arXiv subjects

Zhe Fang

Publications and source records attributed to Zhe Fang.

6 recordsLinked to original sources

arXiv2026

Agentic AI Enhances Physician Trust in Clinical Decision Making

Medical AI has shifted from reasoning to agentic AI, a new paradigm that autonomously invokes external tools during reasoning, rendering intermediate reasoning steps and tool outputs transparent to users. Although proven to outperform previous models, physician trust in agentic AI remains largely unexplored. To address this, three physicians evaluated 315 multimodal clinical cases quantifying both process-oriented cognitive trust and outcome-oriented behavioral reliance. Comparing agentic AI against non-agentic baselines, physicians exhibited significantly higher cognitive and behavioral trust for the agentic model (P < 0.001). Specifically, on treatment planning tasks, physicians trusted the agentic reasoning most, preferring it in 89.57% of cases. Furthermore, process-oriented cognitive trust is significantly associated with outcome-oriented behavioral reliance (P < 0.001). However, measurable over-reliance on incorrect agentic outputs still exists, highlighting the inherent limitations of decision-logic transparency alone and underscoring the continuous need for rigorous clinician oversight.

cs.CY↗

arXiv2026

Benchmarking and Enabling Efficient Chinese Medical Retrieval via Asymmetric Encoders

Effective medical text retrieval requires both high accuracy and low latency. While LLM-based embedding models possess powerful retrieval capabilities, their prohibitive latency and high computational cost limit their application in real-time scenarios. Furthermore, the lack of comprehensive and high-fidelity benchmarks hinders progress in Chinese medical text retrieval. In this work, we introduce the Chinese Medical Text Embedding Benchmark (CMedTEB), a benchmark spanning three kinds of practical embedding tasks: retrieval, reranking, and semantic textual similarity (STS). Distinct from purely automated datasets, CMedTEB is curated via a rigorous multi-LLM voting pipeline validated by clinical experts, ensuring gold-standard label quality while effectively mitigating annotation noise. On this foundation, we propose the Chinese Medical Asymmetric REtriever (CARE), an asymmetric architecture that pairs a lightweight BERT-style encoder for online query encoding with a powerful LLM-based encoder for offline document encoding. However, optimizing such an asymmetric retriever with two structurally different encoders presents distinctive challenges. To address this, we introduce a novel two-stage training strategy that progressively bridges the query and document representations. Extensive experiments demonstrate that CARE surpasses state-of-the-art symmetric models on CMedTEB, achieving superior retrieval performance without increasing inference latency.

cs.IR↗

arXiv2026

LiveMedBench: A Contamination-Free Medical Benchmark for LLMs with Automated Rubric Evaluation

The deployment of Large Language Models (LLMs) in high-stakes clinical settings demands rigorous and reliable evaluation. However, existing medical benchmarks remain static, suffering from two critical limitations: (1) data contamination, where test sets inadvertently leak into training corpora, leading to inflated performance estimates; and (2) temporal misalignment, failing to capture the rapid evolution of medical knowledge. Furthermore, current evaluation metrics for open-ended clinical reasoning often rely on either shallow lexical overlap (e.g., ROUGE) or subjective LLM-as-a-Judge scoring, both inadequate for verifying clinical correctness. To bridge these gaps, we introduce LiveMedBench, a continuously updated, contamination-free, and rubric-based benchmark that weekly harvests real-world clinical cases from online medical communities, ensuring strict temporal separation from model training data. We propose a Multi-Agent Clinical Curation Framework that filters raw data noise and validates clinical integrity against evidence-based medical principles. For evaluation, we develop an Automated Rubric-based Evaluation Framework that decomposes physician responses into granular, case-specific criteria, achieving substantially stronger alignment with expert physicians than LLM-as-a-Judge. To date, LiveMedBench comprises 2,756 real-world cases spanning 38 medical specialties and multiple languages, paired with 16,702 unique evaluation criteria. Extensive evaluation of 38 LLMs reveals that even the best-performing model achieves only 39.2%, and 84% of models exhibit performance degradation on post-cutoff cases, confirming pervasive data contamination risks. Error analysis further identifies contextual application-not factual knowledge-as the dominant bottleneck, with 35-48% of failures stemming from the inability to tailor medical knowledge to patient-specific constraints.

cs.AI↗

arXiv2024

First-order spin magnetohydrodynamics

Based on recent papers, we discuss the formulation of the first-order relativistic spin magnetohydrodynamics (MHD) with the totally antisymmetric spin current and properties of the anisotropic linear waves awaken near an equilibrium configuration. We show that there appears a critical angle in the momentum direction of the linear waves, where a pair of propagating modes turns into purely diffusive modes. Due to this critical behavior, polynomial solutions do not fully capture the angle dependence of the linear waves.

hep-ph↗

arXiv2024

Anisotropic linear waves and breakdown of the momentum expansion in spin magnetohydrodynamics

We formulate spin magnetohydrodynamics (MHD) by including the magnetic-flux and total angular momentum conservation laws in the hydrodynamic framework. To specify the local angular momentum conservation, we choose the totally antisymmetric spin current. The entropy-current analysis allows for ten dissipative first-order transport coefficients including anisotropic spin relaxation rates and the conversion rate between a vorticity (shear) to a symmetric stress (antisymmetric torque), as well as anisotropic viscosities and resistivities. By employing the linear-mode analysis, we solve the first-order spin MHD equations to determine the dispersion relations with the complete information of anisotropy retained. Our analytic solutions indicate that the small-momentum expansion is spoiled by blow up of the higher-order terms when the angle between the momentum and the magnetic field approaches the right angle. This also reveals the existence of another expansion parameter, and, in light of it, we provide solutions in an alternative series expression beyond the critical angle. We confirm that these two series expansions work well in the appropriate angle ranges as compared with numerical results. Building on our findings regarding the breakdown of the small-momentum expansion in first-order theory, we proceed to discussing how these first-order solutions are modified when we include the relaxation dynamics for dissipative modes with the Israel-Stewart framework. We find that, due to the presence of the critical behavior in the first-order solutions, there remains a diffusive window even after the relaxation dynamics is introduced.

hep-ph↗

arXiv2024

Analytic solutions for the linearized first-order magnetohydrodynamics and implications for causality and stability

We address the linear-mode analysis performed near an equilibrium configuration in the fluid rest frame with a dynamical magnetic field perturbed on a constant configuration. We develop a simple and general algorithm for an analytic solution search that works on an order-by-order basis in the derivative expansion. This method can be applied to general sets of hydrodynamic equations. Applying our method to the first-order relativistic magnetohydrodynamics, we demonstrate that the method finds a complete set of solutions. We obtain two sets of analytic solutions for the four and two coupled modes with seven dissipative transport coefficients. The former set has been missing in the literature for a long time due to the difficulties originating from coupled degrees of freedom and strong anisotropy provided by a magnetic field. The newly developed method resolves these difficulties. We also find that the small-momentum expansions of the solutions break down when the momentum direction is nearly perpendicular to an equilibrium magnetic field due to the presence of another small quantity, that is, a trigonometric function representing the anisotropy. We elaborate on the angle dependence of the solutions and provide alternative series representations that work near the right angle. This identifies the origin of a discrepancy found in recent works. Finally, we discuss the issues of causality and stability based on our analytic solutions and recent developments in the literature.

physics.plasm-ph↗