MedConcept: Unsupervised Concept Discovery for Interpretability in Medical VLMs

• URL: http://arxiv.org/abs/2604.11868v1
• Date: Mon, 13 Apr 2026 17:53:41 GMT
• Title: MedConcept: Unsupervised Concept Discovery for Interpretability in Medical VLMs
• Abstract summary: Interpretability of medical Vision-Language models (VLMs) is essential for trustworthy clinical deployment.<n>We introduce MedConcept, a framework that uncovers latent medical concepts in a fully unsupervised manner.<n>MedConcept translates pretrained VLM representations into pseudo-report-style summaries, enabling physician-level inspection of internal model reasoning.
• Score: 2.2615829161440546
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: While medical Vision-Language models (VLMs) achieve strong performance on tasks such as tumor or organ segmentation and diagnosis prediction, their opaque latent representations limit clinical trust and the ability to explain predictions. Interpretability of these multimodal representations are therefore essential for the trustworthy clinical deployment of pretrained medical VLMs. However, current interpretability methods, such as gradient- or attention-based visualizations, are often limited to specific tasks such as classification. Moreover, they do not provide concept-level explanations derived from shared pretrained representations that can be reused across downstream tasks. We introduce MedConcept, a framework that uncovers latent medical concepts in a fully unsupervised manner and grounds them in clinically verifiable textual semantics. MedConcept identifies sparse neuron-level concept activations from pretrained VLM representations and translates them into pseudo-report-style summaries, enabling physician-level inspection of internal model reasoning. To address the lack of quantitative evaluation in concept-based interpretability, we introduce a quantitative semantic verification protocol that leverages an independent pretrained medical LLM as a frozen external evaluator to assess concept alignment with radiology reports. We define three concept scores, Aligned, Unaligned, and Uncertain, to quantify semantic support, contradiction, or ambiguity relative to radiology reports and use them exclusively for post hoc evaluation. These scores provide a quantitative baseline for assessing interpretability in medical VLMs. All codes, prompt and data to be released on acceptance. Ke

Related papers

• Concept-Grounded Reasoning with Prompt-Driven Localization for Interpretable Structured Report Generation [66.62662330558906]
  CORAL is a multimodal framework that integrates spatial grounding and concept-level supervision into a unified reasoning process.<n>CORAL employs a prompt-driven medical segmentation model to localize lesions and predicts multi-class clinical attributes through a Concept Bottleneck module.<n>Experiments on BUS-CoT and IU X-ray datasets demonstrate consistent improvements in diagnostic accuracy, concept consistency, and report quality over strong general-purpose and medical MLLMs.
  arXiv  Detail & Related papers  (2026-09-14T10:21:23Z)
• MedLayBench-V: A Large-Scale Benchmark for Expert-Lay Semantic Alignment in Medical Vision Language Models [2.0221473384338884]
  We introduce MedLayBench-V, the first large-scale multimodal benchmark dedicated to expert-lay semantic alignment.<n>MedLayBench-V provides a verified foundation for training and evaluating next-generation Med-VLMs capable of bridging the communication divide between clinical experts and patients.
  arXiv  Detail & Related papers  (2026-04-07T11:39:41Z)
• MedicalBench: Evaluating Large Language Models Toward Improved Medical Concept Extraction [1.1371912210771806]
  We present MedicalBench, a benchmark for medical concept extraction with evidence grounding.<n>Built from MIMIC-IV discharge summaries and human-verified ICD-10 codes, the dataset is curated through a large language model (LLM) triage pipeline.<n>MedicalBench provides the first systematic benchmark for implicit, evidence-grounded medical concept extraction.
  arXiv  Detail & Related papers  (2026-04-05T14:11:00Z)
• Vision-Language Models Encode Clinical Guidelines for Concept-Based Medical Reasoning [3.2723884637302443]
  Concept Bottleneck Models (CBMs) are a prominent framework for interpretable AI.<n>We propose MedCBR, a concept-based reasoning framework that integrates clinical guidelines with vision-language and reasoning models.<n>Our framework achieves interpretability and forms an end-to-end bridge from medical image analysis to decision-making.
  arXiv  Detail & Related papers  (2026-03-09T20:39:46Z)
• Concept-Enhanced Multimodal RAG: Towards Interpretable and Accurate Radiology Report Generation [12.226029763256962]
  Radiology Report Generation through Vision-Language Models (VLMs) promises to reduce documentation burden, improve reporting consistency, and accelerate clinical adoption.<n>Existing research treats interpretability and accuracy as separate objectives, with concept-based explainability techniques focusing primarily on transparency.<n>We present Concept-Enhanced Multimodal RAG (CEMRAG), a unified framework that decomposes visual representations into interpretable clinical concepts.
  arXiv  Detail & Related papers  (2026-02-17T15:18:07Z)
• Benchmarking Egocentric Clinical Intent Understanding Capability for Medical Multimodal Large Language Models [48.95516224614331]
  We introduce MedGaze-Bench, the first benchmark leveraging clinician gaze as a Cognitive Cursor to assess intent understanding across surgery, emergency simulation, and diagnostic interpretation.<n>Our benchmark addresses three fundamental challenges: visual homogeneity of anatomical structures, strict temporal-causal dependencies in clinical, and implicit adherence to safety protocols.
  arXiv  Detail & Related papers  (2026-01-11T02:20:40Z)
• Training and Evaluation of Guideline-Based Medical Reasoning in LLMs [7.814266948607376]
  Machine learning for early prediction in medicine has recently shown breakthrough performance.<n>The goal of this paper is to teach LLMs to follow medical consensus guidelines step-by-step in their reasoning and prediction process.
  arXiv  Detail & Related papers  (2025-12-03T14:39:02Z)
• Weakly Supervised Concept Learning with Class-Level Priors for Interpretable Medical Diagnosis [9.002305517166635]
  Prior-guided Concept Predictor (PCP) is a weakly supervised framework that enables concept answer prediction without explicit supervision or reliance on language models.<n>PCP improves concept-level F1-score by over 33% compared to zero-shot baselines.
  arXiv  Detail & Related papers  (2025-11-03T00:43:04Z)
• Uncertainty-Driven Expert Control: Enhancing the Reliability of Medical Vision-Language Models [52.2001050216955]
  Existing methods aim to enhance the performance of Medical Vision Language Model (MedVLM) by adjusting model structure, fine-tuning with high-quality data, or through preference fine-tuning.<n>We propose an expert-in-the-loop framework named Expert-Controlled-Free Guidance (Expert-CFG) to align MedVLM with clinical expertise without additional training.
  arXiv  Detail & Related papers  (2025-07-12T09:03:30Z)
• Uncertainty-aware Medical Diagnostic Phrase Identification and Grounding [72.18719355481052]
  We introduce a novel task called Medical Report Grounding (MRG)<n>MRG aims to directly identify diagnostic phrases and their corresponding grounding boxes from medical reports in an end-to-end manner.<n>We propose uMedGround, a robust and reliable framework that leverages a multimodal large language model to predict diagnostic phrases.
  arXiv  Detail & Related papers  (2024-04-10T07:41:35Z)
• Robust and Interpretable Medical Image Classifiers via Concept Bottleneck Models [49.95603725998561]
  We propose a new paradigm to build robust and interpretable medical image classifiers with natural language concepts. Specifically, we first query clinical concepts from GPT-4, then transform latent image features into explicit concepts with a vision-language model.
  arXiv  Detail & Related papers  (2023-10-04T21:57:09Z)
• Enriching Unsupervised User Embedding via Medical Concepts [51.17532619610099]
  Unsupervised user embedding aims to encode patients into fixed-length vectors without human supervisions. Medical concepts extracted from the clinical notes contain rich connections between patients and their clinical categories. We propose a concept-aware unsupervised user embedding that jointly leverages text documents and medical concepts from two clinical corpora.
  arXiv  Detail & Related papers  (2022-03-20T18:54:05Z)

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.