# RetMIL: Retentive Multiple Instance Learning for Histopathological Whole Slide Image Classification

arXiv: https://arxiv.org/abs/2403.10858 · HTML: https://mlsanity.com/paper/2403.10858/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- Builds on 18 of its 28 references that are in the map.
- First posted on arXiv 16 Mar 2024.

## Builds on (most cited first, 18 of 18)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Emerging Properties in Self-Supervised Vision Transformers](https://mlsanity.com/paper/2104.14294/) (2021)
- [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://mlsanity.com/paper/2104.09864/) (2021)
- [Gaussian Error Linear Units (GELUs)](https://mlsanity.com/paper/1606.08415/) (2016)
- [Group Normalization](https://mlsanity.com/paper/1803.08494/) (2018)
- [Attention-based Deep Multiple Instance Learning](https://mlsanity.com/paper/1802.04712/) (2018)
- [Data Efficient and Weakly Supervised Computational Pathology on Whole Slide Images](https://mlsanity.com/paper/2004.09666/) (2020)
- [Dual-stream Multiple Instance Learning Network for Whole Slide Image Classification with Self-supervised Contrastive Learning](https://mlsanity.com/paper/2011.08939/) (2020)
- [Scaling Vision Transformers to Gigapixel Images via Hierarchical Self-Supervised Learning](https://mlsanity.com/paper/2206.02647/) (2022)
- [Searching for Activation Functions](https://mlsanity.com/paper/1710.05941/) (2017)
- [Benchmarking Self-Supervised Learning on Diverse Pathology Datasets](https://mlsanity.com/paper/2212.04690/) (2022)
- [BRACS: A Dataset for BReAst Carcinoma Subtyping in H&E Histology Images](https://mlsanity.com/paper/2111.04740/) (2021)
- [Dynamic Graph Representation with Knowledge-aware Attention for Histopathology Whole Slide Image Analysis](https://mlsanity.com/paper/2403.07719/) (2024)
- [Diagnose Like a Pathologist: Transformer-Enabled Hierarchical Attention-Guided Multiple Instance Learning for Whole Slide Image Classification](https://mlsanity.com/paper/2301.08125/) (2023)
- [HIGT: Hierarchical Interaction Graph-Transformer for Whole Slide Image Analysis](https://mlsanity.com/paper/2309.07400/) (2023)
- [SparseConvMIL: Sparse Convolutional Context-Aware Multiple Instance Learning for Whole Slide Image Classification](https://mlsanity.com/paper/2105.02726/) (2021)
- [Multi-Scale Prototypical Transformer for Whole Slide Image Classification](https://mlsanity.com/paper/2307.02308/) (2023)

## Built on by (most cited first, 7 of 7)

- [Diagnostic Text-guided Representation Learning in Hierarchical Classification for Pathological Whole Slide Image](https://mlsanity.com/paper/2411.10709/) (2024)
- [Towards a Comprehensive Benchmark for Pathological Lymph Node Metastasis in Breast Cancer Sections](https://mlsanity.com/paper/2411.10752/) (2024)
- [Can We Simplify Slide-level Fine-tuning of Pathology Foundation Models?](https://mlsanity.com/paper/2502.20823/) (2025)
- [Multi-View Stenosis Classification Leveraging Transformer-Based Multiple-Instance Learning Using Real-World Clinical Data](https://mlsanity.com/paper/2602.02067/) (2026)
- [Cross-Modal Prototype Allocation: Unsupervised Slide Representation Learning via Patch-Text Contrast in Computational Pathology](https://mlsanity.com/paper/2503.20190/) (2025)
- [Fourier Transform Multiple Instance Learning for Whole Slide Image Classification](https://mlsanity.com/paper/2510.15138/) (2025)
- [WSD-MIL: Window Scale Decay Multiple Instance Learning for Whole Slide Image Classification](https://mlsanity.com/paper/2512.19982/) (2025)

## Abstract (arXiv)

Histopathological whole slide image (WSI) analysis with deep learning has become a research focus in computational pathology. The current paradigm is mainly based on multiple instance learning (MIL), in which approaches with Transformer as the backbone are well discussed. These methods convert WSI tasks into sequence tasks by representing patches as tokens in the WSI sequence. However, the feature complexity brought by high heterogeneity and the ultra-long sequences brought by gigapixel size makes Transformer-based MIL suffer from the challenges of high memory consumption, slow inference speed, and lack of performance. To this end, we propose a retentive MIL method called RetMIL, which processes WSI sequences through hierarchical feature propagation structure. At the local level, the WSI sequence is divided into multiple subsequences. Tokens of each subsequence are updated through a parallel linear retention mechanism and aggregated utilizing an attention layer. At the global level, subsequences are fused into a global sequence, then updated through a serial retention mechanism, and finally the slide-level representation is obtained through a global attention pooling. We conduct experiments on two public CAMELYON and BRACS datasets and an public-internal LUNG dataset, confirming that RetMIL not only achieves state-of-the-art performance but also significantly reduces computational overhead. Our code will be accessed shortly.