Robustifying pathology foundation models via fine-tuning

• URL: http://arxiv.org/abs/2607.22861v1
• Date: Fri, 24 Jul 2026 19:05:21 GMT
• Title: Robustifying pathology foundation models via fine-tuning
• Abstract summary: Pathology foundation models (FMs) produce powerful tile-level representations which remain sensitive to scanner and staining variability.<n>We develop a novel fine-tuning recipe that improves the robustness of pathology FMs to acquisition factors.
• Score: 39.255502531644204
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Pathology foundation models (FMs) produce powerful tile-level representations which remain sensitive to scanner and staining variability, undermining deployment across laboratories. We develop a novel fine-tuning recipe that improves the robustness of pathology FMs to acquisition factors. Applied to ten different FMs, our fine-tuning strategy consistently improves robustness for every model as well as downstream performance, with no observed trade-off. On average, it raises the PathoROB robustness index by 23% (from 0.72 to 0.87) and increases the overall cross-benchmark performance by 43% on Patho-Bench, HEST and THUNDER combined, with individual gains reaching up to 72% in robustness (Phikon-v2) and 76% in performance (Midnight-12k). We publicly release the fine-tuned versions of Phikon-v2 (Phaet) and Midnight-12k (Mascaret) at https://huggingface.co/wearewaiv/models.

Related papers

• Robust Lightweight Deep Learning Models for Oral Cancer Screening [2.1173478851307252]
  Oral cancer is a leading cause of mortality in low-to-middle-income countries.<n>We present the optimisation of lightweight deep learning models for smartphone-based oral cancer screening.
  arXiv  Detail & Related papers  (2026-08-21T19:36:25Z)
• Modern Backbones Improve Multi-task DETR for Mammography Classification and Lesion Localization [0.6982666633050445]
  Joint exam-level prediction and candidate-region localization may improve the usefulness of AI support in mammography.<n>We study this setting using a multi-task DETR framework, where shared representations support both image-level malignancy prediction and lesion localization.<n>Modern backbones consistently outperformed older ResNet-style features, with ConvNeXtV2 and DINOv3 giving the strongest overall results.
  arXiv  Detail & Related papers  (2026-08-10T16:22:07Z)
• Cross-Architectural Mixture-of-Experts with Adaptive Soft Routing for Plant Leaf Disease Classification [0.06875312133832077]
  This study proposes an adaptive soft Mixture-of-Experts (MoE) framework with cross-architectural routing.<n>MoE framework integrates EfficientNet-B0, DenseNet-121, and Swin-Tiny to exploit complementary multi-scale, local, and global features.
  arXiv  Detail & Related papers  (2026-06-22T14:58:42Z)
• RepSAM: Bridging Foundation Models to Robotic Vision via Representation-Guided Adaptation [0.0]
  RepSAM is a representation-guided parameter-efficient fine-tuning framework for adapting foundation models to robotic vision.<n>RepSAM achieves 97.9% of full fine-tuning performance (89.0% vs. 90.9% mIoU) while reducing trainable parameters by 158x (from 632M to 4.0M).
  arXiv  Detail & Related papers  (2026-05-25T06:56:42Z)
• VS-DDPM: Efficient Low-Cost Diffusion Model for Medical Modality Translation [5.653007117827187]
  Diffusion models produce high-quality synthetic data but suffer from slow inference.<n>We propose 3D Variable-Step Denoising Diffusion Probabilistic Model (VS-DDPM)<n>We tested our approach on four tasks (missing MRI, tumor removal, MRI-to-sCT, and CBCT-to-sCT) within the BraTS2025 and SynthRAD2025 challenges.
  arXiv  Detail & Related papers  (2026-04-24T18:32:22Z)
• CONFIDE: Hallucination Assessment for Reliable Biomolecular Structure Prediction and Design [46.12506067241116]
  We present CODE (Chain of Diffusion Embeddings), a self evaluating metric to quantify topological frustration.<n>We propose CONFIDE, a unified evaluation framework that combines energetic and topological perspectives.<n>By combining data driven embeddings with theoretical insight, CODE and CONFIDE outperform existing metrics across a wide range of biomolecular systems.
  arXiv  Detail & Related papers  (2025-11-20T03:38:46Z)
• Analyzing the Effects of Supervised Fine-Tuning on Model Knowledge from Token and Parameter Levels [72.3670919950349]
  Large language models (LLMs) acquire substantial world knowledge during pre-training.<n>Post-training techniques such as supervised fine-tuning (SFT) shape this knowledge change behavior.<n>We evaluate closed-book question answering (CBQA) performance across five LLMs from the LLaMA-2 and LLaMA-3 families.
  arXiv  Detail & Related papers  (2025-09-20T09:40:32Z)
• Ensemble YOLO Framework for Multi-Domain Mitotic Figure Detection in Histopathology Images [0.7541656202645494]
  Two modern one-stage detectors, YOLOv5 and YOLOv8, were trained on MIDOG++, CMC, and CCMCT datasets.<n>YOLOv5 achieved higher precision (84.3%), while YOLOv8 offered improved recall (82.6%)<n>Our ensemble ranked 5th with an F1 score of 79.2%, precision of 73.6%, and recall of 85.8%, confirming that the proposed strategy generalizes effectively across unseen test data.
  arXiv  Detail & Related papers  (2025-09-03T02:43:02Z)
• Skywork Open Reasoner 1 Technical Report [51.403686909760914]
  We present Skywork-OR1, an effective and scalable reinforcement learning (RL) implementation for long Chain-of-Thought (CoT) models.<n>Building on the DeepSeek-R1-Distill model series, our RL approach achieves notable performance gains.<n>Our Skywork-OR1-32B model surpasses both DeepSeek-R1 and Qwen3-32B on the AIME24 and AIME25 benchmarks.
  arXiv  Detail & Related papers  (2025-05-28T12:56:04Z)
• Mutual-Taught for Co-adapting Policy and Reward Models [43.11214888109746]
  We propose Mutual-Taught, a self-training method that iteratively improves both the policy model and the reward model.<n> Experimental results demonstrate that this iterative approach leads to consistent improvements in both models.
  arXiv  Detail & Related papers  (2025-05-17T04:34:23Z)
• Contrastive Adapters for Foundation Model Group Robustness [11.961187296775185]
  Existing methods to improve robustness require retraining, which can be prohibitively expensive on large foundation models. We propose contrastive adapting, which trains adapters with contrastive learning to bring sample embeddings close to both their ground-truth class embeddings and other sample embeddings in the same class. Our approach consistently improves group robustness, raising worst-group accuracy by 8.5 to 56.0 pp over zero-shot.
  arXiv  Detail & Related papers  (2022-07-14T19:40:55Z)
• ERNIE-SPARSE: Learning Hierarchical Efficient Transformer Through Regularized Self-Attention [48.697458429460184]
  Two factors, information bottleneck sensitivity and inconsistency between different attention topologies, could affect the performance of the Sparse Transformer. This paper proposes a well-designed model named ERNIE-Sparse. It consists of two distinctive parts: (i) Hierarchical Sparse Transformer (HST) to sequentially unify local and global information, and (ii) Self-Attention Regularization (SAR) to minimize the distance for transformers with different attention topologies.
  arXiv  Detail & Related papers  (2022-03-23T08:47:01Z)
• Towards a Competitive End-to-End Speech Recognition for CHiME-6 Dinner Party Transcription [73.66530509749305]
  In this paper, we argue that, even in difficult cases, some end-to-end approaches show performance close to the hybrid baseline. We experimentally compare and analyze CTC-Attention versus RNN-Transducer approaches along with RNN versus Transformer architectures. Our best end-to-end model based on RNN-Transducer, together with improved beam search, reaches quality by only 3.8% WER abs. worse than the LF-MMI TDNN-F CHiME-6 Challenge baseline.
  arXiv  Detail & Related papers  (2020-04-22T19:08:33Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.