arXiv subjects

Nan Liu

Publications and source records attributed to Nan Liu.

At least 55 records · Page 3Linked to original sources

arXiv2025

Gender Bias in Large Language Models for Healthcare: Assignment Consistency and Clinical Implications

The integration of large language models (LLMs) into healthcare holds promise to enhance clinical decision-making, yet their susceptibility to biases remains a critical concern. Gender has long influenced physician behaviors and patient outcomes, raising concerns that LLMs assuming human-like roles, such as clinicians or medical educators, may replicate or amplify gender-related biases. Using case studies from the New England Journal of Medicine Challenge (NEJM), we assigned genders (female, male, or unspecified) to multiple open-source and proprietary LLMs. We evaluated their response consistency across LLM-gender assignments regarding both LLM-based diagnosis and models' judgments on the clinical relevance or necessity of patient gender. In our findings, diagnoses were relatively consistent across LLM genders for most models. However, for patient gender's relevance and necessity in LLM-based diagnosis, all models demonstrated substantial inconsistency across LLM genders, particularly for relevance judgements. Some models even displayed a systematic female-male disparity in their interpretation of patient gender. These findings present an underexplored bias that could undermine the reliability of LLMs in clinical practice, underscoring the need for routine checks of identity-assignment consistency when interacting with LLMs to ensure reliable and equitable AI-supported clinical care.

cs.CL↗

arXiv2025

SIM-Shapley: A Stable and Computationally Efficient Approach to Shapley Value Approximation

Explainable artificial intelligence (XAI) is essential for trustworthy machine learning (ML), particularly in high-stakes domains such as healthcare and finance. Shapley value (SV) methods provide a principled framework for feature attribution in complex models but incur high computational costs, limiting their scalability in high-dimensional settings. We propose Stochastic Iterative Momentum for Shapley Value Approximation (SIM-Shapley), a stable and efficient SV approximation method inspired by stochastic optimization. We analyze variance theoretically, prove linear $Q$-convergence, and demonstrate improved empirical stability and low bias in practice on real-world datasets. In our numerical experiments, SIM-Shapley reduces computation time by up to 85% relative to state-of-the-art baselines while maintaining comparable feature attribution quality. Beyond feature attribution, our stochastic mini-batch iterative framework extends naturally to a broader class of sample average approximation problems, offering a new avenue for improving computational efficiency with stability guarantees. Code is publicly available at https://github.com/nliulab/SIM-Shapley.

stat.ML↗

arXiv2025

EVLF-FM: Explainable Vision Language Foundation Model for Medicine

Despite the promise of foundation models in medical AI, current systems remain limited - they are modality-specific and lack transparent reasoning processes, hindering clinical adoption. To address this gap, we present EVLF-FM, a multimodal vision-language foundation model (VLM) designed to unify broad diagnostic capability with fine-grain explainability. The development and testing of EVLF-FM encompassed over 1.3 million total samples from 23 global datasets across eleven imaging modalities related to six clinical specialties: dermatology, hepatology, ophthalmology, pathology, pulmonology, and radiology. External validation employed 8,884 independent test samples from 10 additional datasets across five imaging modalities. Technically, EVLF-FM is developed to assist with multiple disease diagnosis and visual question answering with pixel-level visual grounding and reasoning capabilities. In internal validation for disease diagnostics, EVLF-FM achieved the highest average accuracy (0.858) and F1-score (0.797), outperforming leading generalist and specialist models. In medical visual grounding, EVLF-FM also achieved stellar performance across nine modalities with average mIOU of 0.743 and Acc@0.5 of 0.837. External validations further confirmed strong zero-shot and few-shot performance, with competitive F1-scores despite a smaller model size. Through a hybrid training strategy combining supervised and visual reinforcement fine-tuning, EVLF-FM not only achieves state-of-the-art accuracy but also exhibits step-by-step reasoning, aligning outputs with visual evidence. EVLF-FM is an early multi-disease VLM model with explainability and reasoning capabilities that could advance adoption of and trust in foundation models for real-world clinical deployment.

cs.CV↗

arXiv2025

AURAD: Anatomy-Pathology Unified Radiology Synthesis with Progressive Representations

Medical image synthesis has become an essential strategy for augmenting datasets and improving model generalization in data-scarce clinical settings. However, fine-grained and controllable synthesis remains difficult due to limited high-quality annotations and domain shifts across datasets. Existing methods, often designed for natural images or well-defined tumors, struggle to generalize to chest radiographs, where disease patterns are morphologically diverse and tightly intertwined with anatomical structures. To address these challenges, we propose AURAD, a controllable radiology synthesis framework that jointly generates high-fidelity chest X-rays and pseudo semantic masks. Unlike prior approaches that rely on randomly sampled masks-limiting diversity, controllability, and clinical relevance-our method learns to generate masks that capture multi-pathology coexistence and anatomical-pathological consistency. It follows a progressive pipeline: pseudo masks are first generated from clinical prompts conditioned on anatomical structures, and then used to guide image synthesis. We also leverage pretrained expert medical models to filter outputs and ensure clinical plausibility. Beyond visual realism, the synthesized masks also serve as labels for downstream tasks such as detection and segmentation, bridging the gap between generative modeling and real-world clinical applications. Extensive experiments and blinded radiologist evaluations demonstrate the effectiveness and generalizability of our method across tasks and datasets. In particular, 78% of our synthesized images are classified as authentic by board-certified radiologists, and over 40% of predicted segmentation overlays are rated as clinically useful. All code, pre-trained models, and the synthesized dataset will be released upon publication.

eess.IV↗

arXiv2025

Benchmarking Foundation Models with Multimodal Public Electronic Health Records

Foundation models have emerged as a powerful approach for processing electronic health records (EHRs), offering flexibility to handle diverse medical data modalities. In this study, we present a comprehensive benchmark that evaluates the performance, fairness, and interpretability of foundation models, both as unimodal encoders and as multimodal learners, using the publicly available MIMIC-IV database. To support consistent and reproducible evaluation, we developed a standardized data processing pipeline that harmonizes heterogeneous clinical records into an analysis-ready format. We systematically compared eight foundation models, encompassing both unimodal and multimodal models, as well as domain-specific and general-purpose variants. Our findings demonstrate that incorporating multiple data modalities leads to consistent improvements in predictive performance without introducing additional bias. Through this benchmark, we aim to support the development of effective and trustworthy multimodal artificial intelligence (AI) systems for real-world clinical applications. Our code is available at https://github.com/nliulab/MIMIC-Multimodal.

cs.LG↗

arXiv2025

Enabling Inclusive Systematic Reviews: Incorporating Preprint Articles with Large Language Model-Driven Evaluations

Background. Systematic reviews in comparative effectiveness research require timely evidence synthesis. Preprints accelerate knowledge dissemination but vary in quality, posing challenges for systematic reviews. Methods. We propose AutoConfidence (automated confidence assessment), an advanced framework for predicting preprint publication, which reduces reliance on manual curation and expands the range of predictors, including three key advancements: (1) automated data extraction using natural language processing techniques, (2) semantic embeddings of titles and abstracts, and (3) large language model (LLM)-driven evaluation scores. Additionally, we employed two prediction models: a random forest classifier for binary outcome and a survival cure model that predicts both binary outcome and publication risk over time. Results. The random forest classifier achieved AUROC 0.692 with LLM-driven scores, improving to 0.733 with semantic embeddings and 0.747 with article usage metrics. The survival cure model reached AUROC 0.716 with LLM-driven scores, improving to 0.731 with semantic embeddings. For publication risk prediction, it achieved a concordance index of 0.658, increasing to 0.667 with semantic embeddings. Conclusion. Our study advances the framework for preprint publication prediction through automated data extraction and multiple feature integration. By combining semantic embeddings with LLM-driven evaluations, AutoConfidence enhances predictive performance while reducing manual annotation burden. The framework has the potential to facilitate incorporation of preprint articles during the appraisal phase of systematic reviews, supporting researchers in more effective utilization of preprint resources.

cs.CL↗

arXiv2025

Multimodal, Multi-Disease Medical Imaging Foundation Model (MerMED-FM)

Current artificial intelligence models for medical imaging are predominantly single modality and single disease. Attempts to create multimodal and multi-disease models have resulted in inconsistent clinical accuracy. Furthermore, training these models typically requires large, labour-intensive, well-labelled datasets. We developed MerMED-FM, a state-of-the-art multimodal, multi-specialty foundation model trained using self-supervised learning and a memory module. MerMED-FM was trained on 3.3 million medical images from over ten specialties and seven modalities, including computed tomography (CT), chest X-rays (CXR), ultrasound (US), pathology patches, color fundus photography (CFP), optical coherence tomography (OCT) and dermatology images. MerMED-FM was evaluated across multiple diseases and compared against existing foundational models. Strong performance was achieved across all modalities, with AUROCs of 0.988 (OCT); 0.982 (pathology); 0.951 (US); 0.943 (CT); 0.931 (skin); 0.894 (CFP); 0.858 (CXR). MerMED-FM has the potential to be a highly adaptable, versatile, cross-specialty foundation model that enables robust medical imaging interpretation across diverse medical disciplines.

eess.IV↗

arXiv2025

Soliton resolution for the coupled complex short pulse equation

We address the long-time asymptotics of the solution to the Cauchy problem of ccSP (coupled complex short pulse) equation on the line for decaying initial data that can support solitons. The ccSP system describes ultra-short pulse propagation in optical fibers, which is a completely integrable system and posses a $4\times4$ matrix Wadati--Konno--Ichikawa type Lax pair. Based on the $\bar{\partial}$-generalization of the Deift--Zhou steepest descent method, we obtain the long-time asymptotic approximations of the solution in two kinds of space-time regions under a new scale $(ζ,t)$. The solution of the ccSP equation decays as a speed of $O(t^{-1})$ in the region $ζ/t>\varepsilon$ with any $\varepsilon>0$; while in the region $ζ/t<-\varepsilon$, the solution is depicted by the form of a multi-self-symmetric soliton/composite breather and $t^{-1/2}$ order term arises from self-symmetric soliton/composite breather-radiation interactions as well as an residual error order $O(t^{-1}\ln t)$.

nlin.SI↗

arXiv2025

MMLU-ProX: A Multilingual Benchmark for Advanced Large Language Model Evaluation

Existing large language model (LLM) evaluation benchmarks primarily focus on English, while current multilingual tasks lack parallel questions that specifically assess cross-linguistic reasoning abilities. This dual limitation makes it challenging to comprehensively assess LLMs' performance in the multilingual setting. To fill this gap, we introduce MMLU-ProX, a comprehensive benchmark covering 29 languages, built on an English benchmark. Each language version consists of 11,829 identical questions, enabling direct cross-linguistic comparisons. Additionally, to meet efficient evaluation needs, we provide a lite version containing 658 questions per language. To ensure the high quality of MMLU-ProX, we employ a rigorous development process that involves multiple powerful LLMs for translation, followed by expert review to ensure accurate expression, consistent terminology, and cultural relevance. Building on this, we systematically evaluate 36 state-of-the-art LLMs, including reasoning-enhanced and multilingual-optimized LLMs. The results reveal significant disparities in the multilingual capabilities of LLMs: While they perform well in high-resource languages, their performance declines markedly in low-resource languages, with gaps of up to 24.3%. Through MMLU-ProX, we aim to advance the development of more inclusive AI systems and promote equitable access to technology across global contexts.

cs.CL↗

arXiv2025

The Evolving Landscape of Generative Large Language Models and Traditional Natural Language Processing in Medicine

Natural language processing (NLP) has been traditionally applied to medicine, and generative large language models (LLMs) have become prominent recently. However, the differences between them across different medical tasks remain underexplored. We analyzed 19,123 studies, finding that generative LLMs demonstrate advantages in open-ended tasks, while traditional NLP dominates in information extraction and analysis tasks. As these technologies advance, ethical use of them is essential to ensure their potential in medical applications.

cs.CL↗

arXiv2025

Toward Fair Federated Learning under Demographic Disparities and Data Imbalance

Ensuring fairness is critical when applying artificial intelligence to high-stakes domains such as healthcare, where predictive models trained on imbalanced and demographically skewed data risk exacerbating existing disparities. Federated learning (FL) enables privacy-preserving collaboration across institutions, but remains vulnerable to both algorithmic bias and subgroup imbalance - particularly when multiple sensitive attributes intersect. We propose FedIDA (Fed erated Learning for Imbalance and D isparity A wareness), a framework-agnostic method that combines fairness-aware regularization with group-conditional oversampling. FedIDA supports multiple sensitive attributes and heterogeneous data distributions without altering the convergence behavior of the underlying FL algorithm. We provide theoretical analysis establishing fairness improvement bounds using Lipschitz continuity and concentration inequalities, and show that FedIDA reduces the variance of fairness metrics across test sets. Empirical results on both benchmark and real-world clinical datasets confirm that FedIDA consistently improves fairness while maintaining competitive predictive performance, demonstrating its effectiveness for equitable and privacy-preserving modeling in healthcare. The source code is available on GitHub.

cs.CY↗

arXiv2025

Reverse time-to-death as time-scale in time-to-event analysis for studies of advanced illness and palliative care

Background: Incidence of adverse outcome events rises as patients with advanced illness approach end-of-life. Exposures that tend to occur near end-of-life, e.g., use of wheelchair, oxygen therapy and palliative care, may therefore be found associated with the incidence of the adverse outcomes. We propose a strategy for time-to-event analysis to mitigate the time-varying confounding. Methods: We propose a concept of reverse time-to-death (rTTD) and its use for the time-scale in time-to-event analysis. We used data on community-based palliative care uptake (exposure) and emergency department visits (outcome) among patients with advanced cancer in Singapore to illustrate. We compare the results against that of the common practice of using time-on-study (TOS) as time-scale. Results: Graphical analysis demonstrated that cancer patients receiving palliative care had higher rate of emergency department visits than non-recipients mainly because they were closer to end-of-life, and that rTTD analysis made comparison between patients at the same time-to-death. Analysis of emergency department visits in relation to palliative care using TOS time-scale showed significant increase in hazard ratio estimate when observed time-varying covariates were omitted from statistical adjustment (change-in-estimate=0.38; 95% CI 0.15 to 0.60). There was no such change in otherwise the same analysis using rTTD (change-in-estimate=0.04; 95% CI -0.02 to 0.11), demonstrating the ability of rTTD time-scale to mitigate confounding that intensifies in relation to time-to-death. Conclusion: Use of rTTD as time-scale in time-to-event analysis provides a simple and robust approach to control time-varying confounding in studies of advanced illness, even if the confounders are unmeasured.

stat.ME↗

arXiv2025

Characterizing the Optimal Memory-Rate Tradeoff in Secure Coded Caching for Small Buffer or Small Rate

We consider the secure coded caching problem proposed by Ravindrakumar et. al where no user can obtain information about files other than the one requested. We first propose three new schemes for the three cases of cache size $M=1$, $N=2$ files and arbitrary $K$ users, delivery rate $ R=1$, arbitrary $N$ files and $K$ users, and the general case for arbitrary $N$ files and $K$ users, respectively. Then we derive converse results by characterizing new properties of secure coded caching schemes. As a result, we characterize the two end-points of the optimal memory-rate tradeoff curve for arbitrary number of users and files. Furthermore, for the case of $N=2$ files and arbitrary number of users, we also characterize a segment of the optimal memory-rate tradeoff curve, where the cache size is relatively small.

cs.IT↗

arXiv2025

Demand Private Coded Caching: Small Cache Size

We investigate the demand private coded caching problem, which is an $(N,K)$ coded caching problem with $N$ files, $K$ users, each equipped with a cache of size $M$, and an additional privacy constraint on user demands, i.e., each user can not gain any information about the demands of other users. We focus on scenarios where the size of users' caches is small, aiming to further characterize the fundamental limits of this problem. We first present a new virtual-user-based achievable scheme for arbitrary number of users and files, and two MDS-code-based achievable schemes for the case $N \le K$. With a newly derived converse bound for the case $N \le K$, these proposed schemes lead to the optimal memory-rate tradeoff of the demand private coded caching problem for $M \in \big[0, \frac{N}{(K+1)(N-1)} \big] $ where $N \le K \le 2N-2$, and the optimal memory-rate tradeoff for $M \in \big[0, \frac{1}{K+1} \big] $ where $ K > 2N-2$. Moreover, for the case of 2 files and arbitrary number of users, by deriving another new converse bound, the optimal memory-rate tradeoff is characterized for $M\in \big[0,\frac{2}{K}\big] \cup \big[\frac{2(K-1)}{K+1},2\big]$. Finally, we provide the optimal memory-rate tradeoff of the demand private coded caching problem for 2 files and 3 users.

cs.IT↗

arXiv2025

Capacity of Hierarchical Secure Coded Gradient Aggregation with Straggling Communication Links

The growing privacy concerns in distributed learning have led to the widespread adoption of secure aggregation techniques in distributed machine learning systems, such as federated learning. Motivated by a coded gradient aggregation problem in a user-helper-master hierarchical network setting with straggling communication links, we formulate a new secure hierarchical coded gradient aggregation problem. In our setting, $ K $ users communicate with the master through an intermediate layer of $ N $ helpers, who can communicate with each other. With a resiliency threshold of $ N_r $ for straggling communication links, and at most $ T $ colluding helpers and any number of colluding users, the master aims to recover the sum of all users' gradients while remaining unaware of any individual gradient that exceeds the expected sum. In addition, helpers cannot infer more about users' gradients than what is already known by the colluding users. We propose an achievable scheme where users' upload messages are based on a globally known Vandermonde matrix, and helper communication is facilitated using an extended Vandermonde matrix with special structural properties. A matching converse bound is also derived, establishing the optimal result for this hierarchical coded gradient aggregation problem.

cs.IT↗

arXiv2025

seeBias: A Comprehensive Tool for Assessing and Visualizing AI Fairness

Fairness in artificial intelligence (AI) prediction models is increasingly emphasized to support responsible adoption in high-stakes domains such as health care and criminal justice. Guidelines and implementation frameworks highlight the importance of both predictive accuracy and equitable outcomes. However, current fairness toolkits often evaluate classification performance disparities in isolation, with limited attention to other critical aspects such as calibration. To address these gaps, we present seeBias, an R package for comprehensive evaluation of model fairness and predictive performance. seeBias offers an integrated evaluation across classification, calibration, and other performance domains, providing a more complete view of model behavior. It includes customizable visualizations to support transparent reporting and responsible AI implementation. Using public datasets from criminal justice and healthcare, we demonstrate how seeBias supports fairness evaluations, and uncovers disparities that conventional fairness metrics may overlook. The R package is available on GitHub, and a Python version is under development.

cs.LG↗

arXiv2025

Joint Active and Passive Beamforming Design for IRS-aided MIMO ISAC Based on Sensing Mutual Information

In this paper, we investigate the intelligent reflecting surface (IRS)/reconfigurable intelligent surface (RIS)-aided integrated sensing and communication (ISAC) system based on sensing mutual information (MI). Specifically, the base station (BS) perceives the sensing target via the reflected sensing signal by the IRS, while communicating with the users simultaneously. Our aim is to maximize the sensing MI, subject to the quality of service (QoS) constraints for all communication users, the transmit power constraint at the BS, and the unit-modulus constraint on the IRS's passive reflection. We solve this problem under two cases: one simplified case assuming a line-of-sight (LoS) channel between the BS and IRS and no clutter interference to sensing, and the other generalized case considering the Rician fading channel of the BS-IRS link and the presence of clutter interference to sensing. For the first case, we prove that the dedicated sensing beamformer is unnecessary for improving sensing MI and develop a low-complexity iterative algorithm to jointly optimize the BS and IRS active/passive beamformers. Then, for the second case, we propose an alternative iterative algorithm, which can also be applied to the first case, to solve the beamforming design problem under the general setup. Numerical results are provided to validate the performance of the proposed algorithms, as compared to various benchmark schemes.

eess.SP↗

arXiv2025

Regulatory Science Innovation for Generative AI and Large Language Models in Health and Medicine: A Global Call for Action

The integration of generative AI (GenAI) and large language models (LLMs) in healthcare presents both unprecedented opportunities and challenges, necessitating innovative regulatory approaches. GenAI and LLMs offer broad applications, from automating clinical workflows to personalizing diagnostics. However, the non-deterministic outputs, broad functionalities and complex integration of GenAI and LLMs challenge existing medical device regulatory frameworks, including the total product life cycle (TPLC) approach. Here we discuss the constraints of the TPLC approach to GenAI and LLM-based medical device regulation, and advocate for global collaboration in regulatory science research. This serves as the foundation for developing innovative approaches including adaptive policies and regulatory sandboxes, to test and refine governance in real-world settings. International harmonization, as seen with the International Medical Device Regulators Forum, is essential to manage implications of LLM on global health, including risks of widening health inequities driven by inherent model biases. By engaging multidisciplinary expertise, prioritizing iterative, data-driven approaches, and focusing on the needs of diverse populations, global regulatory science research enables the responsible and equitable advancement of LLM innovations in healthcare.

cs.CY↗