# Role-Guided MOE for Encoder-Level Pathology Representation Learning in WSI Classification

arXiv: https://arxiv.org/abs/2609.34897 · HTML: https://mlsanity.com/paper/2609.34897/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Medical and biomedical MoE, one of 12 directions in Mixture-of-experts.
- Builds on 24 of its 36 references that are in the map.
- First posted on arXiv 28 Sep 2026.

## Placed in

- Mixture-of-experts / [Medical and biomedical MoE](https://mlsanity.com/moe/medical-and-biomedical-moe/)

## Builds on (most cited first, 24 of 24)

- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [Training data-efficient image transformers & distillation through attention](https://mlsanity.com/paper/2012.12877/) (2020)
- [Parameter-Efficient Transfer Learning for NLP](https://mlsanity.com/paper/1902.00751/) (2019)
- [Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer](https://mlsanity.com/paper/1701.06538/) (2017)
- [Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity](https://mlsanity.com/paper/2101.03961/) (2021)
- [Prototypical Networks for Few-shot Learning](https://mlsanity.com/paper/1703.05175/) (2017)
- [Reproducible scaling laws for contrastive language-image learning](https://mlsanity.com/paper/2212.07143/) (2022)
- [Attention-based Deep Multiple Instance Learning](https://mlsanity.com/paper/1802.04712/) (2018)
- [DeepSeekMoE: Towards Ultimate Expert Specialization in Mixture-of-Experts Language Models](https://mlsanity.com/paper/2401.06066/) (2024)
- [Scaling Vision with Sparse Mixture of Experts](https://mlsanity.com/paper/2106.05974/) (2021)
- [Data Efficient and Weakly Supervised Computational Pathology on Whole Slide Images](https://mlsanity.com/paper/2004.09666/) (2020)
- [TransMIL: Transformer based Correlated Multiple Instance Learning for Whole Slide Image Classification](https://mlsanity.com/paper/2106.00908/) (2021)
- [Multimodal Whole Slide Foundation Model for Pathology](https://mlsanity.com/paper/2411.19666/) (2024)
- [On the Representation Collapse of Sparse Mixture of Experts](https://mlsanity.com/paper/2204.09179/) (2022)
- [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)
- [Feature Re-Embedding: Towards Foundation Model-Level Performance in Computational Pathology](https://mlsanity.com/paper/2402.17228/) (2024)
- [Dynamic Mixture of Experts: An Auto-Tuning Approach for Efficient Transformer Models](https://mlsanity.com/paper/2405.14297/) (2024)
- [Benchmarking Pathology Foundation Models: Adaptation Strategies and Scenarios](https://mlsanity.com/paper/2410.16038/) (2024)
- [SPIDER: A Comprehensive Multi-Organ Supervised Pathology Dataset and Baseline Models](https://mlsanity.com/paper/2503.02876/) (2025)
- [Free Lunch in Pathology Foundation Model: Task-specific Model Adaptation with Concept-Guided Feature Enhancement](https://mlsanity.com/paper/2411.09894/) (2024)
- [PathoTune: Adapting Visual Foundation Model to Pathological Specialists](https://mlsanity.com/paper/2403.16497/) (2024)
- [M4: Multi-Proxy Multi-Gate Mixture of Experts Network for Multiple Instance Learning in Histopathology Image Analysis](https://mlsanity.com/paper/2407.17267/) (2024)

## Abstract (arXiv)

Whole slide image classification is a fundamental task in computational pathology, where patch representation quality directly affects downstream aggregation and slide-level discriminability. Pathology foundation models are widely adopted as frozen feature extractors for WSI classification; however, their fixed encoders may produce representations insufficiently adapted to target-specific tissue patterns and discriminative cues. Fine-tuning can improve target adaptation, but introduces a trade-off between pathology-specific representation capacity and adaptation efficiency, particularly in data-scarce settings. To address this, we propose a pathology role-guided mixture-of-experts feed-forward network (MoE-FFN) framework for efficient encoder-level representation learning. We design a two-stage training paradigm to establish and adapt pathology-aware expert specialization. In source-domain expert initialization, pathology-specific priors are distilled from a frozen Virchow2 teacher into a lightweight DINOv2-small student, while role prototypes serve as weak pathological anchors to encourage distinct expert functions. MoE-FFN blocks are introduced into selected high-level transformer layers to provide transformation diversity for heterogeneous pathological patterns. In target-domain adaptation, the initialized experts are refined through asymmetric prototype-guided optimization, enhancing task-relevant positive evidence and separating confusable hard negatives. The resulting encoder extracts offline patch representations that can be directly integrated with standard MIL aggregators. Experiments on the public BRACS dataset and a private PAROTID WSI dataset across five representative backbones demonstrate consistent improvements over the strongest baseline.