mlsanityData to 13 Oct 2026

Additive MIL: Intrinsically Interpretable Multiple Instance Learning for Pathology

arXiv 2206.01794 ↗First posted 3 Jun 2022cs.CV cs.LGcited by 58

Not on the map. Why?

Abstract (arXiv)

Multiple Instance Learning (MIL) has been widely applied in pathology towards solving critical problems such as automating cancer diagnosis and grading, predicting patient prognosis, and therapy response. Deploying these models in a clinical setting requires careful inspection of these black boxes during development and deployment to identify failures and maintain physician trust. In this work, we propose a simple formulation of MIL models, which enables interpretability while maintaining similar predictive performance. Our Additive MIL models enable spatial credit assignment such that the contribution of each region in the image can be exactly computed and visualized. We show that our spatial credit assignment coincides with regions used by pathologists during diagnosis and improves upon classical attention heatmaps from attention MIL models. We show that any existing MIL model can be made additive with a simple change in function composition. We also show how these models can debug model failures, identify spurious features, and highlight class-wise regions of interest, enabling their use in high-stakes environments such as clinical decision-making.

Key facts. Cited by 58 papers in the map (data through 13 Oct 2026). Builds on 25 of its 56 references that are in the map. First posted on arXiv 3 Jun 2022.

Builds on

25 of 56 references.

References · 25

  1. 2014Adam: A Method for Stochastic Optimizationcited by 37,093
  2. 2017A Unified Approach to Interpreting Model Predictionscited by 6,515
  3. 2018Sanity Checks for Saliency Mapscited by 1,152
  4. 2018ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Designcited by 1,134
  5. 2018Attention-based Deep Multiple Instance Learningcited by 1,083
  6. 2020Data Efficient and Weakly Supervised Computational Pathology on Whole Slide Imagescited by 739
  7. 2021TransMIL: Transformer based Correlated Multiple Instance Learning for Whole Slide Image Classificationcited by 551
  8. 2020Dual-stream Multiple Instance Learning Network for Whole Slide Image Classification with Self-supervised Contrastive Learningcited by 474
  9. 2019The many Shapley values for model explanationcited by 343
  10. 2017The (Un)reliability of saliency methodscited by 332
  11. 2020Neural Additive Models: Interpretable Machine Learning with Neural Netscited by 318
  12. 2019Explaining individual predictions when features are dependent: More accurate approximations to Shapley valuescited by 207
13 more
  1. 2020Whole Slide Images based Cancer Survival Prediction using Attention Guided Deep Multiple Instance Learning Networkscited by 178
  2. 2016Deep Learning for Identifying Metastatic Breast Cancercited by 93
  3. 2020On the Tractability of SHAP Explanationscited by 92
  4. 2020General Pitfalls of Model-Agnostic Interpretation Methods for Machine Learning Modelscited by 74
  5. 2019Exploratory Not Explanatory: Counterfactual Analysis of Saliency Maps for Deep Reinforcement Learningcited by 60
  6. 2020Towards falsifiable interpretability researchcited by 53
  7. 2019Deep learning-based survival prediction for multiple cancer types using histopathology imagescited by 47
  8. 2020A Multi-resolution Model for Histopathology Image Classification and Localization with Multiple Instance Learningcited by 35
  9. 2020Data Valuation for Medical Imaging Using Shapley Value: Application on A Large-scale Chest X-ray Datasetcited by 34
  10. 2019Resolving challenges in deep learning-based analyses of histopathological images using explanation methodscited by 32
  11. 2020Artificial Intelligence Assistance Significantly Improves Gleason Grading of Prostate Biopsies by Pathologistscited by 9
  12. 2020Improving Interpretability for Computer-aided Diagnosis tools on Whole Slide Imaging with Multiple Instance Learning and Gradient-based Explanationscited by 8
  13. 2022Rethinking Machine Learning Model Evaluation in Pathologycited by 6

Built on by

Cited by year (2026 to 13 Oct)

Citing papers per year: 2023: 12; 2024: 19; 2025: 11; 2026: 16.

58 papers.

Sort:

By area:

  1. 2023Visual Language Pretrained Multiple Instance Zero-Shot Transfer for Histopathology Imagescited by 133
  2. 2023Artificial Intelligence for Digital and Computational Pathologycited by 107
  3. 2024Morphological Prototyping for Unsupervised Slide Representation Learning in Computational Pathologycited by 70
  4. 2024Transcriptomics-guided Slide Representation Learning in Computational Pathologycited by 51
  5. 2023The Rise of AI Language Pathologists: Exploring Two-level Prompt Learning for Few-shot Weakly-supervised Whole Slide Image Classificationcited by 41
  6. 2023SI-MIL: Taming Deep MIL for Self-Interpretability in Gigapixel Histopathologycited by 25
  7. 2023Inherently Interpretable Time Series Classification via Multiple Instance Learningcited by 23
  8. 2024PLUTO: Pathology-Universal Transformercited by 17
  9. 2024xMIL: Insightful Explanations for Multiple Instance Learning in Histopathologycited by 16
  10. 2023Label-Efficient Deep Learning in Medical Image Analysis: Challenges and Future Directionscited by 15
  11. 2024Learning biologically relevant features in a pathology foundation model using sparse autoencoderscited by 14
  12. 2023Shapley Values-enabled Progressive Pseudo Bag Augmentation for Whole Slide Image Classificationcited by 12
38 more
  1. 2024Rethinking Transformer for Long Contextual Histopathology Whole Slide Image Analysiscited by 12
  2. 2024Free Lunch in Pathology Foundation Model: Task-specific Model Adaptation with Concept-Guided Feature Enhancementcited by 11
  3. 2023Disambiguated Attention Embedding for Multi-Instance Partial-Label Learningcited by 10
  4. 2025Label-free Concept Based Multiple Instance Learning for Gigapixel Histopathologycited by 8
  5. 2024From Correlation to Causation: Max-Pooling-Based Multi-Instance Learning Leads to More Robust Whole Slide Image Classificationcited by 7
  6. 2024MergeUp-augmented Semi-Weakly Supervised Learning for WSI Classificationcited by 7
  7. 2024Ordinal Multiple-instance Learning for Ulcerative Colitis Severity Estimation with Selective Aggregated Transformercited by 7
  8. 2024A self-supervised framework for learning whole slide representationscited by 6
  9. 2024Learning from Partial Label Proportions for Whole Slide Image Segmentationcited by 6
  10. 2025Self-Supervision Enhances Instance-based Multiple Instance Learning Methods in Digital Pathology: A Benchmark Studycited by 6
  11. 2023SC-MIL: Supervised Contrastive Multiple Instance Learning for Imbalanced Classification in Pathologycited by 5
  12. 2024Detecting Domain Shift in Multiple Instance Learning for Digital Pathology Using Fr'echet Domain Distancecited by 3
  13. 2025PathVQ: Reforming Computational Pathology Foundation Model for Whole Slide Image Analysis via Vector Quantizationcited by 3
  14. 2023Mixed Models with Multiple Instance Learningcited by 2
  15. 2025Prototype-Based Multiple Instance Learning for Gigapixel Whole Slide Image Classificationcited by 2
  16. 2025Continual Multiple Instance Learning with Enhanced Localization for Histopathological Whole Slide Image Analysiscited by 2
  17. 2025Learning from Majority Label: A Novel Problem in Multi-class Multiple-Instance Learningcited by 2
  18. 2026Deja Vu in Plots: Leveraging Cross-Session Evidence with Retrieval-Augmented LLMs for Live Streaming Risk Assessmentcited by 2
  19. 2026Live or Lie: Action-Aware Capsule Multiple Instance Learning for Risk Assessment in Live Streaming Platformscited by 2
  20. 2026GrapHist: Graph Self-Supervised Learning for Histopathologycited by 2
  21. 2026Beyond Attention Heatmaps: How to Get Better Explanations for Multiple Instance Learning Models in Histopathologycited by 2
  22. 2023Case-level Breast Cancer Prediction for Real Hospital Settingscited by 1
  23. 2024Beyond accuracy: quantifying the reliability of Multiple Instance Learning for Whole Slide Image classificationcited by 1
  24. 2024Advancing Histopathology with Deep Learning Under Data Scarcity: A Decade in Reviewcited by 1
  25. 2025Multimodal Distillation-Driven Ensemble Learning for Long-Tailed Histopathology Whole Slide Images Analysiscited by 1
  26. 2026Are Compact Rationales Free? Measuring Tile Selection Headroom in Frozen WSI-MILcited by 1
  27. 2023Enhancing Instance-Level Image Classification with Set-Level Labelscited by 0
  28. 2024Finding Regions of Interest in Whole Slide Images Using Multiple Instance Learningcited by 0
  29. 2024BAHOP: Similarity-based Basin Hopping for A fast hyper-parameter search in WSI classificationcited by 0
  30. 2024Explaining Black-box Model Predictions via Two-level Nested Feature Attributions with Consistency Propertycited by 0
  31. 2024CARMIL: Context-Aware Regularization on Multiple Instance Learning models for Whole Slide Imagescited by 0
  32. 2025LadderMIL: Multiple Instance Learning with Coarse-to-Fine Self-Distillationcited by 0
  33. 2025Explainable Classifier for Malignant Lymphoma Subtyping via Cell Graph and Image Fusioncited by 0
  34. 2025Leveraging Vision-Language Embeddings for Zero-Shot Learning in Histopathology Imagescited by 0
  35. 2025Histomorphology-Guided Prototypical Multi-Instance Learning for Breast Cancer WSI Classificationcited by 0
  36. 2026Explainable histomorphology-based survival prediction of glioblastoma, IDH-wildtypecited by 0
  37. 2026Enabling Progressive Whole-slide Image Analysis with Multi-scale Pyramidal Networkcited by 0
  38. 2026PA-MIL: Phenotype-Aware Multiple Instance Learning Guided by Language Prompting and Genotype-to-Phenotype Relationshipscited by 0

50 most cited of 58 All 58 (JSON)