A curated list of Distribution Shift papers/articles and recent advancements.
Data distribution shift refers to the phenomenon in supervised learning when the data a model works with changes over time, which causes this model’s predictions to become less accurate as time passes. The distribution of the data the model is trained on is called the source distribution. This repo contains a curated list of Distribution Shift papers/articles and recent advancements in Machine learning.
- Enhancing Model Robustness and Fairness with Causality: A Regularization Approach
- Does Robustness Improve Fairness? Approaching Fairness with Word Substitution Robustness Methods for Text Classification
- [2001.08103] Secure and Robust Machine Learning for Healthcare: A Survey
- [2103.08291] Robust Machine Learning in Critical Care -- Software Engineering and Medical Perspectives
- [2108.00402] Style Curriculum Learning for Robust Medical Image Segmentation
- [2108.12242] Deep learning models are not robust against noise in clinical text
- [2210.00589] Uncertainty estimations methods for a deep learning model to aid in clinical decision-making -- a clinician's perspective
- [2209.15042] Generalizability of Adversarial Robustness Under Distribution Shifts
- [2209.09423] Fairness and robustness in anti-causal prediction
- [2209.09631] De-Identification of French Unstructured Clinical Notes for Machine Learning Tasks
- [2207.00769] Test-time Adaptation with Calibration of Medical Image Classification Nets for Label Distribution Shift
- Performance deterioration of deep neural networks for lesion classification in mammography due to distribution shift: an analysis based on artificially created distribution shift
- [2109.01668] How Reliable Are Out-of-Distribution Generalization Methods for Medical Image Segmentation?
- [1910.13681] The Domain Shift Problem of Medical Image Segmentation and Vendor-Adaptation by Unet-GAN
- Secure and Robust Machine Learning for Healthcare: A Survey
- AIMI Research Meeting: Rethink Robustness of Deep Learning Models for Medical Image Analysis - Yuyin Zhou, PhD
- Identification of robust deep neural network models of longitudinal clinical measurements
- Robustness of AI-based prognostic and systems health management - ScienceDirect
- The impact of domain shift on the calibration of fine-tuned models
- [2110.01955] Distribution Mismatch Correction for Improved Robustness in Deep Neural Networks
- agrawal20a.pdf
- nestor19a.pdf
- liu21f.pdf
- cheng20a.pdf
- EHR Foundation Models Improve Robustness in the Presence of Temporal Distribution Shift
- Maintaining fairness across distribution shift: do we have viable solutions for real-world applications?
- Characterizing the Value of Information in Medical Notes
- subbaswamy19a.pdf
- Forecasting Patient Outcomes in Kidney Exchange
- https://openreview.net/pdf?id=AVTfiZgV64X
- [1910.00199] Saliency is a Possible Red Herring When Diagnosing Poor Generalization
- [2007.00644] Measuring Robustness to Natural Distribution Shifts in Image Classification
- [2206.14467] Single-domain Generalization in Medical Image Segmentation via Test-time Adaptation from Shape Dictionary
- [2205.13723] DLTTA: Dynamic Learning Rate for Test-time Adaptation on Cross-domain Medical Images
- [2203.06060] ROOD-MRI: Benchmarking the robustness of deep learning segmentation models to out-of-distribution and corrupted data in MRI
- [2112.13734] Multi-Domain Balanced Sampling Improves Out-of-Distribution Generalization of Chest X-ray Pathology Prediction Models
- [2110.14019] Reliable and Trustworthy Machine Learning for Health Using Dataset Shift Detection
- [2110.09276] Natural Attribute-based Shift Detection
- [2109.13230] The Impact of Domain Shift on Left and Right Ventricle Segmentation in Short Axis Cardiac MR Images
- Adapting Event Extractors to Medical Data: Bridging the Covariate Shift
- Distinguishing Clinical Sentiment: The Importance of Domain Adaptation in Psychiatric Patient Health Records
- 2022.clinicalnlp-1.10.pdf
- Investigating the Challenges of Temporal Relation Extraction from Clinical Text
- 2022.findings-acl.192.pdf
- 2022.findings-acl.18.pdf
- otles21a.pdf
- pfisterer22a.pdf
- caldas21a.pdf
- zhang13d.pdf
- Review for NeurIPS paper: What went wrong and when? Instance-wise feature importance for time-series black-box models
- 08fa43588c2571ade19bc0fa5936e028-Paper.pdf
- 908075ea2c025c335f4865f7db427062-Paper.pdf
- Domain Generalization via Model-Agnostic Learning of Semantic Features
- [2208.03392] Federated Learning for Medical Applications: A Taxonomy, Current Trends, Challenges, and Future Research Directions
- Adapting on the Fly to Test Time Distribution Shift – The Berkeley Artificial Intelligence Research Blog
- Estimating Generalization under Distribution Shifts via Domain-Invariant Representations
- [2207.11486] Time Series Prediction under Distribution Shift using Differentiable Forgetting
- https://www.ml.cmu.edu/research/phd-dissertation-pdfs/yw4_phd_ml_2021.pdf
- statistics - Distribution Shift vs Transfer Learning - Data Science Stack Exchange
- Failing Loudly: An Empirical Study of Methods for Detecting Dataset Shift
- [PDF] A Fine-Grained Analysis on Distribution Shift
- The Many Faces of Robustness: A Critical Analysis of Out-of-Distribution Generalization
- 4.7. Environment and Distribution Shift — Dive into Deep Learning 1.0.0-alpha1.post0 documentation
- [2207.00476] Online Reflective Learning for Robust Medical Image Segmentation
- [2207.01059] Identifying the Context Shift between Test Benchmarks and Production Data
- [2206.05498] A Review of Causality for Learning Algorithms in Medical Image Analysis
- [2206.08023] AMOS: A Large-Scale Abdominal Multi-Organ Benchmark for Versatile Medical Image Segmentation
- darestani21a.pdf
- Conditional Synthetic Data Generation for Robust Machine Learning Applications with Limited Pandemic Data
- Multi-Domain Ensembles for Domain Generalization
- Optimal Representations for Covariate Shifts
- Fourier-Based Augmentations for Improved Robustness and Uncertainty Calibration
- Investigating Shifts in GAN Output-Distributions
- Exploring Covariate and Concept Shift for Out-of-Distribution Detection
- Unsupervised Attribute Alignment for Characterizing Distribution Shift
- BEDS-Bench: Behavior of EHR-models under Distributional Shift - A Benchmark
- How Does Contrastive Pre-training Connect Disparate Domains?
- Ensembles and Cocktails: Robust Finetuning for Natural Language Generation
- Distribution Shift in Airline Customer Behavior during COVID-19
- PixMix: Dreamlike Pictures Comprehensively Improve Safety Measures
- Avoiding Spurious Correlations: Bridging Theory and Practice
- MEMO: Test Time Robustness via Adaptation and Augmentation
- Understanding Post-hoc Adaptation for Improving Subgroup Robustness
- Diurnal or Nocturnal? Federated Learning from Periodically Shifting Distributions
- An Empirical Investigation of Model-to-Model Distribution Shifts in Trained Convolutional Filters
- Maximum Mean Discrepancy for Generalization in the Presence of Distribution and Missingness Shift
- An Empirical Study of Pre-trained Vision Models on Out-of-distribution Generalization
- Reliable Graph Neural Networks for Drug Discovery Under Distributional Shift
- A benchmark with decomposed distribution shifts for 360 monocular depth estimation
- Leveraging Unlabeled Data to Predict Out-of-Distribution Performance
- Quantifying and Alleviating Distribution Shifts in Foundation Models on Review Classification
- A fine-grained analysis of robustness to distribution shifts
- A Closer Look at Distribution Shifts and Out-of-Distribution Generalization on Graphs
- Is Importance Weighting Incompatible with Interpolating Classifiers?
- Benchmarking Bayesian Deep Learning on Diabetic Retinopathy Detection Tasks
- A Benchmark for Text Quantification Learning Under Real-World Temporal Distribution Shift
- MetaShift: A Dataset of Datasets for Evaluating Contextual Distribution Shifts and Training Conflicts
- [2112.13885] MedShift: identifying shift data for medical dataset curation
- [2207.00769] Test-time Adaptation with Calibration of Medical Image Classification Nets for Label Distribution Shift
- [2207.05796] Estimating Test Performance for AI Medical Devices under Distribution Shift with Conformal Prediction
- [2206.15274] Exposing and addressing the fragility of neural networks in digital pathology
- [2205.09723] Robust and Efficient Medical Imaging with Self-Supervision
- [2208.03217] Distance-based detection of out-of-distribution silent failures for Covid-19 lung lesion segmentation
- [2203.05574] On-the-Fly Test-time Adaptation for Medical Image Segmentation
- [2202.02833] CheXstray: Real-time Multi-Modal Data Concordance for Drift Detection in Medical Imaging AI
- [2202.05271] A Field of Experts Prior for Adapting Neural Networks at Test Time
- [2201.07317] A Privacy-Preserving Unsupervised Domain Adaptation Framework for Clinical Text Analysis
- [2110.06866] Bayesian logistic regression for online recalibration and revision of risk prediction models with performance guarantees
- [2107.14317] Temporal Dependencies in Feature Importance for Time Series Predictions
- Analysis of Machine Learning Models Predicting Quality of Life for Cancer Patients
- CSUR5405-111
- Data Distribution Shifts and Monitoring
- Shifting the distribution
- Principles for Tackling Distribution Shift: Pessimism, Adaptation, and Anticipation - YouTube
- Zachary C. Lipton: Deep Learning Under Distribution Shift - YouTube
- Preventing dataset shift from breaking machine-learning biomarkers
- How robust are pre-trained models to distribution shift?
- [1911.00677] Fairness Violations and Mitigation under Covariate Shift
- f9a2ae9ee8021aeb70a8f2deeab247a324b8200e.pdf
- attachment
- towards-explaining-image-based-shifts.pdf
- [2110.11328] A Fine-Grained Analysis on Distribution Shift
- [2205.12753] An Empirical Study on Distribution Shift Robustness From the Perspective of Pre-Training and Data Augmentation
- [2202.01034] Maintaining fairness across distribution shift: do we have viable solutions for real-world applications?
- [2207.05796] Estimating Test Performance for AI Medical Devices under Distribution Shift with Conformal Prediction
- EHR Foundation Models Improve Robustness in the Presence of Temporal Distribution Shift
- SOoD: Self-Supervised Out-of-Distribution Detection Under Domain Shift for Multi-Class Colorectal Cancer Tissue Types
- subbaswamy21a.pdf
- machine learning - Difference between distribution shift and data shift, concept drift and model drift - Cross Validated
- Understanding Dataset Shift. How to make sure your models are not…
- Distribution Shift Framework
- 4.7. Environment and Distribution Shift — Dive into Deep Learning 1.0.0-alpha1.post0 documentation
- https://www.ml.cmu.edu/research/phd-dissertation-pdfs/yw4_phd_ml_2021.pdf
- Mechanical MNIST – Distribution Shift
- microsoft/distribution-shift-latent-representations
- NeurIPS DistShift Workshop 2021
- Types of Out-of-Distribution Texts and How to Detect Them
- Looking Beyond Label Noise: Shifted Label Distribution Matters in Distantly Supervised Relation Extraction
- To Annotate or Not? Predicting Performance Drop under Domain Shift
- 2022.acl-long.223.pdf
- 2022.repl4nlp-1.1.pdf
- 2022.findings-acl.68.pdf
- 2022.acl-long.74.pdf
- 2022.naacl-main.339.pdf
- [2103.17171] Spectral decoupling allows training transferable neural networks in medical imaging
- [2102.08660] CheXternal: Generalization of Deep Learning Models for Chest X-ray Interpretation to Photos of Chest X-rays and External Clinical Settings
- [2012.10564] Computer-aided abnormality detection in chest radiographs in a clinical setting via domain-adaptation
- [2011.11750] Federated Semi-Supervised Learning for COVID Region Segmentation in Chest CT using Multi-National Data from China, Italy, Japan
- [2010.06667] Chasing Your Long Tails: Differentially Private Prediction in Health Care Settings
- [2007.02035] Shape-aware Meta-learning for Generalizing Prostate MRI Segmentation to Unseen Domains
- [2006.00327] Probabilistic self-learning framework for Low-dose CT Denoising
- Metric Learning in Optimal Transport for Domain Adaptation
- [2202.10808] Combating Distribution Shift for Accurate Time Series Forecasting via Hypernetworks
- [2207.11486] Time Series Prediction under Distribution Shift using Differentiable Forgetting
- [2204.10049] On Distribution Shift in Learning-based Bug Detectors
- [2206.00129] Fairness Transferability Subject to Bounded Distribution Shift
- [2208.06604] Combating Label Distribution Shift for Active Domain Adaptation
- [2209.11459] TeST: Test-time Self-Training under Distribution Shift
- [2202.06523] MetaShift: A Dataset of Datasets for Evaluating Contextual Distribution Shifts and Training Conflicts
- [2207.05796] Estimating Test Performance for AI Medical Devices under Distribution Shift with Conformal Prediction
- [2206.13089] Agreement-on-the-Line: Predicting the Performance of Neural Networks under Distribution Shift
- [2210.00084] Contrastive Graph Few-Shot Learning
- [2210.01360] Learning an Invertible Output Mapping Can Mitigate Simplicity Bias in Neural Networks
- [2210.01979] GAPX: Generalized Autoregressive Paraphrase-Identification X
- [2210.03103] Env-Aware Anomaly Detection: Ignore Style Changes, Stay True to Content!
- [2210.01930] Benchmarking Learnt Radio Localisation under Distribution Shift
- [2209.01332] Class-Specific Channel Attention for Few-Shot Learning
- [2209.01321] Deep Stable Representation Learning on Electronic Health Records
- [2209.15177] Domain Generalization -- A Causal Perspective
- [2209.03620] Black-Box Audits for Group Distribution Shifts
- [2209.05706] Non-Parametric Temporal Adaptation for Social Media Topic Classification
- [2209.05779] Test-Time Adaptation with Principal Component Analysis
- Unlearn Dataset Bias in Natural Language Inference by Fitting the Residual
- 2022.findings-naacl.13.pdf
- 2022.acl-long.256.pdf
- 2022.naacl-srw.6.pdf
- Test-Time Training Can Close the Natural Distribution Shift Performance Gap in Deep Learning Based Compressed Sensing
- zhou21g.pdf
- he22a.pdf
- Estimating Generalization under Distribution Shifts via Domain-Invariant Representations
- LTF: A Label Transformation Framework for Correcting Target Shift
- 07ac7cd13fd0eb1654ccdbd222b81437-Paper.pdf
- 8b9e7ab295e87570551db122a04c6f7c-Paper.pdf
- d1f255a373a3cef72e03aa9d980c7eca-Paper.pdf
- 219e052492f4008818b8adb6366c7ed6-Paper.pdf
- Provably Efficient Q-learning with Function Approximation via Distribution Shift Error Checking Oracle
- f5e536083a438cec5b64a4954abc17f1-Paper.pdf
- Domain Adaptation by Using Causal Inference to Predict Invariant Conditional Distributions
- dfbfa7ddcfffeb581f50edcf9a0204bb-Paper.pdf
- https://www.ijcai.org/Proceedings/15/Papers/147.pdf
- https://www.ijcai.org/proceedings/2022/0232.pdf
- https://www.ijcai.org/proceedings/2022/0484.pdf
- https://www.ijcai.org/proceedings/2022/0595.pdf
- https://www.ijcai.org/proceedings/2021/0644.pdf
- https://www.ijcai.org/proceedings/2022/0392.pdf
- https://www.ijcai.org/proceedings/2022/0501.pdf
- https://www.ijcai.org/proceedings/2022/0514.pdf
- https://www.ijcai.org/proceedings/2021/0367.pdf
- https://www.ijcai.org/proceedings/2022/0240.pdf
- https://openreview.net/pdf?id=Dl4LetuLdyK
- https://openreview.net/pdf?id=BUQD1tJ2UwK
- e2d52448d36918c575fa79d88647ba66-Paper.pdf
- fang22a.pdf
- kang22a.pdf
- zhao21b.pdf
- yu22i.pdf
- Can Autonomous Vehicles Identify, Recover From,and Adapt to Distribution Shifts?
- d8330f857a17c53d217014ee776bfd50-Paper.pdf
- 0d441de75945e5acbc865406fc9a2559-Paper.pdf
- 73fed7fd472e502d8908794430511f4d-Paper.pdf
- 8420d359404024567b5aefda1231af24-Paper.pdf
- 621461af90cadfdaf0e8d4cc25129f91-Paper.pdf
- [2209.08253] Mitigating Both Covariate and Conditional Shift for Domain Generalization
- [2209.00652] Towards Optimization and Model Selection for Domain Generalization: A Mixup-guided Solution
- [2209.01501] Meta-Learning with Less Forgetting on Large-Scale Non-Stationary Task Distributions
- [2209.02408] Robustness and invariance properties of image classifiers
- [2209.08745] Importance Tempering: Group Robustness for Overparameterized Models
- RECORD: Resource Constrained Semi-Supervised Learning under Distribution Shift
- Towards Reliable Multimodal Stress Detection under Distribution Shift
- Confidence May Cheat: Self-Training on Graph Neural Networks under Distribution Shift
- Active Model Adaptation Under Unknown Shift
- Balance-Subsampled Stable Prediction Across Unknown Test Data
- Focused Context Balancing for Robust Offline Policy Evaluation
- 3511598
- A Critical Reassessment of the Saerens-Latinne-Decaestecker Algorithm for Posterior Probability Adjustment
- Understanding the Effect of Out-of-distribution Examples and Interactive Explanations on Human-AI Decision Making
- Co-training Disentangled Domain Adaptation Network for Leveraging Popularity Bias in Recommenders
- CausPref: Causal Preference Learning for Out-of-Distribution Recommendation
- Off-Policy Actor-critic for Recommender Systems
- HybridRepair: towards annotation-efficient repair for deep learning models
- Decoupled Reinforcement Learning to Stabilise Intrinsically-Motivated Exploration
- Neural Statistics for Click-Through Rate Prediction
- Influence Function for Unbiased Recommendation
- DCAF-BERT: A Distilled Cachable Adaptable Factorized Model For Improved Ads CTR Prediction
- Making Adversarially-Trained Language Models Forget with Model Retraining: A Case Study on Hate Speech Detection
- 3455716.3455805
- AdaRNN
- Transfer String Kernel for Cross-Context DNA-Protein Binding Prediction
- Real Negatives Matter: Continuous Training with Real Negatives for Delayed Feedback Modeling
- A New Generation of Perspective API: Efficient Multilingual Character-level Transformers
- Robust Self-Supervised Structural Graph Neural Network for Social Network Prediction
- Stable Prediction across Unknown Environments
- 3460120.3484776
- Fairness Violations and Mitigation under Covariate Shift
- Quantifying the Performance of Adversarial Training on Language Models with Distribution Shifts
- FedRS
- [2208.02896] Interpretable Distribution Shift Detection using Optimal Transport
- [2206.05480] CodeS: A Distribution Shift Benchmark Dataset for Source Code Learning
- [2206.08871] How robust are pre-trained models to distribution shift?
- [2207.08977] Calibrated ensembles can mitigate accuracy tradeoffs under distribution shift
- [2202.02339] Discovering Distribution Shifts using Latent Space Representations
- [2207.04075] Models Out of Line: A Fourier Lens on Distribution Shift Robustness