Medical
-
Inductive Visual Logic for Few-Shot Out-Of-Distribution Adaptation in VLMs
2026-09-29 18:25:04Hung-Jen Chen, Yu-Heng Ho, Ting-Yao Huang, Po-Hsiang Hsu, Li-Yu Chen, Chun-Yi Lee, Min SunAbstract
Generative vision-language models (VLMs) such as Qwen-VL and LLaVA achieve strong zero-shot performance on tasks overlapping with their pretraining distribution, yet fail on specialized domains where the required discriminative features were never learned, a regime we term distant out-of-distribution (OOD). Standard adaptation methods cannot overcome this representational absence because they operate within the encoder's existing feature space. However, VLMs retain a robust descriptive capacity even when discrimination collapses: a model that cannot classify a medical scan can still articulate its visual patterns. Exploiting this asymmetry, we introduce Inductive Visual Logic (IVL), a training-free framework that constructs classification knowledge from the model's surviving descriptive ability. IVL extracts visual traits from few-shot support images through dual-mode prompting, combining semantic descriptions with primitive visual observations, and organizes them into per-class trait dictionaries. At inference, hierarchical filtering identifies spatially grounded trait evidence for classification. Across multiple distant-OOD benchmarks, IVL achieves the highest aggregate accuracy under two VLM backbones while producing interpretable, trait-traceable predictions.
Abstract (translated)
URL
https://arxiv.org/abs/2609.38362
PDF
-
Aegis: Generative Gradient Masking for Privacy-Preserving Medical Federated Learning
2026-09-29 18:05:52Chaoyu Zhang, Shanghao Shi, Heng Jin, Ning Wang, Y. Thomas Hou, Wenjing LouAbstract
Federated learning (FL) has become a foundational paradigm for multi-institutional medical AI, allowing hospitals and research centers to jointly train diagnostic models without exchanging patient records. This privacy promise, however, is increasingly contested: a malicious or honest-but-curious server can launch model inversion attacks (MIAs) that reconstruct private patient images directly from shared model updates, and recent scalable, closed-form attacks penetrate even secure aggregation at clinically realistic batch sizes. Existing defenses face an unsatisfactory dilemma. Gradient-perturbation methods such as differential privacy and pruning trade away the diagnostic accuracy on which clinical reliability depends, while cryptographic protocols add system complexity yet still leave updates exposed to these scalable attacks. We propose Aegis, a principled client-side defense that breaks this dilemma without perturbing patient data or modifying the FL protocol. Our key insight is that the success of every known MIA is fundamentally bounded by the local batch size relative to the model's leakage capacity; once this limit is exceeded, distinct samples collide and reconstructions collapse into indistinguishable mixtures. Aegis turns this universal bottleneck into a defense: each client superimposes onto its real update a masking gradient computed on locally synthesized, task-relevant data, deliberately pushing the effective batch beyond the attack's recovery capacity. We complement the design with theoretical convergence guarantees under standard convex assumptions and evaluate Aegis on MNIST, CIFAR-10, and three MedMNIST modalities (chest X-ray, abdominal CT, colon pathology). Aegis neutralizes three state-of-the-art MIAs while preserving model utility and incurring only modest overhead, offering a practical privacy primitive for medical FL.
Abstract (translated)
URL
https://arxiv.org/abs/2609.38339
PDF
-
Evaluation Choices Shape Biomedical ML Claims: A Pediatric Pneumonia Benchmark Case Study
2026-09-29 15:42:15Bhanu Prakash Vangala, Sowmya Guda, Latha Peddi, Navya VangalaAbstract
Biomedical machine learning papers often compress model performance into one headline number. That number can look like a property of the model even when it depends strongly on how the benchmark was evaluated. We study this problem on the widely used Kermany pediatric chest radiograph dataset using nine image classifiers and a controlled evaluation protocol. Under the same protocol, the eight pretrained backbones differ by only 0.026 AUROC. In contrast, changing whether the backbone is frozen or fine-tuned changes AUROC by 0.044 on average, and changing the decision threshold changes balanced accuracy by 0.090 on average. The official test split is also measurably different from the training pool: a partition classifier distinguishes them at AUC 0.697, rising to 0.898 for normal radiographs. Most strikingly, a classifier using only file properties, with no image anatomy, reaches 0.992 balanced accuracy within the training pool but falls to 0.496 on the official test split. Validation-fitted thresholds and calibration also transfer imperfectly. These results show that a high benchmark score can support different conclusions when the split, training policy, threshold, metric, calibration, and uncertainty are not communicated with it. We end with a seven-item reporting recommendation in which each item is tied to an effect measured in the study
Abstract (translated)
URL
https://arxiv.org/abs/2609.37848
PDF
-
A Proposed Rubric for Evaluating Expressed Clinical Reasoning in Large Language Model Responses
2026-09-29 15:19:42Zhangshu Joshua Jiang, Zina Ibrahim, James T. TeoAbstract
Rubrics support the structured evaluation of language models. We propose a rubric for assessing expressed clinical reasoning in model responses, drawing on three bodies of work: medical education assessment frameworks (ART, SCT, Key Feature Problems and OSCE); clinical LLM benchmarks (MedR-Bench, HealthBench, TIMER-Bench, this http URL, PrIME-LLM and PatientSafeBench); and general LLM reasoning evaluation research, including the Factuality-Validity-Coherence-Utility taxonomy, FaithCoT-Bench and C2-Faith. We use groundedness as a clinically oriented adaptation of the taxonomy's factuality category. The rubric brings these concepts together in a multidimensional framework for scoring free-text responses to gold-standard clinical vignettes. It includes provisional behavioural anchors, applicability rules and a separate flag for case-specific safety-critical errors. General-domain frameworks inform its design but are not treated as validated clinical instruments. The rubric does not replace case-specific reference criteria or the task-specific metrics of existing benchmarks. It has not yet been tested for inter-rater reliability, construct validity or clinical utility. Its immediate purpose is to make evaluation decisions explicit and open to scrutiny before empirical testing.
Abstract (translated)
URL
https://arxiv.org/abs/2609.37788
PDF
-
Multi-Site Real-World Performance of Commercial AI for Pulmonary and Incidental Pulmonary Embolism Detection
2026-09-29 15:00:27Aawez Mansuri, Mohammadreza Chavoshi, Theodorus Dapamede, Wasif Bala, Beatrice Brown-Mulry, Rohan Isaac, Bardia Khosravi, Hanzhou Li, Frank Li, John T. Moon, Chad Robichaux, Dan I. G. Cohen-Addad, Ninad V. Salastekar, Janice Newsome, Judy W. Gichoya, Hari TrivediAbstract
Pulmonary embolism (PE) is a leading cause of cardiovascular mortality, yet the real-world performance of FDA-cleared AI detection models remains incompletely characterized. We retrospectively evaluated two FDA-cleared AI algorithms from a single commercial platform (Aidoc Medical BriefCase), one for PE triage on dedicated CT pulmonary angiography (CTPA; n = 30,678) and one for incidental PE (iPE) detection on routine contrast-enhanced CTs (n = 37,191), across a 17-facility academic health system. Reference-standard labels were extracted from radiology reports using a validated LLM pipeline (97% accuracy, kappa = 0.94). The PE model achieved 86.8% sensitivity and 99.1% specificity, with sensitivity declining from 99.3% for saddle emboli to 72.9% for subsegmental PE, and from 89.7% for acute to 65.3% for non-acute PE. The iPE model achieved 73.5% sensitivity and 99.8% specificity. Both models demonstrated lower sensitivity than FDA-clearance benchmarks while exceeding cleared specificity, with diminishing performance for peripheral and non-acute emboli mirroring known human reader limitations and underscoring the need for standardized post-market surveillance of AI-enabled medical devices.
Abstract (translated)
URL
https://arxiv.org/abs/2609.37750
PDF
-
Med-RADIO: Reducing All Medical Domains Into One via Multi-Teacher Distillation
2026-09-29 14:30:37Chu Zhang, Haoyu Jiang, Hongyuan Zhang, Hongbin Liu, Dong YiAbstract
The rapid expansion of large-scale medical datasets and computational resources has driven significant progress in medical foundation models. Given the inherent heterogeneity of medical imaging modalities, current research mainly follows two paths: specialized models optimized for specific modalities, and generalist models designed to handle multiple modalities. However, medical generalist models suffer from both insufficient training data scale relative to natural image generalists and inadequate domain-specific depth relative to medical specialists. Empirically, generalist models establish a cross-modality performance baseline, while specialists define the performance ceiling within their respective domains. To elevate this baseline toward these ceilings, we propose Med-RADIO, a medical multi-teacher distillation framework that Reduces All Domains Into One by compressing complementary expertise from multiple domain-specific teachers into a unified medical vision foundation model. Our method curates both generalist and specialist teachers, allocates modality-aligned distillation streams to reorganize generalist pretraining data so it matches specialist domains, and uses a balanced loss to prevent any single teacher from dominating the distillation process. On internal and external classification benchmarks spanning five modalities, Med-RADIO improves over strong medical generalists under linear probing and remains competitive with representative specialists on most evaluated modalities. Code is available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2609.37682
PDF
-
Correct, Don't Delete: Mitigating Emergent Misalignment with Corrective Supervision
Abstract
Fine-tuning a language model on a narrow set of harmful demonstrations, such as bad medical advice, can make it broadly misaligned on unrelated questions, a phenomenon known as emergent misalignment (EM). The usual defense is to find the offending rows and delete them, but a row locator failed our held-out test and deleting rows helps less than expected. We ask a different question: given a fixed set of poisoned rows, is it better to correct them than to remove them? We fine-tune Qwen2.5-14B-Instruct on a mixture of bad medical advice and benign chat data, select a quarter of the poison rows in advance, and either delete them or replace each with a corrected answer to the same prompt, keeping everything else the same. Replacing the rows cuts the EM rate by about a third and improves answers on held-out medical questions, while deleting the same rows has little measurable effect. The advantage is larger when half the poison rows are corrected, and it holds on a second base model and a second misaligned model organism. The content of the replacement appears to matter: paraphrasing the rows while keeping their bad advice shows no clear benefit, and the correct answers distributed with the dataset appear to do about as well as our rewriter's. Realigning an already-poisoned model with further fine-tuning is known to work, but which data does the work has not been compared directly. We find that a short round of training on corrections beats the same amount of training on generic chat data, that corrections on other medical prompts do roughly as well as corrections of the poisoned prompts themselves, and that instructing the correction writer to model a careful, harm-avoiding assistant adds no measurable benefit over plain corrections. In the settings we tested, correcting harmful training data reduces EM more than deleting it.
Abstract (translated)
URL
https://arxiv.org/abs/2609.37624
PDF
-
TomoTransformer: Towards a Foundation Model for CT Reconstruction
2026-09-29 13:53:27AmirEhsan Khorashadizadeh, Benjam\'in B\'ejarAbstract
Supervised deep learning has advanced sparse-view tomographic reconstruction. However, conventional models, which typically map filtered back-projection (FBP) images or sinograms to clean reconstructions, are brittle under distribution shifts. Because they require retraining whenever projection counts and angles, detector resolutions, or data distributions change, their deployment in real-world applications remains limited. To address this, we introduce TomoTransformer, a transformer-based architecture that treats each \textit{local} filtered projection as an individual token and predicts missing views via self-attention. Crucially, TomoTransformer operates in a \emph{back-projection space} that separates projections across spatial locations, making view interpolation geometrically well-posed and invariant to detector size. This design yields a single foundation model that can process any number of input projections, at arbitrary angular locations and detector dimensions, and query any number of target angles without retraining. Trained on a large-scale dataset spanning diverse medical CT anatomies and natural images, TomoTransformer generalizes effectively across anatomies, materials, and resolutions. Extensive evaluations on several benchmark sparse-view datasets show that TomoTransformer significantly outperforms concurrent multi-purpose models like ViewTrans and matches or exceeds strong protocol-specific baselines, while remaining fully agnostic to the number of input and target projections. Furthermore, the model demonstrates robust zero-shot generalization on real experimental nanoscale brain data collected from an X-ray synchrotron, showcasing its practical utility for real-world applications.
Abstract (translated)
URL
https://arxiv.org/abs/2609.37605
PDF
-
SAM Meets VLM: Parameter-Decoupled Full-Parameter Training for Unified Medical Reasoning and Segmentation
2026-09-29 11:27:05Xuyang Cao, Enyou Liu, Jun Zhao, Zhuoyun Liu, Jintao Fei, LeoAbstract
Medical multimodal large language models (MLLMs) are increasingly expected not only to answer clinical questions, but also to localize the visual evidence behind their predictions. A common strategy connects a vision--language model (VLM) with SAM-style segmentation through a special <SEG> token, yet full-parameter training of this unified architecture is difficult because image-level reasoning and pixel-level segmentation impose different requirements on the shared representation space. To address this issue, we propose a parameter-decoupled training framework for unified medical reasoning and segmentation. The framework treats the <SEG> hidden state as a semantic-to-spatial prompt for the mask decoder and encourages it to become separable from generic language states, reducing ambiguous segmentation prompts and potential disruption to reasoning representations. It first performs medical shallow alignment to adapt visual features to clinical language without disturbing the LLM; then controlled instruction tuning shapes separable <SEG> prompt states, monitored by the Davies--Bouldin Index (DBI), while scaling segmentation gradients entering the language backbone; finally, the SAM branch is specialized with the VLM frozen to improve mask precision without altering reasoning parameters. Experiments on medical referring segmentation, grounding, visual QA, and textual QA benchmarks show that our framework achieves strong language-conditioned segmentation while preserving competitive reasoning ability. Ablations show that two-phase instruction tuning, gradient scaling, and segmentation specialization all contribute to the model.
Abstract (translated)
URL
https://arxiv.org/abs/2609.37283
PDF
-
Chinese-Jev: Bringing System One Model to Chinese-Language Tasks
2026-09-29 08:02:27Zexiao Wang, Zihao Zhang, Xudong Wang, Pan Wang, Ziyi Ye, Haoyu Zhao, Zuxuan Wu, Shuicheng YanAbstract
System One models such as Jev offer an efficient alternative to generative language models for tasks that require decisions rather than open-ended responses. However, existing Jev models exhibit limited Chinese-language decision accuracy, restricting their utility in both general and specialized settings. In this paper, we introduce Chinese-Jev, a System One model that addresses this gap through a unified data processing and training pipeline. Our data processing protocol converts heterogeneous Chinese-language annotations into probability targets over candidate options, enabling a shared training formulation across domains and question formats. To enable efficient inference, Chinese-Jev adopts a lightweight encoder-only backbone for text encoding and learns to score candidate answers through decision-oriented training. To address the misalignment between the pre-training distribution and downstream Chinese-language scenarios, we first train the model on a general-purpose corpus of 10 million examples, then fine-tune it separately for the medical, legal, and financial domains. To evaluate decision accuracy and calibration in both general and domain-specific Chinese-language settings, we introduce Chinese-Jev Bench (CJ-Bench). After first-stage pre-training, Chinese-Jev exceeds the accuracy of the closed-source Jev model by 1.24% on general-domain tasks while achieving a 20.3x speedup. Subsequent domain-specific fine-tuning yields a 4.0% accuracy improvement over Jev in medicine and achieves 92% of Jev's average accuracy across specialized domains, with a 17x speedup and an average latency of only 15 ms per example. We further demonstrate on-device deployment of an INT8-quantized model on mobile devices, achieving an inference latency of approximately 1.0 second per decision. The project is available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2609.36965
PDF
-
OpenTumorBoard: A Real-World Benchmark of Multidisciplinary Tumor Board Discussion Trajectories
2026-09-29 07:50:22Anqi Li, Zhixuan Ge, Yixuan Duan, Jiarong Qian, Chi-Yu Chen, MingYu Lu, Huan-Yu Hsu, Yu Gu, Yue Guo, Sheng Wang, Wei Qiu, Hanwen XuAbstract
Multidisciplinary tumor boards integrate multimodal clinical observations and longitudinal patient histories through specialist discussions, yet benchmarks rarely capture these real-world trajectories. We introduce OpenTumorBoard, a benchmark with 611 patient cases and 19,157 discussion turns across ten specialist roles, transcribed from 12,534 minutes of publicly available tumor board recordings on YouTube. The benchmark evaluates two settings: SPECIALIST TURN, in which an LLM responds to a clinically significant question posed during a real discussion, and BOARD SIMULATION, in which it generates an entire back-and-forth discussion and reaches a consensus on therapy recommendations, surgical plans, next actions and clinical trial matching. Evaluation of 14 general-purpose frontier and medical LLMs reveals substantial limitations: the best models score 3.43 out of 5 in clinical equivalence to specialist answers and 2.78 out of 5 in alignment with recorded board conclusions. Supervised finetuning and reinforcement learning improve performance on a held-out test set, suggesting that real-world discussion trajectories can support model adaptation. Three M.D. experts review a subset of the benchmark, finding high information coverage and factuality of patient cases and strong fidelity of extracted consensus conclusions. We will release OpenTumorBoard and its automated curation pipeline to support the development and evaluation of LLMs for multidisciplinary, personalized cancer decision-making.
Abstract (translated)
URL
https://arxiv.org/abs/2609.32810
PDF
-
STAR-GRPO: Canonical Anchoring and Reliability-First Advantages against Representation-Dependent Reward Hacking
2026-09-29 07:25:06Wan Tian, Zhongyi Li, Xiang Xu, Minhao Zou, Yijie Peng, Fuzhen ZhuangAbstract
Reward hacking occurs when policy optimization exploits a brittle reward interface or an overly permissive proxy objective, improving the training score without improving the underlying response quality. This phenomenon is amplified in group-relative policy optimization: an unsupported reward can shift the group baseline and alter the updates of other rollouts, while post-hoc or purely relative weighting cannot represent group-wide uncertainty. We propose \emph{Self-Tuned Anchored Reliability Group-Relative Policy Optimization} (STAR-GRPO), a reliability-first advantage estimator based on paired assessments of the same rollout. STAR separates the quality signal from its learning influence: score disagreement determines rollout reliability, relative reliability enters a self-tuned robust location--scale fit before group normalization, and absolute group reliability attenuates the resulting bounded advantage. The analysis establishes coordinate and second-moment bounds, characterizes exact centering through the weighted location equation, and gives reliability-dependent attenuation guarantees for outlying rewards. We evaluate STAR-GRPO in two complementary reward-hacking regimes. In token-interface exploitation, STAR prevents runaway optimization of the deployed-interface score while improving the canonical quality signal. In rubric-proxy overoptimization for medical reasoning, STAR improves independent semantic evaluation, narrows the proxy--judge discrepancy, and reduces overclaim while optimizing the same task proxy. Together, these results show that reliability-first normalization offers a principled way to limit unsupported reward influence on both group baselines and policy updates, while retaining the task reward as the optimization target.
Abstract (translated)
URL
https://arxiv.org/abs/2609.36900
PDF
-
RareDx: Controlled Knowledge Integration and Graph-Grounded Policy Optimization for Rare-Disease Diagnosis
2026-09-29 07:14:46Bo Zhang, Yuchen Wang, Dongbai Li, Matthew Yu Heng Wong, Qingkai Zeng, Lijun Wang, Tien-Yin Wong, Peng Cui, Tianyu LiuAbstract
Rare-disease diagnosis is a long-tail reasoning problem: phenotypes are incomplete, individual disorders are sparsely documented, and relevant evidence is distributed across ontologies, gene annotations, and biomedical text. Language models consequently favor common conditions, miss rare candidates, or produce plausible but invalid names. We introduce RareDx, which couples controlled evidence use with knowledge-graph-grounded policy optimization. RareDx-Harness normalizes heterogeneous records into one ranked-diagnosis task and compares direct inference, static retrieval, adaptive tools, and structured phenotype-gene-disease reasoning over a shared knowledge layer. The training pipeline combines Top-10 post-training with RareDx-KGPO, our knowledge-graph-grounded policy optimization method. Its reward projects predictions into a canonical disease graph and integrates curated graded relevance, ontology proximity, biomedical similarity, and phenotype consistency. Vocabulary and output-budget constraints prevent dense partial credit from rewarding fabricated or overlong differentials. Across eight benchmarks, the complete RareDx system centered on Qwen3.5-9B reaches 38.34 macro Hit@10, 1.60 points above GPT-5.5 under the archived protocol; a disjoint validation-selection audit retains a 6.80-point routing gain over Direct on held-out cases. The 27B system reaches 23.53/36.56/40.76 at Hit@1/5/10. Controlled ablations show that retrieval is not uniformly helpful and that controlled routing is central to the gain. These results indicate that structured medical knowledge can turn a compact model into a competitive diagnostic ranker across heterogeneous long-tail settings in clinical practice.
Abstract (translated)
URL
https://arxiv.org/abs/2609.35549
PDF
-
Cyclostationary Phase Conditioning for Medical Time Series Diffusion
2026-09-29 07:12:03Samuel Ruiperez-Campillo, Michele Copetti, Jorge da Silva Goncalves, Sonia Laguna, Thomas Hofmann, Julia E. VogtAbstract
Many physiological time series, such as cardiac and brain recordings, exhibit cyclostationarity: their statistics vary periodically with an underlying cycle phase. Corruption from motion, poor contact, and physiological interference obscures morphology needed for diagnosis, making signal restoration essential. Existing diffusion approaches condition on corrupted observations alone and must learn cyclic structure implicitly. We instead propose two inductive biases which encode cyclostationarity: a shift-covariant wavelet representation and dense per-sample phase conditioning inferred from the corrupted input. We further introduce a training-free cyclostationarity index that quantifies phase structure and predicts when phase conditioning will help. Finally, we propose antithetic coupling of reverse trajectories to reduce sampling variance while achieving comparable performance with fivefold fewer network evaluations. Across modalities, our results show that explicitly encoding measurable cyclic structure improves physiological time-series restoration.
Abstract (translated)
URL
https://arxiv.org/abs/2609.34965
PDF
-
Applying Language Models in Clinical Medicine: Recent Trends and Perspectives
Abstract
The use and applicability of artificial intelligence (AI) in medical research and clinical practice has received increasing attention in the literature over recent years. The emergence of large language models (LLMs) has expanded discussions in regards to applications of AI within healthcare. While traditional deep learning based AI applications in medicine have often focused on specific and defined tasks, LLMs offer broader capabilities and flexibility in working with available data,. At the same time of writing, the integration of LLMs into medical settings raises important questions regarding their reliability, accuracy, transparency, safety, and appropriate role in a medical setting. This text presents and discusses recent talks and articles concerning the application of LLMs in medicine, with particular emphasis on their potential utility in research and clinical practice. It considers both the opportunities offered by these technologies and the challenges associated with their implementation, aiming to provide a perspective on the current and emerging role of LLMs within the medical field.
Abstract (translated)
URL
https://arxiv.org/abs/2609.34780
PDF
-
Reconstructing the Vocal Tract with Differentiable Acoustic Simulation
2026-09-29 05:11:04Eric Ming Chen, Jin Woo Lee, Vincent SitzmannAbstract
The vocal tract is the region of the human body responsible for filtering one's voice to create speech. In this paper, we present a differentiable and GPU accelerated acoustic simulator for the vocal tract. The differentiable simulator synthesizes speech by propagating sound along an acoustic tube model of the vocal tract, and via its gradients, can solve the inverse problem: reconstructing the shape of the vocal tract solely from the sound it produces. Although the inverse mapping between geometry and sound is notoriously non-convex, we discover that gradient descent succeeds with three technical contributions: (1) we design a frequency domain formulation of the vocal tract's fluid dynamics that is 70x more GPU parallelizable than finite differences in time, (2) we integrate a differentiable model for turbulence to synthesize consonants, and (3) similar to prior work in implicit neural representations (INRs) and neural fields, we find that parameterizing the geometry with a neural network accelerates convergence and escapes local minima that trap discrete representations. Because the simulator is differentiable, it is readily integrated with other deep learning pipelines to enable novel linguistics and medical imaging applications. (1) We demonstrate self-supervised autoencoding of vocal tract shapes across 11 languages, and (2) we couple our simulator with a generative model of MRI (magnetic resonance imaging) images to reconstruct one's moving vocal tract from only their speech without paired data.
Abstract (translated)
URL
https://arxiv.org/abs/2609.36737
PDF
-
How Medical VLMs Underutilize Their Vision Encoders: A Dermatology Perspective
2026-09-29 02:42:07Janet Wang, Yunbei Zhang, Xiao Wang, Jihun HammAbstract
Medical Vision-Language Models (VLMs) show significant promise for clinical image understanding, offering accurate diagnosis with interpretable reasoning. However, a critical performance gap exists between their strong vision encoders and the full multimodal model: in dermatology, the MedSigLIP encoder outperforms MedGemma by an average of 10.26 percentage points even when both use zero target-task labels; few-shot linear probing provides further evidence of strong visual representations. This gap motivates an investigation of how visual information is used in end-to-end diagnosis and why plausible-sounding predictions can lack grounding in image evidence. Using dermatology as our primary testbed, we systematically investigate three hypotheses for this phenomenon. We further provide a mechanistic analysis of the model's internal attention patterns, showing that a simple describe-then-decide prompting strategy increases vision attention by 30-40% during generation. Task-specific fine-tuning improves dermatology classification but reduces cross-domain medical question-answering performance in our evaluation. To address these challenges, we combine label-free prompting with low-label encoder-assisted reranking while keeping the VLM frozen. We validate the interventions across five VLM backbones in dermatology and provide supporting representation and attention analyses across additional medical modalities.
Abstract (translated)
URL
https://arxiv.org/abs/2609.36557
PDF
-
Reliability Testing of Medical Model Performance under Distributed Deployment
2026-09-29 02:14:18Yifei Wang, Xiaohan Zhang, Youtao Ding, Tianlin Li, Xiaoyu Zhang, Yida Yang, Li PanAbstract
Distributed inference has become an indispensable part of deploying medical models under practical latency, memory, and throughput constraints. Although modern frameworks improve serving efficiency through tensor parallelism, mixed precision, kernel fusion, and multi-device communication, they are generally assumed to preserve the behavior observed during centralized HuggingFace evaluation. This assumption creates an evaluation-deployment mismatch: a model may pass offline evaluation but produce a different output after the execution stack changes. To address this mismatch, we propose a testing framework and an improved, distributed-execution-sensitive medical-model benchmark that evaluates the same checkpoint and input under a centralized HuggingFace reference and matched distributed deployments. Extensive experiments across language, vision, and multimodal medical models show that execution changes can produce measurable output disagreements. Across supported visual settings, the test success rate ranges from 0.21 to 0.43 for single-modality models and from 0.32 to 0.98 for multimodal models. The benchmark is aimed at extending medical-model evaluation from capability and security to evaluation-deployment consistency.
Abstract (translated)
URL
https://arxiv.org/abs/2609.36525
PDF
-
BRIDGE: Bilevel Retrieval-Credit-Aware Agentic Reinforcement Learning
2026-09-29 01:59:03Quan Xiao, Mingda Liu, Gaowen Liu, Katsuki Fujisawa, Tianyi ChenAbstract
Agentic reinforcement learning (ARL) with verifiable rewards improves the ability of large language models (LLMs) to tackle knowledge-intensive tasks by learning to interleave search and reasoning. However, most existing ARL methods optimize only LLM-generated tokens and treat retrieved evidence as environment observations. This creates an information-credit gap: failures caused by missing or misleading evidence are attributed to the LLM policy rather than to the retriever, which motivates training the LLM and the retriever jointly. In this paper, we show that retrieval and LLM policy learning are order-sensitive: adapting the retriever before optimizing the policy yields a larger reward gain than the reverse order. To preserve this hierarchy while allowing both components to co-adapt, we formulate retrieval-augmented agentic RL as a bilevel optimization problem. To solve it efficiently, we introduce BRIDGE, a memory-efficient first-order bilevel method motivated by a loss-landscape analysis of the RL and retrieval objectives. Across seven open-domain QA benchmarks, BRIDGE achieves the highest average accuracy with both 3B and 7B backbones, improving the multi-hop average over the strongest baseline by 9.6 and 3.4 EM points, respectively. It also achieves the best averaged answer accuracy and reasoning quality across medical QA benchmarks.
Abstract (translated)
URL
https://arxiv.org/abs/2609.36505
PDF
-
EyeVQA: Benchmarking Ophthalmic Vision-Language Models from Recognition to Spatial Grounding
2026-09-29 01:27:10Gujie Shao, Zixun Xie, Xuechun Xing, Ruixiang Wang, Ziyun Lan, Yanlin Qi, Gangyi Zhang, Yuxin Yang, Dawei Li, Haiming TangAbstract
Vision-language models (VLMs) have shown increasing potential for medical image understanding, yet their capabilities in ophthalmic imaging remain insufficiently characterized. Existing ophthalmic datasets are typically designed for individual diseases or specialized tasks, making it difficult to systematically evaluate whether VLMs can move beyond disease recognition toward comparative reasoning and fine-grained spatial grounding. We introduce EyeVQA, a unified visual question answering benchmark for comprehensive evaluation of ophthalmic VLMs. EyeVQA is constructed from 21 available ophthalmic datasets and contains 20,000 question-answer pairs spanning six disease groups and seven question types: Single-Choice, Multi-Select, Variable-Select, True-False, Ranking, Point Location, and Bounding Box. Gold answers are deterministically derived from source-provided diagnoses, severity grades, clinical findings, segmentation masks, bounding boxes, and anatomical landmarks, enabling reproducible evaluation without relying on model-generated annotations. Notably, 44.5% of the questions require reasoning across multiple images, extending evaluation beyond conventional single-image medical VQA. We benchmark fourteen representative general-purpose, scientific, and medically specialized VLMs under a unified zero-shot protocol. The best-performing model only achieves an overall score of 62.8, while substantial gaps remain in spatial grounding and cross-task generalization. These results highlight the limitations of current VLMs in comprehensive ophthalmic visual understanding and establish EyeVQA as a diagnostic benchmark for developing more reliable and spatially grounded ophthalmic multimodal models. The project page is available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2609.32352
PDF
