arXiv subjects

Yao Zhang

Publications and source records attributed to Yao Zhang.

At least 109 records · Page 6Linked to original sources

arXiv2025

CX-Mind: A Pioneering Multimodal Large Language Model for Interleaved Reasoning in Chest X-ray via Curriculum-Guided Reinforcement Learning

Chest X-ray (CXR) imaging is one of the most widely used diagnostic modalities in clinical practice, encompassing a broad spectrum of diagnostic tasks. Recent advancements have seen the extensive application of reasoning-based multimodal large language models (MLLMs) in medical imaging to enhance diagnostic efficiency and interpretability. However, existing multimodal models predominantly rely on "one-time" diagnostic approaches, lacking verifiable supervision of the reasoning process. This leads to challenges in multi-task CXR diagnosis, including lengthy reasoning, sparse rewards, and frequent hallucinations. To address these issues, we propose CX-Mind, the first generative model to achieve interleaved "think-answer" reasoning for CXR tasks, driven by curriculum-based reinforcement learning and verifiable process rewards (CuRL-VPR). Specifically, we constructed an instruction-tuning dataset, CX-Set, comprising 708,473 images and 2,619,148 samples, and generated 42,828 high-quality interleaved reasoning data points supervised by clinical reports. Optimization was conducted in two stages under the Group Relative Policy Optimization framework: initially stabilizing basic reasoning with closed-domain tasks, followed by transfer to open-domain diagnostics, incorporating rule-based conditional process rewards to bypass the need for pretrained reward models. Extensive experimental results demonstrate that CX-Mind significantly outperforms existing medical and general-domain MLLMs in visual understanding, text generation, and spatiotemporal alignment, achieving an average performance improvement of 25.1% over comparable CXR-specific models. On real-world clinical dataset (Rui-CXR), CX-Mind achieves a mean recall@1 across 14 diseases that substantially surpasses the second-best results, with multi-center expert evaluations further confirming its clinical utility across multiple dimensions.

cs.LG↗

arXiv2025

Understanding the Supply Chain and Risks of Large Language Model Applications

The rise of Large Language Models (LLMs) has led to the widespread deployment of LLM-based systems across diverse domains. As these systems proliferate, understanding the risks associated with their complex supply chains is increasingly important. LLM-based systems are not standalone as they rely on interconnected supply chains involving pretrained models, third-party libraries, datasets, and infrastructure. Yet, most risk assessments narrowly focus on model or data level, overlooking broader supply chain vulnerabilities. While recent studies have begun to address LLM supply chain risks, there remains a lack of benchmarks for systematic research. To address this gap, we introduce the first comprehensive dataset for analyzing and benchmarking LLM supply chain security. We collect 3,859 real-world LLM applications and perform interdependency analysis, identifying 109,211 models, 2,474 datasets, and 9,862 libraries. We extract model fine-tuning paths, dataset reuse, and library reliance, mapping the ecosystem's structure. To evaluate security, we gather 1,555 risk-related issues-50 for applications, 325 for models, 18 for datasets, and 1,229 for libraries from public vulnerability databases. Using this dataset, we empirically analyze component dependencies and risks. Our findings reveal deeply nested dependencies in LLM applications and significant vulnerabilities across the supply chain, underscoring the need for comprehensive security analysis. We conclude with practical recommendations to guide researchers and developers toward safer, more trustworthy LLM-enabled systems.

cs.SE↗

arXiv2025

Large-Vocabulary Segmentation for Medical Images with Text Prompts

This paper aims to build a model that can Segment Anything in 3D medical images, driven by medical terminologies as Text prompts, termed as SAT. Our main contributions are three-fold: (i) We construct the first multimodal knowledge tree on human anatomy, including 6502 anatomical terminologies; Then, we build the largest and most comprehensive segmentation dataset for training, collecting over 22K 3D scans from 72 datasets, across 497 classes, with careful standardization on both image and label space; (ii) We propose to inject medical knowledge into a text encoder via contrastive learning and formulate a large-vocabulary segmentation model that can be prompted by medical terminologies in text form; (iii) We train SAT-Nano (110M parameters) and SAT-Pro (447M parameters). SAT-Pro achieves comparable performance to 72 nnU-Nets -- the strongest specialist models trained on each dataset (over 2.2B parameters combined) -- over 497 categories. Compared with the interactive approach MedSAM, SAT-Pro consistently outperforms across all 7 human body regions with +7.1% average Dice Similarity Coefficient (DSC) improvement, while showing enhanced scalability and robustness. On 2 external (cross-center) datasets, SAT-Pro achieves higher performance than all baselines (+3.7% average DSC), demonstrating superior generalization ability.

eess.IV↗

arXiv2025

Coalitions on the Fly in Cooperative Games

In this work, we examine a sequential setting of a cooperative game in which players arrive dynamically to form coalitions and complete tasks either together or individually, depending on the value created. Upon arrival, a new player as a decision maker faces two options: forming a new coalition or joining an existing one. We assume that players are greedy, i.e., they aim to maximize their rewards based on the information available at their arrival. The objective is to design an online value distribution policy that incentivizes players to form a coalition structure that maximizes social welfare. We focus on monotone and bounded cooperative games. Our main result establishes an upper bound of $\frac{3\mathsf{min}}{\mathsf{max}}$ on the competitive ratio for any irrevocable policy (i.e., one without redistribution), and proposes a policy that achieves a near-optimal competitive ratio of $\min\left\{\frac{1}{2}, \frac{3\mathsf{min}}{\mathsf{max}}\right\}$, where $\mathsf{min}$ and $\mathsf{max}$ denote the smallest and largest marginal contribution of any sub-coalition of players respectively. Finally, we also consider non-irrevocable policies, with alternative bounds only when the number of players is limited.

cs.GT↗

arXiv2025

H$_2$O$_2$RAM: A High-Performance Hierarchical Doubly Oblivious RAM

The combination of Oblivious RAM (ORAM) with Trusted Execution Environments (TEE) has found numerous real-world applications due to their complementary nature. TEEs alleviate the performance bottlenecks of ORAM, such as network bandwidth and roundtrip latency, and ORAM provides general-purpose protection for TEE applications against attacks exploiting memory access patterns. The defining property of this combination, which sets it apart from traditional ORAM designs, is its ability to ensure that memory accesses, both inside and outside of TEEs, are made oblivious, thus termed doubly oblivious RAM (O$_2$RAM). Efforts to develop O$_2$RAM with enhanced performance are ongoing. In this work, we propose H$_2$O$_2$RAM, a high-performance doubly oblivious RAM construction. The distinguishing feature of our approach, compared to the existing tree-based doubly oblivious designs, is its first adoption of the hierarchical framework that enjoys inherently better data locality and parallelization. While the latest hierarchical solution, FutORAMa, achieves concrete efficiency in the classic client-server model by leveraging a relaxed assumption of sublinear-sized client-side private memory, adapting it to our scenario poses challenges due to the conflict between this relaxed assumption and our doubly oblivious requirement. To this end, we introduce several new efficient oblivious components to build a high-performance hierarchical O$_2$RAM (H$_2$O$_2$RAM). We implement our design and evaluate it on various scenarios. The results indicate that H$_2$O$_2$RAM reduces execution time by up to $\sim 10^3$ times and saves memory usage by $5\sim44$ times compared to state-of-the-art solutions.

cs.CR↗

arXiv2025

It Only Gets Worse: Revisiting DL-Based Vulnerability Detectors from a Practical Perspective

With the growing threat of software vulnerabilities, deep learning (DL)-based detectors have gained popularity for vulnerability detection. However, doubts remain regarding their consistency within declared CWE ranges, real-world effectiveness, and applicability across scenarios. These issues may lead to unreliable detection, high false positives/negatives, and poor adaptability to emerging vulnerabilities. A comprehensive analysis is needed to uncover critical factors affecting detection and guide improvements in model design and deployment. In this paper, we present VulTegra, a novel evaluation framework that conducts a multidimensional comparison of scratch-trained and pre-trained-based DL models for vulnerability detection. VulTegra reveals that state-of-the-art (SOTA) detectors still suffer from low consistency, limited real-world capabilities, and scalability challenges. Contrary to common belief, pre-trained models are not consistently better than scratch-trained models but exhibit distinct strengths in specific contexts.Importantly, our study exposes the limitations of relying solely on CWE-based classification and identifies key factors that significantly affect model performance. Experimental results show that adjusting just one such factor consistently improves recall across all seven evaluated detectors, with six also achieving better F1 scores. Our findings provide deeper insights into model behavior and emphasize the need to consider both vulnerability types and inherent code features for effective detection.

cs.SE↗

arXiv2025

Global Calderón-Zygmund estimates for asymptotically convex fully nonlinear Grad-Mercier type equations

In this paper, we consider the following Dirichlet problem for the fully nonlinear elliptic equation of Grad-Mercier type under asymptotic convexity conditions \begin{equation*} \left\{ \begin{array}{ll} F(D^2u(x),Du(x),u(x),x)=g(|\{y\in Ω:u(y)\ge u(x)\}|)+f(x) & \text{in } Ω, u=ψ&\text{on } \partial Ω. \end{array} \right. \end{equation*} In order to overcome the non-convexity of the operator $F$ and the nonlocality of the nonhomogeneous term $g$, we apply the compactness methods and frozen technique to prove the existence of the $W^{2,p}$-viscosity solutions and the global $W^{2,p}$ estimate. As an application, we derive a Cordes-Nirenberg type continuous estimate up to boundary. Furthermore, we establish a global BMO estimate for the second derivatives of solutions by using an asymptotic approach, thereby refining the borderline case of Calderón-Zygmund estimates.

math.AP↗

arXiv2025

Confidence Scoring for LLM-Generated SQL in Supply Chain Data Extraction

Large Language Models (LLMs) have recently enabled natural language interfaces that translate user queries into executable SQL, offering a powerful solution for non-technical stakeholders to access structured data. However, one of the limitation that LLMs do not natively express uncertainty makes it difficult to assess the reliability of their generated queries. This paper presents a case study that evaluates multiple approaches to estimate confidence scores for LLM-generated SQL in supply chain data retrieval. We investigated three strategies: (1) translation-based consistency checks; (2) embedding-based semantic similarity between user questions and generated SQL; and (3) self-reported confidence scores directly produced by the LLM. Our findings reveal that LLMs are often overconfident in their own outputs, which limits the effectiveness of self-reported confidence. In contrast, embedding-based similarity methods demonstrate strong discriminative power in identifying inaccurate SQL.

stat.AP↗

arXiv2025

SwarmAgentic: Towards Fully Automated Agentic System Generation via Swarm Intelligence

The rapid progress of Large Language Models has advanced agentic systems in decision-making, coordination, and task execution. Yet, existing agentic system generation frameworks lack full autonomy, missing from-scratch agent generation, self-optimizing agent functionality, and collaboration, limiting adaptability and scalability. We propose SwarmAgentic, a framework for fully automated agentic system generation that constructs agentic systems from scratch and jointly optimizes agent functionality and collaboration as interdependent components through language-driven exploration. To enable efficient search over system-level structures, SwarmAgentic maintains a population of candidate systems and evolves them via feedback-guided updates, drawing inspiration from Particle Swarm Optimization (PSO). We evaluate our method on six real-world, open-ended, and exploratory tasks involving high-level planning, system-level coordination, and creative reasoning. Given only a task description and an objective function, SwarmAgentic outperforms all baselines, achieving a +261.8% relative improvement over ADAS on the TravelPlanner benchmark, highlighting the effectiveness of full automation in structurally unconstrained tasks. This framework marks a significant step toward scalable and autonomous agentic system design, bridging swarm intelligence with fully automated system multi-agent generation. Our code is publicly released at https://yaoz720.github.io/SwarmAgentic/.

cs.AI↗

arXiv2025

FedNano: Toward Lightweight Federated Tuning for Pretrained Multimodal Large Language Models

Multimodal Large Language Models (MLLMs) excel in tasks like multimodal reasoning and cross-modal retrieval but face deployment challenges in real-world scenarios due to distributed multimodal data and strict privacy requirements. Federated Learning (FL) offers a solution by enabling collaborative model training without centralizing data. However, realizing FL for MLLMs presents significant challenges, including high computational demands, limited client capacity, substantial communication costs, and heterogeneous client data. Existing FL methods assume client-side deployment of full models, an assumption that breaks down for large-scale MLLMs due to their massive size and communication demands. To address these limitations, we propose FedNano, the first FL framework that centralizes the LLM on the server while introducing NanoEdge, a lightweight module for client-specific adaptation. NanoEdge employs modality-specific encoders, connectors, and trainable NanoAdapters with low-rank adaptation. This design eliminates the need to deploy LLM on clients, reducing client-side storage by 95%, and limiting communication overhead to only 0.01% of the model parameters. By transmitting only compact NanoAdapter updates, FedNano handles heterogeneous client data and resource constraints while preserving privacy. Experiments demonstrate that FedNano outperforms prior FL baselines, bridging the gap between MLLM scale and FL feasibility, and enabling scalable, decentralized multimodal AI systems.

cs.LG↗

arXiv2025

SCOP: Evaluating the Comprehension Process of Large Language Models from a Cognitive View

Despite the great potential of large language models(LLMs) in machine comprehension, it is still disturbing to fully count on them in real-world scenarios. This is probably because there is no rational explanation for whether the comprehension process of LLMs is aligned with that of experts. In this paper, we propose SCOP to carefully examine how LLMs perform during the comprehension process from a cognitive view. Specifically, it is equipped with a systematical definition of five requisite skills during the comprehension process, a strict framework to construct testing data for these skills, and a detailed analysis of advanced open-sourced and closed-sourced LLMs using the testing data. With SCOP, we find that it is still challenging for LLMs to perform an expert-level comprehension process. Even so, we notice that LLMs share some similarities with experts, e.g., performing better at comprehending local information than global information. Further analysis reveals that LLMs can be somewhat unreliable -- they might reach correct answers through flawed comprehension processes. Based on SCOP, we suggest that one direction for improving LLMs is to focus more on the comprehension process, ensuring all comprehension skills are thoroughly developed during training.

cs.CL↗

arXiv2025

Bidirectional magnetization switching of a ferrimagnetic insulator with monochiral molecules

Recent studies have demonstrated that magnetization switching in ferromagnets can be achieved through adsorbing chiral molecules on the surface without the need for current or external magnetic fields, offering a low-power mechanism for applications in spintronic devices. Opposite chirality molecules cause opposite direction reversals of magnetization through the chiral-induced spin selectivity (CISS) mechanism. In this study, we demonstrate bidirectional magnetization switching in thin films of ferrimagnetic insulator TmIG using a single chirality molecule - a Cu metallopolymer of d-leucine. Through UV-VIS circular dichroism and X-ray absorption spectroscopy, we determined that switching between different magnetic orientations is associated with interactions of the d-leucine with the two distinct sublattices of the Fe ions in the TmIG, at octahedral and tetrahedral sites. Our study demonstrates the unexpected versatility of the CISS mechanism for magnetization switching in ferrimagnets using single chirality materials, thereby expanding the potential applications of chiral molecule adsorption-induced magnetization flipping.

physics.app-ph↗

arXiv2025

Knowledge Guided Encoder-Decoder Framework: Integrating Multiple Physical Models for Agricultural Ecosystem Modeling

Agricultural monitoring is critical for ensuring food security, maintaining sustainable farming practices, informing policies on mitigating food shortage, and managing greenhouse gas emissions. Traditional process-based physical models are often designed and implemented for specific situations, and their parameters could also be highly uncertain. In contrast, data-driven models often use black-box structures and does not explicitly model the inter-dependence between different ecological variables. As a result, they require extensive training data and lack generalizability to different tasks with data distribution shifts and inconsistent observed variables. To address the need for more universal models, we propose a knowledge-guided encoder-decoder model, which can predict key crop variables by leveraging knowledge of underlying processes from multiple physical models. The proposed method also integrates a language model to process complex and inconsistent inputs and also utilizes it to implement a model selection mechanism for selectively combining the knowledge from different physical models. Our evaluations on predicting carbon and nitrogen fluxes for multiple sites demonstrate the effectiveness and robustness of the proposed model under various scenarios.

cs.LG↗

arXiv2025

Unlocking User-oriented Pages: Intention-driven Black-box Scanner for Real-world Web Applications

Black-box scanners have played a significant role in detecting vulnerabilities for web applications. A key focus in current black-box scanning is increasing test coverage (i.e., accessing more web pages). However, since many web applications are user-oriented, some deep pages can only be accessed through complex user interactions, which are difficult to reach by existing black-box scanners. To fill this gap, a key insight is that web pages contain a wealth of semantic information that can aid in understanding potential user intention. Based on this insight, we propose Hoyen, a black-box scanner that uses the Large Language Model to predict user intention and provide guidance for expanding the scanning scope. Hoyen has been rigorously evaluated on 12 popular open-source web applications and compared with 6 representative tools. The results demonstrate that Hoyen performs a comprehensive exploration of web applications, expanding the attack surface while achieving about 2x than the coverage of other scanners on average, with high request accuracy. Furthermore, Hoyen detected over 90% of its requests towards the core functionality of the application, detecting more vulnerabilities than other scanners, including unique vulnerabilities in well-known web applications. Our data/code is available at https://hoyen.tjunsl.com/

cs.CR↗

arXiv2025

Lifecycle Management of Optical Networks with Dynamic-Updating Digital Twin: A Hybrid Data-Driven and Physics-Informed Approach

Digital twin (DT) techniques have been proposed for the autonomous operation and lifecycle management of next-generation optical networks. To fully utilize potential capacity and accommodate dynamic services, the DT must dynamically update in sync with deployed optical networks throughout their lifecycle, ensuring low-margin operation. This paper proposes a dynamic-updating DT for the lifecycle management of optical networks, employing a hybrid approach that integrates data-driven and physics-informed techniques for fiber channel modeling. This integration ensures both rapid calculation speed and high physics consistency in optical performance prediction while enabling the dynamic updating of critical physical parameters for DT. The lifecycle management of optical networks, covering accurate performance prediction at the network deployment and dynamic updating during network operation, is demonstrated through simulation in a large-scale network. Up to 100 times speedup in prediction is observed compared to classical numerical methods. In addition, the fiber Raman gain strength, amplifier frequency-dependent gain profile, and connector loss between fiber and amplifier on C and L bands can be simultaneously updated. Moreover, the dynamic-updating DT is verified on a field-trial C+L-band transmission link, achieving a maximum accuracy improvement of 1.4 dB for performance estimation post-device replacement. Overall, the dynamic-updating DT holds promise for driving the next-generation optical networks towards lifecycle autonomous management.

physics.optics↗

arXiv2025

Distillation-Supervised Convolutional Low-Rank Adaptation for Efficient Image Super-Resolution

Convolutional neural networks (CNNs) have been widely used in efficient image super-resolution. However, for CNN-based methods, performance gains often require deeper networks and larger feature maps, which increase complexity and inference costs. Inspired by LoRA's success in fine-tuning large language models, we explore its application to lightweight models and propose Distillation-Supervised Convolutional Low-Rank Adaptation (DSCLoRA), which improves model performance without increasing architectural complexity or inference costs. Specifically, we integrate ConvLoRA into the efficient SR network SPAN by replacing the SPAB module with the proposed SConvLB module and incorporating ConvLoRA layers into both the pixel shuffle block and its preceding convolutional layer. DSCLoRA leverages low-rank decomposition for parameter updates and employs a spatial feature affinity-based knowledge distillation strategy to transfer second-order statistical information from teacher models (pre-trained SPAN) to student models (ours). This method preserves the core knowledge of lightweight models and facilitates optimal solution discovery under certain conditions. Experiments on benchmark datasets show that DSCLoRA improves PSNR and SSIM over SPAN while maintaining its efficiency and competitive image quality. Notably, DSCLoRA ranked first in the Overall Performance Track of the NTIRE 2025 Efficient Super-Resolution Challenge. Our code and models are made publicly available at https://github.com/Yaozzz666/DSCF-SR.

cs.CV↗

arXiv2025

The Tenth NTIRE 2025 Efficient Super-Resolution Challenge Report

This paper presents a comprehensive review of the NTIRE 2025 Challenge on Single-Image Efficient Super-Resolution (ESR). The challenge aimed to advance the development of deep models that optimize key computational metrics, i.e., runtime, parameters, and FLOPs, while achieving a PSNR of at least 26.90 dB on the $\operatorname{DIV2K\_LSDIR\_valid}$ dataset and 26.99 dB on the $\operatorname{DIV2K\_LSDIR\_test}$ dataset. A robust participation saw \textbf{244} registered entrants, with \textbf{43} teams submitting valid entries. This report meticulously analyzes these methods and results, emphasizing groundbreaking advancements in state-of-the-art single-image ESR techniques. The analysis highlights innovative approaches and establishes benchmarks for future research in the field.

cs.CV↗

arXiv2025

FedBiP: Heterogeneous One-Shot Federated Learning with Personalized Latent Diffusion Models

One-Shot Federated Learning (OSFL), a special decentralized machine learning paradigm, has recently gained significant attention. OSFL requires only a single round of client data or model upload, which reduces communication costs and mitigates privacy threats compared to traditional FL. Despite these promising prospects, existing methods face challenges due to client data heterogeneity and limited data quantity when applied to real-world OSFL systems. Recently, Latent Diffusion Models (LDM) have shown remarkable advancements in synthesizing high-quality images through pretraining on large-scale datasets, thereby presenting a potential solution to overcome these issues. However, directly applying pretrained LDM to heterogeneous OSFL results in significant distribution shifts in synthetic data, leading to performance degradation in classification models trained on such data. This issue is particularly pronounced in rare domains, such as medical imaging, which are underrepresented in LDM's pretraining data. To address this challenge, we propose Federated Bi-Level Personalization (FedBiP), which personalizes the pretrained LDM at both instance-level and concept-level. Hereby, FedBiP synthesizes images following the client's local data distribution without compromising the privacy regulations. FedBiP is also the first approach to simultaneously address feature space heterogeneity and client data scarcity in OSFL. Our method is validated through extensive experiments on three OSFL benchmarks with feature space heterogeneity, as well as on challenging medical and satellite image datasets with label heterogeneity. The results demonstrate the effectiveness of FedBiP, which substantially outperforms other OSFL methods.

cs.LG↗