arXiv subjects

Yufan Wang

Publications and source records attributed to Yufan Wang.

17 recordsLinked to original sources

arXiv2026

From Analytics to Tumor Boards: An Evidence-Linked Multi-Agent Workflow for Oncology Feature Extraction

Clinically relevant oncology information is distributed across heterogeneous, longitudinal documentation, creating substantial abstraction burden and requiring accurate attribution across specimens, tumors, biomarkers, and time points, while manual cancer-registry abstraction can require 27.2 minutes per case, highlighting the need for scalable methods that preserve clinical context while converting documentation into structured data. We evaluate an oncology information-extraction workflow in which OncoLens supplies multi-source, oncology-aware document selection, aggregation, and normalization from integrated EHRs, while the NimbleMind Multi-Agent System (nMAS) is a configurable oncology information-extraction workflow that extracts clinically relevant structured fields from fragmented oncology documentation. The extraction task uses a clinician-informed schema of 328 attributes spanning report metadata, diagnosis, staging, and cancer-type-specific information. nMAS separates clinician-defined field specifications from model execution and combines complexity-aware extraction, report-level consolidation, and source-grounded validation. The retrospective evaluation included 230 de-identified oncology documents from 40 patients and 418 clinician-reviewed document-field pairs containing 1,126 non-empty reference values. Evaluation focused on fields identified by clinicians as present in the source documents rather than exhaustively annotating all 328 schema fields. nMAS achieved a rank-weighted value-level precision of 82.6%, recall of 87.5%, and F1 of 85.0%, compared with an F1 of 66.4% for an independently implemented UMA-style MiniMax M2.5 comparator. These findings support the feasibility of using a configurable, source-grounded extraction workflow to convert fragmented oncology documentation into reusable structured data.

cs.AI↗

arXiv2026

A Source-Grounded Framework for Constructing and Evaluating Progressive Multimodal Diagnostic Dialogues from Clinical Case Reports

Clinical diagnosis requires progressive integration of patient history, physical examination, laboratory findings, medical images, and diagnostic-informative tests. However, most multimodal medical benchmarks evaluate fixed inputs or endpoint answers, while fully interactive diagnostic agents conflate evidence selection with evidence interpretation. We present a source-grounded framework to construct progressive multimodal diagnostic dialogues from case reports and an evaluation strategy for assessing MLLMs on final diagnosis, diagnostic reasoning, and image-finding interpretation. Evaluation on 24 internal medicine case reports showed that our framework can accurately convert case reports into reference dialogues, achieving a diagnosis F1 of 0.99 and a reasoning-quality score of 4.79 out of 5. Evaluation on two frontier MLLMs (o4-mini and Claude Haiku 4.5) achieved reasoning-quality scores of 2.75 and 2.50, respectively, with substantially lower diagnosis, reasoning, and image-finding F1 scores. The results demonstrate that fluent responses do not necessarily reflect evidence-grounded clinical reasoning and highlight the utility of the proposed framework for evaluating multimodal diagnostic reasoning.

cs.CL↗

arXiv2026

Trust but Verify:Evidence-Linked Multi-Agent Clinical Information Extraction in Pathology

Clinical feature extraction from pathology reports is challenging because relevant evidence may be distributed across coded and narrative fields and depend on specimen attribution, negation, ancillary findings, and diagnostic context. We retrospectively evaluated the NimbleMind Multi-Agent System (nMAS), a configurable workflow that separates clinician-defined field specifications from extraction models and returns report-level predictions with source-linked evidence. The study included 54 dummy gastric biopsy pathology reports from Singapore and four binary target fields, yielding 216 feature-case decisions. nMAS correctly classified 213 of 216 decisions (98.61\%), and all evidence spans associated with correct predictions occurred verbatim in the corresponding source reports. All three errors occurred in the two context-dependent \textit{H. pylori}-related fields requiring negation handling or diagnostic attribution. A single-model UMA-style comparator produced the similar label-level performance and error pattern. These findings do not demonstrate predictive superiority for the multi-agent architecture.Rather, the contribution of nMAS lies in workflow integration and traceability through configurable field specifications, complexity-based routing, report-level aggregation, and source-text validation within a clinician-reviewable workflow. Larger multi-institutional studies should assess generalizability, semantic evidence quality, adaptation effort, and clinician verification time.

cs.AI↗

arXiv2026

Evaluating Multi-Turn Multimodal Diagnostic Reasoning on Challenging Real-World Clinical Cases

Clinical diagnostic evaluation should not only assess whether models can provide correct diagnoses, but also reflect the realities of clinical practice, including progressive disclosure of multimodal information, dynamic updating of diagnostic hypotheses, and continuous refinement of clinical reasoning. However, existing evaluations of multimodal large language models (MLLMs) typically rely on single-turn or isolated tasks, making it difficult to fully capture the complexity of real-world clinical diagnosis. To bridge this gap, we developed ClinMM-Bench, the largest multi-turn multimodal clinical diagnostic evaluation benchmark to date. ClinMM-Bench contains 1,089 challenging real-world clinical cases and 3,760 medical images across eight specialties. We systematically evaluated 15 representative MLLMs using a two-level evaluation framework that assessed both diagnostic accuracy and diagnostic reasoning quality. Results showed that proprietary models achieved the highest overall diagnostic accuracy, but the proportion of completely correct diagnoses remained limited across all models. In terms of diagnostic reasoning quality, current models can identify plausible diagnostic directions but still have considerable limitations in generating reliable diagnostic reasoning. Error analysis further identified five representative failure modes: information synthesis failure, knowledge mapping error, perception error, premature closure, and visual hallucination.

cs.CL↗

arXiv2026

FAF-CD: Frequency-Aware Fusion for Change Detection under Imperfect Multimodal Remote Sensing

Remote sensing change detection for real-world monitoring often relies on imperfect heterogeneous observations, where pre- and post-event images may be asynchronous, cross-sensor, or affected by illumination, seasonal, and modality shifts. This setting is especially challenging for EO-SAR disaster mapping, where nuisance variation can resemble structural damage. We propose FAF-CD, a frequency-aware hybrid framework with a DINOv3-pretrained ConvNeXt encoder and a linear-complexity VMamba-based decoder. Its rectification-aware tri-branch fusion module combines deformable spatial alignment with Fourier and Haar-wavelet comparisons, using adaptive gating to aggregate complementary cues across scales. On BRIGHT validation, a matched heterogeneous EO-SAR adaptation improves clean and perturbed tc-mIoU/tc-mAP over NeXt2Former-CD. FAF-CD also generalizes to binary optical CD, achieving 0.924 cF1 on LEVIR-CD and 0.955 cF1 on WHU-CD, and obtains the best average perturbed cIoU/cF1 on both binary datasets among M-CD and NeXt2Former-CD under pseudo-change-aligned stress tests. It further reduces cost by approximately 24 GFLOPs relative to NeXt2Former-CD while maintaining or improving accuracy.

cs.CV↗

arXiv2026

LLMs Know When They Know, but Do Not Act on It: A Metacognitive Harness for Test-time Scaling

Large language models (LLMs) often expose useful signals of self-monitoring: before solving a problem, they can estimate whether they are likely to succeed, and after solving it, they can judge whether their answer is likely to be correct. However, these signals are typically measured or elicited in isolation, rather than used to control inference. In this work, we ask whether LLMs possess latent metacognitive ability that can be turned into effective test-time control. Inspired by the Nelson--Narens theory from cognitive psychology, we propose a metacognitive harness that separates monitoring from reasoning. For each problem, the model first reports a pre-solve feeling-of-knowing (FOK) signal; after each solve attempt, it reports a post-solve judgment-of-learning (JOL) signal. Rather than treating these signals as passive confidence estimates, the harness turns them into an explicit control interface for reasoning: it decides when to trust the current solution, when to retry with compact metacognitive feedback, and when to pass multiple attempts to a final aggregator. Across text, code, and multimodal reasoning benchmarks, our harness substantially improves a fixed Claude Sonnet-4.6 base model without parameter updates or benchmark-specific fine-tuning. On the evaluated public benchmark snapshots, it raises pooled accuracy from 48.3 to 56.9 and exceeds the strongest listed leaderboard entries on the three primary evaluation settings: HLE-Verified, LiveCodeBench v6, and R-Bench-V. These results suggest that strong LLMs may already possess useful metacognitive ability, but require an explicit control harness to act on it during reasoning.

cs.LG↗

arXiv2026

NeXt2Former-CD: Efficient Remote Sensing Change Detection with Modern Vision Architectures

State Space Models (SSMs) have recently gained traction in remote sensing change detection (CD) for their favorable scaling properties. In this paper, we explore the potential of modern convolutional and attention-based architectures as a competitive alternative. We propose NeXt2Former-CD, an end-to-end framework that integrates a Siamese ConvNeXt encoder initialized with DINOv3 weights, a deformable attention-based temporal fusion module, and a Mask2Former decoder. This design is intended to better tolerate residual co-registration noise and small object-level spatial shifts, as well as semantic ambiguity in bi-temporal imagery. Experiments on LEVIR-CD, WHU-CD, and CDD datasets show that our method achieves the best results among the evaluated methods, improving over recent Mamba-based baselines in both F1 score and IoU. Furthermore, despite a larger parameter count, our model maintains inference latency comparable to SSM-based approaches, suggesting it is practical for high-resolution change detection tasks.

cs.CV↗

arXiv2024

Cold plasma with zirconia nanoparticles for lung cancer via TGF-\b{eta} signaling pathway

Despite advancements in lung cancer therapy, the prognosis for advanced or metastatic patients remains poor, yet many patients eventually develop resistance to standard treatments leading to disease progression and poor survival. Here, we described a combination of CAP and nanoparticles (ZrO2 NPs (zirconium oxide nanoparticle) and 3Y-TZP NPs (3% mol Yttria Tetragonal Zirconia Polycrystal Nanoparticle)) for lung cancer therapy. We found that ZrO2 NPs caused obvious damage to the inside of the lung cancer cells. CAP and ZrO2 NPs mainly affected the mitochondria function, leading to a decrease in mitochondrial membrane potential and ATP levels, and causing endoplasmic reticulum stress and cell nucleus internal DNA damage, etc. CAP combined with ZrO2 NPs (CAP@ZrO2) induced lung cancer cell apoptosis by activating the TGF-\b{eta} pathway. CAP@ZrO2 offers a new therapy for the clinical treatment of lung cancer.

physics.bio-ph↗

arXiv2024

Flavored modular differential equations

Flavored modular differential equations sometimes arise from null states or their descendants in a chiral algebra with continuous flavor symmetry. In this paper we focus on Kac-Moody algebras $\widehat{\mathfrak{g}}_k$ that contain a level-four null state $|\mathcal{N}_T\rangle$ which implements the nilpotency of the Sugawara stress tensor. We study the properties of the corresponding flavored modular differential equations, and show that the equations exhibit almost covariance under modular $S$-transformation, connecting null states and their descendants at different levels. The modular property of the equations fixes the structure of $\mathfrak{g}$ and the level $k$, as well as the flavored characters of all the highest weight representations. Shift property of the equations can generate non-vacuum characters starting from the vacuum character.

hep-th↗

arXiv2024

A Benchmark Grocery Dataset of Realworld Point Clouds From Single View

Fine-grained grocery object recognition is an important computer vision problem with broad applications in automatic checkout, in-store robotic navigation, and assistive technologies for the visually impaired. Existing datasets on groceries are mainly 2D images. Models trained on these datasets are limited to learning features from the regular 2D grids. While portable 3D sensors such as Kinect were commonly available for mobile phones, sensors such as LiDAR and TrueDepth, have recently been integrated into mobile phones. Despite the availability of mobile 3D sensors, there are currently no dedicated real-world large-scale benchmark 3D datasets for grocery. In addition, existing 3D datasets lack fine-grained grocery categories and have limited training samples. Furthermore, collecting data by going around the object versus the traditional photo capture makes data collection cumbersome. Thus, we introduce a large-scale grocery dataset called 3DGrocery100. It constitutes 100 classes, with a total of 87,898 3D point clouds created from 10,755 RGB-D single-view images. We benchmark our dataset on six recent state-of-the-art 3D point cloud classification models. Additionally, we also benchmark the dataset on few-shot and continual learning point cloud classification tasks. Project Page: https://bigdatavision.org/3DGrocery100/.

cs.CV↗

arXiv2023

$\mathcal{N} = 2$ Schur index and line operators

4d $\mathcal{N} = 2$ SCFTs and their invariants can be often enriched by non-local BPS operators. In this paper we study the flavored Schur index of several types of N = 2 SCFTs with and without line operators, using a series of new integration formula of elliptic functions and Eisenstein series. We demonstrate how to evaluate analytically the Schur index for a series of $A_2$ class-$\mathcal{S}$ theories and the $\mathcal{N} = 4$ SO(7) theory. For all $A_1$ class-$\mathcal{S}$ theories we obtain closed-form expressions for SU(2) Wilson line index, and 't Hooft line index in some simple cases. We also observe the relation between the line operator index with the characters of the associated chiral algebras. Wilson line index for some other low rank gauge theories are also studied.

hep-th↗

arXiv2023

Unsupervised Out-of-Distribution Detection with Diffusion Inpainting

Unsupervised out-of-distribution detection (OOD) seeks to identify out-of-domain data by learning only from unlabeled in-domain data. We present a novel approach for this task - Lift, Map, Detect (LMD) - that leverages recent advancement in diffusion models. Diffusion models are one type of generative models. At their core, they learn an iterative denoising process that gradually maps a noisy image closer to their training manifolds. LMD leverages this intuition for OOD detection. Specifically, LMD lifts an image off its original manifold by corrupting it, and maps it towards the in-domain manifold with a diffusion model. For an out-of-domain image, the mapped image would have a large distance away from its original manifold, and LMD would identify it as OOD accordingly. We show through extensive experiments that LMD achieves competitive performance across a broad variety of datasets. Code can be found at https://github.com/zhenzhel/lift_map_detect.

cs.CV↗

arXiv2023

Virtual Therapy Exergame for Upper Extremity Rehabilitation Using Smart Wearable Sensors

Virtual Reality (VR) has been utilized for several applications and has shown great potential for rehabilitation, especially for home therapy. However, these systems solely rely on information from VR hand controllers, which do not fully capture the individual movement of the joints. In this paper, we propose a creative VR therapy exergame for upper extremity rehabilitation using multi-dimensional reaching tasks while simultaneously capturing hand movement from the VR controllers and elbow joint movement from a flexible carbon nanotube sleeve. We conducted a preliminary study with non-clinical participants (n = 12, 7 F). In a 2x2 within-subjects study (orientation (vertical, horizontal) x configuration (flat, curved)), we evaluated the effectiveness and enjoyment of the exergame in different study conditions. The results show that there was a statistically significant difference in terms of task completion time between the two orientations. However, no significant differences were found in the number of mistakes in both orientation and configuration of the virtual exergame. This can lead to customizing therapy while maintaining the same level of intensity. That is, if a patient has restricted lower limb mobility and requires to be seated, they can use the orientations interchangeably. The results of resistance change generated from the carbon nanotube sleeve revealed that the flat configuration in the vertical orientation induced more elbow stretches than the other conditions. Finally, we reported the subjective measures based on questionnaires for usability and user experience in different study conditions. In conclusion, the proposed VR exergame has the potential as a multimodal sensory tool for personalized upper extremity home-based therapy and telerehabilitation.

cs.HC↗

arXiv2023

A principled distributional approach to trajectory similarity measurement

Existing measures and representations for trajectories have two longstanding fundamental shortcomings, i.e., they are computationally expensive and they can not guarantee the `uniqueness' property of a distance function: dist(X,Y) = 0 if and only if X=Y, where $X$ and $Y$ are two trajectories. This paper proposes a simple yet powerful way to represent trajectories and measure the similarity between two trajectories using a distributional kernel to address these shortcomings. It is a principled approach based on kernel mean embedding which has a strong theoretical underpinning. It has three distinctive features in comparison with existing approaches. (1) A distributional kernel is used for the very first time for trajectory representation and similarity measurement. (2) It does not rely on point-to-point distances which are used in most existing distances for trajectories. (3) It requires no learning, unlike existing learning and deep learning approaches. We show the generality of this new approach in three applications: (a) trajectory anomaly detection, (b) anomalous sub-trajectory detection, and (c) trajectory pattern mining. We identify that the distributional kernel has (i) a unique data-dependent property and the above uniqueness property which are the key factors that lead to its superior task-specific performance; and (ii) runtime orders of magnitude faster than existing distance measures.

cs.LG↗

arXiv2022

Surface defects, flavored modular differential equations and modularity

Every 4d $\mathcal{N} = 2$ SCFT $\mathcal{T}$ corresponds to an associated VOA $\mathbb{V}(\mathcal{T})$, which is in general non-rational with a more involved representation theory. Null states in $\mathbb{V}(\mathcal{T})$ can give rise to non-trivial flavored modular differential equations, which must be satisfied by the refined/flavored character of all the $\mathbb{V}(\mathcal{T})$-modules. Taking some $A_1$ theories $\mathcal{T}_{g,n}$ of class-$\mathcal{S}$ as examples, we construct the flavored modular differential equations satisfied by the Schur index. We show that three types of surface defect indices give rise to common solutions to these differential equations, and therefore are sources of $\mathbb{V}(\mathcal{T})$-module characters. These equations transform almost covariantly under modular transformations, ensuring the presence of logarithmic solutions which may correspond to characters of logarithmic modules.

hep-th↗

arXiv2021

Defects, modular differential equations, and free field realization of N = 4 VOAs

For all 4d $\mathcal{N} = 4$ SYM theories with simple gauge groups $G$, we show that the residues of the integrands in the $\mathcal{N} = 4$ Schur indices, which are related to Gukov-Witten type surface defects in the theories, equal the vacuum characters of rank$G$ copies of $bc βγ$ systems that provide the free field realization of associated $\mathcal{N} = 4$ VOAs. This result predicts that these residues, as module characters, are additional solutions to the flavored modular differential equations satisfied by the original Schur index. The prediction is verified in the $G = SU(2)$ case, where an additional logarithmic solution is constructed.

hep-th↗

arXiv2020

Schur correlation functions from q-deformed Yang-Mills

We construct the wave functions in the q-deformed 2d Yang-Mills theory that compute torus correlation functions of affine currents in the VOA associated to a class of 4d $N = 2$ SCFTs. These wave functions are then shown to reduce to the topological correlators of a set of Coulomb branch operators in the $T[SU(N)]$ theory, from which those correlators in the 3d mirror dual of the 4d TN theories can be computed.

hep-th↗