on sequential transfer learning

  2022cited by this paper
• NICO++: Towards Better Benchmarking for Domain Generalization

  2022cited by this paper
• What Language Model Architecture and Pretraining Objective Work Best for Zero-Shot Generalization?

  2022cited by this paper
• A Closer Look at Rehearsal-Free Continual Learning *

  2022cited by this paper
• Dependable Intrusion Detection System for IoT: A Deep Transfer Learning Based Approach

  2022cited by this paper
• Mitigating Neural Network Overconfidence with Logit Normalization

  2022cited by this paper
• Full-Spectrum Out-of-Distribution Detection

  2022cited by this paper
• Bamboo: Building Mega-Scale Vision Dataset Continually with Human–Machine Synergy

  2022cited by this paper
• An Evolutionary Approach to Dynamic Introduction of Tasks in Large-scale Multitask Learning Systems

  2022cited by this paper
• Out-Of-Distribution Detection Is Not All You Need

  2022cited by this paper
• Transfer learning for medical image classification: a literature review

  2022cited by this paper
• Causality Inspired Representation Learning for Domain Generalization

  2022cited by this paper
• Denoising diffusion models for out-of-distribution detection

  2022cited by this paper
• Runtime Monitoring of Deep Neural Networks Using Top-Down Context Models Inspired by Predictive Processing and Dual Process Theory

  2022cited by this paper
• Learning Causally Invariant Representations for Out-of-Distribution Generalization on Graphs

  2022cited by this paper
• Is Out-of-Distribution Detection Learnable?

  2022cited by this paper
• Open-set learning under covariate shift

  2022cited by this paper
• MGFN: Magnitude-Contrastive Glance-and-Focus Network for Weakly-Supervised Video Anomaly Detection

  2022cited by this paper
• Federated Domain Generalization for Image Recognition via Cross-Client Style Transfer

  2022cited by this paper
• Online Continual Learning with Contrastive Vision Transformer

  2022cited by this paper
• Class-Specific Semantic Reconstruction for Open Set Recognition

  2022cited by this paper
• Runtime Monitoring for Out-of-Distribution Detection in Object Detection Neural Networks

  2022cited by this paper
• Transfer learning for raw network traffic detection

  2022cited by this paper
• Class-aware sample reweighting optimal transport for multi-source domain adaptation

  2022cited by this paper
• Training OOD Detectors in their Natural Habitats

  2022cited by this paper
• Domain Adaptation under Open Set Label Shift

  2022cited by this paper
• Unsupervised Continual Learning for Gradually Varying Domains

  2022cited by this paper
• Continual learning with Bayesian model based on a fixed pre-trained feature extractor

  2022cited by this paper
• Sparse Coding in a Dual Memory System for Lifelong Learning

  2022cited by this paper
• A perspective survey on deep transfer learning for fault diagnosis in industrial scenarios: Theories, applications and challenges

  2022cited by this paper
• DATID-3D: Diversity-Preserved Domain Adaptation Using Text-to-Image Diffusion for 3D Generative Model

  2022cited by this paper
• MIC: Masked Image Consistency for Context-Enhanced Domain Adaptation

  2022cited by this paper
• POEM: Out-of-Distribution Detection with Posterior Sampling

  2022cited by this paper
• Visual Prompt Tuning for Generative Transfer Learning

  2022cited by this paper
• Domain Adaptation via Prompt Learning

  2022cited by this paper
• PIVOT: Prompting for Video Continual Learning

  2022cited by this paper
• Open-world Machine Learning: Applications, Challenges, and Opportunities

  2021cited by this paper
• Deep Learning: A Comprehensive Overview on Techniques, Taxonomy, Applications and Research Directions

  2021cited by this paper
• ReAct: Out-of-distribution Detection With Rectified Activations

  2021cited by this paper
• Provable Guarantees for Understanding Out-of-distribution Detection

  2021cited by this paper
• Towards Open-Set Text Recognition via Label-to-Prototype Learning

  2021cited by this paper
• DualNet: Continual Learning, Fast and Slow Supplementary Materials

  2021cited by this paper
• SWAD: Domain Generalization by Seeking Flat Minima

  2021cited by this paper
• Learning to Prompt for Vision-Language Models

  2021cited by this paper
• Exploring Covariate and Concept Shift for Detection and Calibration of Out-of-Distribution Data

  2021cited by this paper
• Statistical Learning Theory

  2021cited by this paper
• Deep transfer learning for conditional shift in regression

  2021cited by this paper
• Adversarial Reciprocal Points Learning for Open Set Recognition

  2021cited by this paper
• Spatial Location Constraint Prototype Loss for Open Set Recognition

  2021cited by this paper

• KL divergence-guided transfer learning for data-driven shield tunneling under distribution shift

  2026cites this paper
• Short bond evaluation method for rapidly assessing the generalization ability of deep neural network potential function models and its effectiveness verification

  2026cites this paper
• A Note on k-NN Gating in RAG

  2026cites this paper
• Cross-Fusion Distance: A Novel Metric for Measuring Fusion and Separability Between Data Groups in Representation Space

  2026cites this paper
• ERIS: An Energy-Guided Feature Disentanglement Framework for Out-of-Distribution Time Series Classification

  2025cites this paper
• Prior Distribution and Model Confidence

  2025cites this paper
• Fundamental limitations of online supervised learning in dynamic control loops

  2025cites this paper
• General OOD Detection via Model-aware and Subspace-aware Variable Priority

  2025cites this paper