# Are Compact Rationales Free? Measuring Tile Selection Headroom in Frozen WSI-MIL

arXiv: https://arxiv.org/abs/2605.12575 · HTML: https://mlsanity.com/paper/2605.12575/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 30 of its 40 references that are in the map.
- First posted on arXiv 12 May 2026.

## Builds on (most cited first, 30 of 30)

- [Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localization](https://mlsanity.com/paper/1610.02391/) (2016)
- [Categorical Reparameterization with Gumbel-Softmax](https://mlsanity.com/paper/1611.01144/) (2016)
- [Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computation](https://mlsanity.com/paper/1308.3432/) (2013)
- [The Concrete Distribution: A Continuous Relaxation of Discrete Random Variables](https://mlsanity.com/paper/1611.00712/) (2016)
- [Attention-based Deep Multiple Instance Learning](https://mlsanity.com/paper/1802.04712/) (2018)
- [Concept Bottleneck Models](https://mlsanity.com/paper/2007.04612/) (2020)
- [Selective Classification for Deep Neural Networks](https://mlsanity.com/paper/1705.08500/) (2017)
- [Data Efficient and Weakly Supervised Computational Pathology on Whole Slide Images](https://mlsanity.com/paper/2004.09666/) (2020)
- [BranchyNet: Fast Inference via Early Exiting from Deep Neural Networks](https://mlsanity.com/paper/1709.01686/) (2017)
- [TransMIL: Transformer based Correlated Multiple Instance Learning for Whole Slide Image Classification](https://mlsanity.com/paper/2106.00908/) (2021)
- [Evaluating the visualization of what a Deep Neural Network has learned](https://mlsanity.com/paper/1509.06321/) (2015)
- [Is Attention Interpretable?](https://mlsanity.com/paper/1906.03731/) (2019)
- [Scaling Vision Transformers to Gigapixel Images via Hierarchical Self-Supervised Learning](https://mlsanity.com/paper/2206.02647/) (2022)
- [Rationalizing Neural Predictions](https://mlsanity.com/paper/1606.04155/) (2016)
- [Interpretable Neural Predictions with Differentiable Binary Variables](https://mlsanity.com/paper/1905.08160/) (2019)
- [Learning to Deceive with Attention-Based Explanations](https://mlsanity.com/paper/1909.07913/) (2019)
- [Rethinking Cooperative Rationalization: Introspective Extraction and Complement Control](https://mlsanity.com/paper/1910.13294/) (2019)
- [Attention-Challenging Multiple Instance Learning for Whole Slide Image Classification](https://mlsanity.com/paper/2311.07125/) (2023)
- [Multiple Instance Learning Framework with Masked Hard Instance Mining for Whole Slide Image Classification](https://mlsanity.com/paper/2307.15254/) (2023)
- [Additive MIL: Intrinsically Interpretable Multiple Instance Learning for Pathology](https://mlsanity.com/paper/2206.01794/) (2022)
- [Multiple Instance Learning for Digital Pathology: A Review on the State-of-the-Art, Limitations & Future Potential](https://mlsanity.com/paper/2206.04425/) (2022)
- [Deep Learning for Whole Slide Image Analysis: An Overview](https://mlsanity.com/paper/1910.11097/) (2019)
- [SI-MIL: Taming Deep MIL for Self-Interpretability in Gigapixel Histopathology](https://mlsanity.com/paper/2312.15010/) (2023)
- [Rationalization for Explainable NLP: A Survey](https://mlsanity.com/paper/2301.08912/) (2023)
- [xMIL: Insightful Explanations for Multiple Instance Learning in Histopathology](https://mlsanity.com/paper/2406.04280/) (2024)
- [Rethinking Multiple Instance Learning for Whole Slide Image Classification: A Bag-Level Classifier is a Good Instance-Level Teacher](https://mlsanity.com/paper/2312.01099/) (2023)
- [How Effective Can Dropout Be in Multiple Instance Learning ?](https://mlsanity.com/paper/2504.14783/) (2025)
- [ASMIL: Attention-Stabilized Multiple Instance Learning for Whole Slide Imaging](https://mlsanity.com/paper/2603.06658/) (2026)
- [Boosting Explainability through Selective Rationalization in Pre-trained Language Models](https://mlsanity.com/paper/2501.03182/) (2025)
- [ReaMIL: Reasoning- and Evidence-Aware Multiple Instance Learning for Whole-Slide Histopathology](https://mlsanity.com/paper/2601.10073/) (2026)

## Built on by (most cited first, 1 of 1)

- [CHARTER: Auditing Reference Substitution in Hierarchical Compact-Evidence Evaluation for Computational Pathology](https://mlsanity.com/paper/2610.07843/) (2026)

## Abstract (arXiv)

Whole-slide image (WSI) multiple instance learning (MIL) classifiers can achieve strong slide-level AUC while leaving the full-bag prediction opaque. Attention scores are widely reused as post-hoc explanations, but high attention can reflect aggregation preference rather than a compact, model-sufficient rationale. We study post-hoc rationale highlighting for frozen WSI-MIL: given a trained classifier, can its slide-level prediction be recovered from a compact, output-consistent tile subset without retraining the backbone? We instantiate this with Finding Optimal Contextual Instances (FOCI), a lightweight rationale-readout layer over a frozen MIL backbone. FOCI is trained with model-output sufficiency and exclusion objectives over keep/drop tile subsets, evaluated with an insertion-style Sequential Reveal Protocol (SRP) adapted to WSI-MIL, and summarized by the Selection Headroom Index (SHI). Across three WSI benchmarks and seven MIL backbones, FOCI reveals that compact rationales are selection-headroom dependent: transformer and multi-branch attention aggregators can admit compact rationales, near-minimal attention-pooling baselines enter a selection-saturation regime, and hard-selection backbones can conflict with an external readout. For TransMIL, relative to its documented CLS-proxy ranking, FOCI reduces the Minimum Sufficient K (MSK) tile count by 32-56% across benchmarks, while ACMIL+FOCI attains the highest mean SHI (+0.465). Deletion-based perturbation and selected-only downstream evaluation provide complementary checks. These results position FOCI as a model-level interpretability and audit layer: selected tiles are not claims of clinical or pathologist-level diagnostic sufficiency, but candidate rationales that offer a compact, reviewable view of when a frozen MIL prediction can be localized to a small output-consistent subset.