搜索医学 AI 论文与资源

ICLR 2026 Poster accepted paper at ICLR 2026. Spectral imaging offers promising applications across diverse domains, including medicine and urban scene understanding, and is already established as a critical modality in remote sensing. However, variability in channel dimensionality and captured wavelengths among spectral cameras impede the development of AI-driven methodologies, leading to camera-specific models with limited generalizability and inadequate cross-camera applicability. To address this bottleneck, we introduce CARL, a model for Camera-Agnostic Representation Learning across RGB, multispectral, and hyperspectral imaging modalities. To enable the conversion of a spectral image with any channel dimensionality to a camera-agnostic representation, we introduce a novel spectral encoder, featuring a self-attention-cross-attention mechanism, to distill salient spectral information into learned spectral representations. Code/project link: https://github.com/IMSY-DKFZ/CARL

ICLR 2026 Poster accepted paper at ICLR 2026. DNA methylation (DNAm) is a key epigenetic modification that regulates gene expression and is pivotal in development and disease. However, profiling DNAm at genome scale is challenging: of $\textasciitilde$28 million CpG sites in the human genome, only about 1–3\% are typically assayed in common datasets due to technological limitations and cost. Recent deep learning approaches, including masking-based generative Transformer models, have shown promise in capturing DNAm–gene expression relationships, but they rely on partially observed DNAm values for unmeasured CpGs and cannot be applied to completely unmeasured samples. To overcome this barrier, we introduce MethylProphet, a gene-guided, context-aware Transformer model for whole-genome DNAm inference without any measured DNAm input.

ICLR 2026 Poster accepted paper at ICLR 2026. Prior-data fitted networks (PFNs) have recently been proposed as a promising way to train tabular foundation models. PFNs are transformers that are pre-trained on synthetic data generated from a prespecified prior distribution and that enable Bayesian inference through in-context learning. In this paper, we introduce CausalFM, a comprehensive framework for training PFN-based foundation models in various causal inference settings. First, we formalize the construction of Bayesian priors for causal inference based on structural causal models (SCMs) in a principled way and derive necessary criteria for the validity of such priors. Building on this, we propose a novel family of prior distributions using causality-inspired Bayesian neural networks that enable CausalFM to perform Bayesian causal inference in various settings, including for back-door, front-door, and instrumental variable adjustment.

ICLR 2026 Poster accepted paper at ICLR 2026. Predicting individualized potential outcomes in sequential decision-making is central for optimizing therapeutic decisions in personalized medicine (e.g., which dosing sequence to give to a cancer patient). However, predicting potential out- comes over long horizons is notoriously difficult. Existing methods that break the curse of the horizon typically lack strong theoretical guarantees such as orthogonality and quasi-oracle efficiency. In this paper, we revisit the problem of predicting individualized potential outcomes in sequential decision-making (i.e., estimating Q-functions in Markov decision processes with observational data) through a causal inference lens.

ICLR 2026 Poster accepted paper at ICLR 2026. Estimating heterogeneous treatment effects (HTEs) from right-censored survival data is critical in high-stakes applications such as precision medicine and individualized policy-making. Yet, the survival analysis setting poses unique challenges for HTE estimation due to censoring, unobserved counterfactuals, and complex identification assumptions. Despite recent advances, from causal survival forests to survival meta-learners and outcome imputation approaches, evaluation practices remain fragmented and inconsistent. We introduce SurvHTE‐Bench, the first comprehensive benchmark for HTE estimation with censored outcomes. The benchmark spans (i) a modular suite of synthetic datasets with known ground truth, systematically varying causal assumptions and survival dynamics, (ii) semi-synthetic datasets that pair real-world covariates with simulated treatments and outcomes, and (iii) real-world datasets from a twin study (with known ground truth) and from an HIV clinical trial.

ICLR 2026 Poster accepted paper at ICLR 2026. Estimating potential outcomes for treatments over time based on observational data is important for personalized decision-making in medicine. However, many existing methods for this task fail to properly adjust for time-varying confounding and thus yield biased estimates. There are only a few neural methods with proper adjustments, but these have inherent limitations (e.g., division by propensity scores that are often close to zero), which result in poor performance. As a remedy, we introduce the iterative G-computation network (IGC-Net). Our IGC-Net is a novel, neural end-to-end model which adjusts for time-varying confounding in order to estimate conditional average potential outcomes (CAPOs) over time.

ICLR 2026 Poster accepted paper at ICLR 2026. In precision medicine, quantitative multi-omic features, topological context, and textual biological knowledge play vital roles in identifying disease-critical signaling pathways and targets, guiding the discovery of novel therapeutics and effective treatment strategies. Existing pipelines capture only one or two of these—numerical omics ignore topological context, text-centric LLMs lack quantitative grounded reasoning, and graph-only models underuse rich node semantics and the generalization power of LLMs—thereby limiting mechanistic interpretability. Although Process Reward Models (PRMs) aim to guide reasoning in LLMs, they remain limited by coarse step definitions, unreliable intermediate evaluation, and vulnerability to reward hacking with added computational cost. These gaps motivate jointly integrating quantitative multi-omic signals, topological structure with node annotations, and literature-scale text via LLMs, using subgraph reasoning as the principle bridge linking numeric evidence, topological knowledge and language context.

ICLR 2026 Poster accepted paper at ICLR 2026. Tasks ranging from sleep staging to clinical diagnosis traditionally rely on standard polysomnography (PSG) devices, bedside monitors and wearable devices, which capture diverse nocturnal biosignals (e.g., EEG, EOG, ECG, SpO$_2$). However, heterogeneity across devices and frequent sensor dropout pose significant challenges for unified modelling of these multimodal signals. We present sleep2vec, a foundation model for diverse and incomplete nocturnal biosignals that learns a shared representation via cross-modal alignment. sleep2vec is contrastively pre-trained on 42,249 overnight recordings spanning nine modalities using a Demography, Age, Site & History-aware InfoNCE objective that incorporates physiological and acquisition metadata (e.g., age, gender, recording site) to dynamically weight negatives and mitigate cohort-specific shortcuts.

ICLR 2026 Poster accepted paper at ICLR 2026. Popular explanation methods often produce unreliable feature importance scores due to missingness bias, a systematic distortion that arises when models are probed with ablated, out-of-distribution inputs. Existing solutions treat this as a deep representational flaw that requires expensive retraining or architectural modifications. In this work, we challenge this assumption and show that missingness bias can be effectively treated as a superficial artifact of the model's output space. We introduce MCal, a lightweight post-hoc method that corrects this bias by fine-tuning a simple linear head on the outputs of a frozen base model.

ICLR 2026 Poster accepted paper at ICLR 2026. Alzheimer’s Disease (AD) is a progressive neurodegenerative disorder and a leading cause of dementia worldwide. Accurate diagnosis requires integrating diverse patient data modalities. With the rapid advancement of foundation models in neurobiology and medicine, integrating foundation models from various modalities has emerged as a promising yet underexplored direction for multi-modal AD diagnosis. A central challenge is enabling effective interaction among these models without disrupting the robust, modality-specific representations learned from large-scale pretraining. To address this, we propose a novel multi-modal framework for AD diagnosis that enables joint interaction among uni-modal foundation models through modality-anchored interaction.

ICLR 2026 Poster accepted paper at ICLR 2026. The application of large language models (LLMs) in the medical field has garnered significant attention, yet their reasoning capabilities in more specialized domains like anesthesiology remain underexplored. To bridge this gap, we introduce AnesSuite, the first comprehensive dataset suite specifically designed for anesthesiology reasoning in LLMs. The suite features AnesBench, an evaluation benchmark tailored to assess anesthesiology-related reasoning across three levels: factual retrieval (System 1), hybrid reasoning (System 1.x), and complex decision-making (System 2). Alongside this benchmark, the suite includes three training datasets that provide an infrastructure for continued pre-training (CPT), supervised fine-tuning (SFT), and reinforcement learning with verifiable rewards (RLVR). Code/project link: https://github.com/MiliLab/AnesSuite

ICLR 2026 Poster accepted paper at ICLR 2026. The goal of personalized medicine is to discover a treatment regimen that optimizes a patient's clinical outcome based on their personal genetic and environmental factors. However, candidate treatments cannot be arbitrarily administered to the patient to assess their efficacy; we often instead have access to an *in silico* surrogate model that approximates the true fitness of a proposed treatment. Unfortunately, such surrogate models have been shown to fail to generalize to previously unseen patient-treatment combinations. We hypothesize that domain-specific prior knowledge—such as medical textbooks and biomedical knowledge graphs—can provide a meaningful alternative signal of the fitness of proposed treatments.

ICLR 2026 Poster accepted paper at ICLR 2026. We present a systematic study of medical-domain interpretability in Large Language Models (LLMs). We study how the LLMs both represent and process medical knowledge through four different interpretability techniques: (1) UMAP projections of intermediate activations, (2) gradient-based saliency with respect to the model weights, (3) layer lesioning/removal and (4) activation patching. We present knowledge maps of five LLMs which show, at a coarse-resolution, where knowledge about patient's ages, medical symptoms, diseases and drugs is stored in the models. In particular for Llama3.3-70B, we find that most medical knowledge is processed in the first half of the model's layers.

步态障碍是帕金森病（PD）中最常见且最具致残性的症状之一，其表现复杂且高度异质。在此，我们提出了一种基于深度学习的框架，利用智能手机录制的视频评估步态障碍。该框架在预测 PD 严重程度方面表现出色，微平均受试者工作特征曲线下面积（AUC）为 0.87，F1 分数为 0.806，与三位临床专家的平均表现相当。此外，它以 73.68%的精度有效区分了药物对步态障碍的整体疗效。特别是，它能够区分统一帕金森病评分量表（UPDRS）分辨率之外的药物诱导的细微粒度步态变化。此外，我们的可解释框架能够提取传统临床使用的运动指标，并发现对疾病进展和药物反应敏感的新数字生物标志物。 这些发现强调了其在临床和家庭环境中高效评估疾病进展的巨大潜力，以及在临床试验中评估疾病修饰效果的潜力，以促进个性化治疗。

系统综述与荟萃分析，评估临床医学中人类与大语言模型协作相对于人类单独工作流的表现，覆盖临床推理、文档和解释等任务；研究指出当前证据仍初步且具有情境依赖性，建议后续开展预注册、务实、多中心并嵌入真实工作流的临床研究。

Nature Medicine article describing a generalist biomedical vision-language foundation model evaluated across multiple biomedical tasks.

Nature Medicine paper on Med-PaLM 2 and expert-level medical question answering with large language models.

This Nature Portfolio / npj Health Systems collection is open for submissions until 2026-06-15. It focuses on software-powered healthcare, including clinical decision-making and AI integration, EHR and health-data infrastructures, telemedicine, remote care, multimodal agentic AI, trustworthiness, responsibility, alignment, and health-system workflows.

This Nature Portfolio / npj Genomic Medicine collection is open for submissions until 2026-06-23. It covers AI-powered genomic medicine, including variant prioritization, pathway inference, AI prediction from clinical assays such as histology, radiology and EHRs, multi-omics, precision oncology, rare diseases, population health, explainability, bias, and clinical implementation.

This Nature Portfolio / npj Digital Medicine collection is open for submissions until 2026-06-03. It invites research on real-world clinical performance of AI, including clinical utility, safety, reliability, equity, generalizability, workflow integration, post-deployment monitoring, transparency, patient-centered outcomes, and clinician-AI interaction.

This Nature Portfolio / npj Digital Medicine collection is open for submissions until 2027-03-24. It seeks clinical AI studies centered on pre-specified actions, including pragmatic trials, cluster-randomized and stepped-wedge designs, generative AI in workflow, causal analyses, target trial emulation, post-market surveillance, drift monitoring, safety, economic evaluation, and lifecycle governance.

This Nature Portfolio / npj Digital Medicine collection is open for submissions until 2026-05-06. It calls for physics-informed machine learning for personalized disease prediction, prevention, and management, including digital twins, physics-informed generative AI, biomedical time-series, signals, images, interpretability, and clinical decision support.

This Nature Portfolio / npj Digital Medicine collection is open for submissions until 2027-04-30. It invites work at the intersection of computational drug repurposing, multimodal biomedical data, and AI, including omics, EHRs, real-world evidence, imaging, digital phenotyping, LLMs, graph neural networks, multimodal transformers, knowledge graphs, generative AI, causal inference, explainability, and clinical translation.

This Nature Portfolio / npj Digital Medicine collection is open for submissions until 2026-07-21. It invites research on AI in sports medicine, including multimodal injury and medical-condition prediction, individualized diagnosis, treatment and rehabilitation, transparent and diverse datasets, open-source explainable AI, and safe AI systems for athlete and exercise health.

This Nature Portfolio / npj Digital Medicine collection is open for submissions until 2026-06-01. It focuses on AI tools that support or deliver therapeutic interventions in mental health, including generative AI therapy bots, reinforcement-learning agents, human-in-the-loop models, clinical validity, safety, ethics, equity, and regulation.

This Nature Portfolio / npj Digital Medicine collection is open for submissions until 2026-07-12. It calls for work on agentic AI in care delivery, including real-time evidence-based decision support, virtual and remote patient care, multimodal and longitudinal clinical data, EHRs, medical imaging, genomics, resource-limited deployment, ethics, regulation, quality, and patient safety.

This Nature Portfolio collection is open for submissions until 2026-05-15. It focuses on AI technologies for population medicine and public health, including infectious-disease early warning, pathogen detection, chronic disease risk stratification, policy simulation, wearable AI, multimodal fusion, federated learning, privacy preservation, and foundation models.

This Diagnostics special issue calls for work on artificial intelligence for health and medicine, in a journal section focused on machine learning and AI in diagnostics. It is relevant to medical imaging, diagnostic pathology and radiology, digital health, rehabilitation, cybersecurity, patient safety, and clinical AI quality. The page lists a manuscript submission deadline of 2026-12-31.

This Frontiers Research Topic calls for work on integrating multi-omics data with clinical information to improve diagnosis, prognosis, and personalized treatment. The page lists a manuscript deadline of 2026-09-14 and is currently accepting articles, making it a relevant journal CFP for clinical translation, multimodal medical AI, and precision medicine.

CCF-Deadlines lists APBC 2026 with papers due 2026-06-15 UTC+0 and conference dates 2026-10-21 to 2026-10-24 in Hsinchu. APBC is directly relevant to medical AI through bioinformatics, computational biology, multi-omics modeling, precision medicine, and AI-assisted biomedical discovery.

IEEE BIBM 2026 covers bioinformatics, biomedicine, and health informatics, including machine learning and AI, biomedical image analysis, biomedical signal analysis, clinical decision support, EHR standards, healthcare knowledge representation, NLP and text mining, and precision medicine. The official CFP lists electronic submission of full papers due 2026-07-05, notification on 2026-09-25, camera-ready on 2026-10-25, and the conference on 2026-12-01 to 2026-12-04 in Dallas.

Stanford Medicine RAISE Health symposium on responsible AI for safe and equitable health care.

Stanford AI in Healthcare Leadership and Strategy is a four-week hybrid program for healthcare leaders. It covers AI strategy, governance, model performance and safety, clinical and operational value, implementation, change management, and responsible scaling of AI solutions in healthcare.

MIT OCW 6.S897 Machine Learning for Healthcare introduces clinical data and machine learning for risk stratification, disease progression modeling, precision medicine, diagnosis, subtype discovery, medical imaging, public health, and clinical workflow improvement.

Harvard Medical School's AI in Clinical Medicine is an online live course for clinicians and allied health professionals. It covers practical AI skills, real-world case studies, clinical integration challenges, ethical considerations, bias, decision-making, and AI regulation.

FutureLearn's Artificial Intelligence for Healthcare course covers AI concepts, tools, ethics, diagnostics, patient care, operational efficiency, personalized medicine, and robotic surgery. The page lists 8 weeks at 2 hours per week and introductory level.

The Broad Institute ML4H Clinical AI Seminar Series features talks from leading experts at the intersection of AI and medicine. Topics include generative and foundation models, ethical and responsible AI, self-supervised learning, medical imaging, digital twins, and real-world clinical applications.

Stanford AIMI Grand Rounds seminar series on artificial intelligence in medicine and imaging.

Harvard Medical School HMX Pro online course on applying artificial intelligence in medicine.

Stanford Center for Artificial Intelligence in Medicine and Imaging course page covering AIMI short courses and programs.

DeepLearning.AI specialization on diagnosis, prognosis, and treatment using medical AI workflows.