arXiv subjects

Junde Wu

Publications and source records attributed to Junde Wu.

At least 19 recordsLinked to original sources

arXiv2026

AI Can Be Easily Persuaded in Clinical Decision Making

As AI becomes increasingly integrated into clinical practice, it is playing a growing role in medical decision making. Medicine, however, is a high stakes and evidence based field, where decisions can directly affect patients' lives. It is therefore important to understand whether AI can maintain objective judgment when others try to persuade it. In this paper, we study how easily AI can be persuaded through controlled experiments. We find that professional authority, national background, institutional affiliation, claimed past performance, multiple physicians, supported clinician views, and repeated pressure can all affect AI decisions. Surprisingly, the same persuasive input changes about 10% more cases when it comes from a senior clinician than from a medical student. Simply claiming a better performance history consistently makes the physician more persuasive. More strikingly, a plausible clinician view can persuade AI away from a correct decision even when it is fabricated to support an incorrect answer. This indicates that AI can be strongly influenced by convincing support without reliably determining whether this view from the clinician is correct. Together, these findings suggest that AI can be easily persuaded by what people say, who says it, and how the opinion is presented. Therefore, it is essential for AI to maintain sound judgment under persuasion, enabling its safe and reliable use in high stakes medical decision making.

cs.CL↗

arXiv2026

J-CoT: Chain-of-Thought in J-Space

Chain-of-thought prompting improves language-model reasoning by carrying intermediate states across successive computation steps. However, relying on natural language as the only recurrent interface is overly restrictive, since many transient computations do not need to be fully verbalized. Existing latent-reasoning methods remove this constraint by recurrently propagating continuous hidden states. However, these methods pass a dense hidden vector as a whole, without an explicit mechanism for selecting and organizing the information needed by the next reasoning step. This motivates an intermediate interface that remains linguistically grounded without requiring a decoded sentence. We introduce \textbf{J-CoT}, a recurrent reasoning framework built on \emph{J-space}, a vocabulary-indexed coordinate system within the model's hidden representations. Within each cycle, the model computes in its full hidden space. At the cycle boundary, J-CoT expresses the intermediate state as vocabulary-indexed coefficients, carries these coefficients forward as a \emph{J-thought}, and maps them back into the model's hidden representation for the next cycle. J-CoT therefore requires neither a fluent intermediate rationale nor recurrence over the complete hidden state. Under matched backbone and inference settings, J-CoT-Zero matches or exceeds the strongest evaluated latent-reasoning baseline on every benchmark, while J-CoT-Train obtains the highest score across the evaluated mathematical, scientific, coding, and structured path-reasoning tasks.

cs.CL↗

arXiv2026

Mean-State Entropy Hierarchies and Classical Communication through Quantum Convolutions

Quantum convolution provides a discrete-variable analogue of classical convolution, with the mean state capturing the stabilizer structure preserved under repeated convolution. We establish a finite-step entropy hierarchy generated by compatible stabilizer dephasings. Along every compatible isotropic flag, the entropy increases toward the mean-state entropy ceiling, while the relative-entropy distance to the mean state decomposes exactly into successive coherence losses and a terminal classical nonuniformity. Optimizing over compatible subspaces yields an intrinsic entropy profile of the state. For quantum convolutional channels, Weyl covariance reduces the one-shot classical communication problem to minimal output entropy. A spectral-transfer argument shows that suitable stabilizer inputs reproduce stabilizer-measurement distributions of the environment as channel-output spectra. This gives a computable Holevo lower bound over all complete stabilizer measurements; its compatible restriction is characterized by the entropy hierarchy and refines the previous mean-state bound of Bu, Gu, and Jaffe. The bound is exact for stabilizer-diagonal environments, for which the Holevo capacity is strongly additive, and yields a single-letter formula for a nonstabilizer qutrit family. The same family also exhibits a coexistence region with simultaneously positive classical and quantum communication rates

quant-ph↗

arXiv2026

Private Capacity of Quantum Channels Induced by Non-stabilizer Environmental States

We investigate the private capacity of quantum channels using the recently proposed quantum convolution theory for discrete-variable quantum systems. We focus on the role of the magic resource played in this framework. Firstly, for a large class of convolutional channels, we find that the private capacity is zero if the fixed environmental state is a stabilizer state. Moreover, we show that the private capacity can be nonzero for some magic environmental states. Furthermore, we show that the private capacity of a discrete beam splitter unitary is upper-bounded by the amount of magic of the environmental state. In addition, if the environmental state exhibits a certain symmetric structure, even if it is magic, the corresponding private capacity will also vanish for a class of convolution. These results emphasize the role of magic resources in quantum communication

quant-ph↗

arXiv2026

From Failure to Feedback: Group Revision Unlocks Hard Cases in Object-Level Grounding

Finetuning Large Vision-Language Models with reinforcement learning has emerged as a promising approach to enhance their capability in object-level grounding. However, existing methods, mainly based on GRPO, assign rewards at the response level. Such sparse reward, often criterion-induced, leads to minimal learning signals when all candidate responses fail in challenging scenarios. In this work, we propose a group-revision optimisation paradigm that enhances learning on hard cases. It begins with a sampled initial response and generates a set of revised candidates to explore improved grounding outcomes. Inspired by reward shaping, we introduce a consolidation process that quantifies each candidate's improvement over the initial attempt and converts it into informative shaping signals. These signals are used to both refine the reward and modulate the advantage, amplifying the influence of high-quality revisions. Our method achieves consistent gains across referring and reasoning segmentation, REC, and counting benchmarks compared with prior GRPO-based models. Our code is available at https://github.com/yyliu01/GroupRevision.

cs.CV↗

arXiv2026

On the Simulation Cost of Quantum Finite Automata

This paper identifies exact probabilistic simulation cost as the natural quantitative measure of quantum advantage for finite automata under strict cutpoints. It gives sharp simulation laws for two representative models. A one-way finite automaton with $c$ classical states and a $q$-dimensional quantum register has exact probabilistic simulation cost $\Theta(cq^2)$, while an $n$-dimensional measure-once one-way quantum finite automaton has worst-case cost $\Theta(n^2)$. The proofs develop a prepare--test framework, in which prefixes generate the relevant real operator degrees of freedom and suffixes convert them into strict-cutpoint tests. The same obstruction is recast through finite sign-rank matrices, clarifying the role of Forster's spectral method. Placed beside the surrounding two-way separations, these results give a clean hierarchy of finite-automata quantum advantage.

quant-ph↗

arXiv2026

Matrix encoding method in variational algorithm of calculating eigenvalues and generalized eigenvalues

We propose a variational method for constructing the eigenvalues and generalized eigenvalues for an arbitrary $N\times N$ complex matrix. The quantum part of our algorithm is based on encoding the matrix elements into the pure state of a quantum system and expressing the loss function with optimization parameters in terms of certain probability amplitudes in the superposition state. The principal step of this algorithm is the measurement of the ancilla state that removes all extra terms from the above superposition and allows to probabilistically construct the required loss function along with its derivatives with respect to the optimization parameters. These output data are used to find the new values of optimization parameters for the next iteration of the loss function in the gradient optimization method. The depth and size of the circuit for this algorithm are, respectively, $O(N^2 \log N)$ and $O(\log N)$.

quant-ph↗

arXiv2026

BioMedArena: An Open-source Toolkit for Building and Evaluating Biomedical Deep Research Agents

Reproducing and comparing deep research agents today is hard: the same backbone evaluated on the same benchmark can report different accuracies across papers because the harness and tool registry differ, and integrating a new model into a comparable evaluation surface costs weeks of model-specific engineering. These are symptoms of a broader reproducibility problem in deep research agent research. Here, we introduce BioMedArena, an open-source toolkit that addresses this reproducibility gap and provides an arena for comparing deep research agents under a shared evaluation environment. BioMedArena decouples six layers of biomedical agent evaluation -- benchmark loading, tool exposure, tool selection, harness mode, context management, and scoring -- and exposes 166 biomedical benchmarks and 75 biomedical tools across 9 functional families. Adding a new model, benchmark, or tool can be accomplished with a few-line provider adapter. Beyond evaluation infrastructure, BioMedArena ships a library of high-quality reference components: 6 agent harnesses (including our proposed Mutual-Evolve) and 6 context-management strategies, any of which can be equipped on any backbone. Equipping these components substantially improves all 12 backbones; on each of 8 representative biomedical benchmarks, the best equipped backbone surpasses prior state-of-the-art (SOTA), by 15.01 percentage points on average. The toolkit, configurations, and per-task traces are available at https://github.com/AI-in-Health/BioMedArena.

cs.AI↗

arXiv2026

The State Cost of Classical Simulation of One-Way General Quantum Finite Automata

Under strict cutpoints, probabilistic finite automata (PFAs) and one-way general quantum finite automata (1gQFAs) recognize the same stochastic languages, shifting the theoretical focus to the state cost required for a classical PFA to simulate a 1gQFA. For an $n$-state ($n\geq 2$) 1gQFA, the state cost upper bound of classical simulation was previously known to be $n^2+3$, while the lower bound remained an open problem widely conjectured to be quadratic. After establishing a well-defined notion of classical simulation, we improve the existing state cost upper bound from $n^2+3$ to $n^2+1$. Subsequently, we introduce the concepts of shattering and dynamic shattering, which are used to determine the memory required for a PFA to recognize a language. Using these techniques, we prove that the state cost lower bound of the simulation reaches $n^2+1$ over a four-letter alphabet. With the upper and lower bounds thus matching, the problem is fully resolved.

cs.FL↗

arXiv2026

Exact Separation of Words via Trace Geometry

A basic question in the study of measure-once quantum finite automata is whether two distinct input words can be separated with certainty. The exact separation problem reduces to a trace-vanishing question in $SU(2)$. The main difficulty lies in the genuinely nonabelian regime, where $u$ and $v$ have the same abelianization. This paper develops a slice-driven framework that converts algebraic invariants of the word -- prefix statistics, metabelian polynomials, and slope specializations -- into explicit low-dimensional families in $SU(2)^2$ on which the trace-vanishing question can be analyzed effectively. A quadratic trace-deficit identity on a principal one-parameter family provides the main algebraic-to-geometric bridge. Building on this framework, the paper establishes three core certified slice criteria: a dihedral criterion, equivalently readable through a signed $a$-count; a quaternionic criterion; and a local one-row criterion. Together with a supplementary interior-point test and a binary-dihedral slice, these results sharply reduce the unresolved portion of the problem to a residual super-degenerate class, while also clarifying the limitations of certification strategies based only on finitely many finite-subgroup evaluations.

cs.FL↗

arXiv2026

MedOpenClaw and MedFlowBench: Auditing Medical Agents in Full-Study Workflows

Medical imaging benchmarks often evaluate VLMs on pre-selected 2D images, slices, crops, or patches, making evaluation closer to visual recognition. Real clinical workflows impose a different burden: readers must search through complete studies, operate imaging software, navigate across slices and magnifications, and document visual evidence that can be audited. We argue that this evidence-producing workflow is a critical missing evaluation axis for medical imaging agents. To study it, we introduce MedFlowBench, a full-study benchmark for VLM agents, together with MedOpenClaw, a controlled and replayable runtime in which agents operate medical imaging viewers such as 3D Slicer and QuPath. In each episode, an agent inspects a complete radiology study or whole-slide pathology image, returns a task answer, and submits structured evidence, including key slices, coordinates, regions of interest, or lesion-state fields. This evidence is automatically checked against withheld masks, annotations, and labels. Across evaluated models, final answer-only scoring gives an overly optimistic picture: when answers must also be supported by correct evidence, performance drops substantially on complex workflows. We further find that adding image-analysis tools does not by itself solve the problem. Tools help when they make a complex procedure simple and reliable, but agents still struggle when they must choose inputs, manage viewer state, and verify intermediate outputs over multiple steps. MedFlowBench exposes whether medical imaging agents can produce auditable evidence from complete studies, rather than plausible answers from selected images.

cs.CV↗

arXiv2026

3DMedAgent: Unified Perception-to-Understanding for 3D Medical Analysis

3D CT analysis spans a continuum from low-level perception to high-level clinical understanding. Existing 3D-oriented analysis methods adopt either isolated task-specific modeling or task-agnostic end-to-end paradigms to produce one-hop outputs, impeding the systematic accumulation of perceptual evidence for downstream reasoning. In parallel, recent multimodal large language models (MLLMs) exhibit improved visual perception and can integrate visual and textual information effectively, yet their predominantly 2D-oriented designs fundamentally limit their ability to perceive and analyze volumetric medical data. To bridge this gap, we propose 3DMedAgent, a unified agent that enables 2D MLLMs to perform general 3D CT analysis without 3D-specific fine-tuning. 3DMedAgent coordinates heterogeneous visual and textual tools through a flexible MLLM agent, progressively decomposing complex 3D analysis into tractable subtasks that transition from global to regional views, from 3D volumes to informative 2D slices, and from visual evidence to structured textual representations. Central to this design, 3DMedAgent maintains a long-term structured memory that aggregates intermediate tool outputs and supports query-adaptive, evidence-driven multi-step reasoning. We further introduce the DeepChestVQA benchmark for evaluating unified perception-to-understanding capabilities in 3D thoracic imaging. Experiments across over 40 tasks demonstrate that 3DMedAgent consistently outperforms general, medical, and 3D-specific MLLMs, highlighting a scalable path toward general-purpose 3D clinical assistants.Code and data are available at \href{https://github.com/jinlab-imvr/3DMedAgent}{https://github.com/jinlab-imvr/3DMedAgent}.

cs.CV↗

arXiv2026

Evo: Autoregressive-Diffusion Large Language Models with Evolving Balance

We introduce \textbf{Evo}, a duality latent trajectory model that bridges autoregressive (AR) and diffusion-based language generation within a continuous evolutionary generative framework. Rather than treating AR decoding and diffusion generation as separate paradigms, Evo reconceptualizes text generation as a latent flow: each token is associated with a vector-valued embedding that evolves over a progression variable $t_i \in [0, 1]$, indicating its semantic maturity. Low $t_i$ values correspond to confident AR-like refinement, while high values invoke diffusion-style planning, allowing the model to adaptively balance AR and diffusion based on uncertainty. Theoretically, we show that both AR and diffusion models emerge as discretizations of a shared probability flow, and we derive Evo's training objective from a unified variational ELBO. The model is implemented as a time-conditioned Transformer governed by a shared vector field, trained end-to-end to jointly infer latent codes and their progression times. During decoding, Evo performs efficient, semantics-aware refinement, achieving high-quality outputs without sacrificing speed. Empirically, Evo 8B achieves state-of-the-art or highly competitive results on 15 diverse benchmarks, including reasoning (GSM8K, ARC-C), code generation (HumanEval, MBPP), and general language understanding, while maintaining fast inference speed. Our results demonstrate that Evo delivers a new paradigm for LLM design with strong generation quality, robust symbolic reasoning, and decoding efficiency.

cs.LG↗

arXiv2026

MedVAR: Towards Scalable and Efficient Medical Image Generation via Next-scale Autoregressive Prediction

Medical image generation is pivotal in applications like data augmentation for low-resource clinical tasks and privacy-preserving data sharing. However, developing a scalable generative backbone for medical imaging requires architectural efficiency, sufficient multi-organ data, and principled evaluation, yet current approaches leave these aspects unresolved. Therefore, we introduce MedVAR, the first autoregressive-based foundation model that adopts the next-scale prediction paradigm to enable fast and scale-up-friendly medical image synthesis. MedVAR generates images in a coarse-to-fine manner and produces structured multi-scale representations suitable for downstream use. To support hierarchical generation, we curate a harmonized dataset of around 440,000 CT and MRI images spanning six anatomical regions. Comprehensive experiments across fidelity, diversity, and scalability show that MedVAR achieves state-of-the-art generative performance and offers a promising architectural direction for future medical generative foundation models.

cs.CV↗

arXiv2026

SWE Context Bench: A Benchmark for Context Learning in Coding

Large language models are increasingly used as coding agents for software engineering tasks. Current benchmarks mainly evaluate whether the agent can correctly solve the request or fix the bugs. They largely treat tasks as independent and do not assess whether agents can reuse previous experience across related problems. As a result, the efficiency gains from reusing the previous experience remains difficult to measure. We introduce SWE-ContextBench, a benchmark designed to explicitly evaluate context understanding and retrieval in coding agents. SWE-ContextBench consists of 1,100 base tasks with another 376 related tasks derived from real dependency and reference relationships among GitHub issues and pull requests. SWE-ContextBench groups base tasks and related tasks with shared context across 51 unique repositories and 9 programming languages. The benchmark evaluates how accurately and efficiently agents solve related issues when prior cases are available in context. Using SWE-ContextBench, we study the behavior of multiple coding agents across varying context reuse settings and retrieval strategies. Our results show that accurately summarized and retrieved previous experience can significantly improve resolution accuracy and reduce runtime and token cost, particularly on harder tasks. In contrast, unfiltered or incorrectly selected context provides limited or negative benefits. These findings highlight the importance of context management and retrieval accuracy, and position SWE-ContextBench as a principled benchmark for studying context learning in coding agents.

cs.SE↗

arXiv2025

Analysis of a nonlinear free-boundary tumor model with three layers

In this paper, we study a nonlinear free boundary problem modeling the growth of spherically symmetric tumors. The tumor consists of a central necrotic core, an intermediate annual quiescent-cell layer, and an outer proliferating-cell layer. The evolution of tumor layers and the movement of the tumor boundary are totally governed by external nutrient supply and conservation of mass. The three-layer structure generates three free boundaries with boundary conditions of different types. We develop a nonlinear analysis method to get over the great difficulty arising from free boundaries and the discontinuity of the nutrient-consumption rate function. By carefully studying the mutual relationships between the free boundaries, we reveal the evolutionary mechanism in tumor growth and the mutual transformation of its internal structures. The existence and uniqueness of the radial stationary solution is proved, and its globally asymptotic stability towards different dormant tumor states is established.

math.AP↗

arXiv2025

MedicalOS: An LLM Agent based Operating System for Digital Healthcare

Decades' advances in digital health technologies, such as electronic health records, have largely streamlined routine clinical processes. Yet, most these systems are still hard to learn and use: Clinicians often face the burden of managing multiple tools, repeating manual actions for each patient, navigating complicated UI trees to locate functions, and spending significant time on administration instead of caring for patients. The recent rise of large language model (LLM) based agents demonstrates exceptional capability in coding and computer operation, revealing the potential for humans to interact with operating systems and software not by direct manipulation, but by instructing agents through natural language. This shift highlights the need for an abstraction layer, an agent-computer interface, that translates human language into machine-executable commands. In digital healthcare, however, requires a more domain-specific abstractions that strictly follow trusted clinical guidelines and procedural standards to ensure safety, transparency, and compliance. To address this need, we present \textbf{MedicalOS}, a unified agent-based operational system designed as such a domain-specific abstract layer for healthcare. It translates human instructions into pre-defined digital healthcare commands, such as patient inquiry, history retrieval, exam management, report generation, referrals, treatment planning, that we wrapped as off-the-shelf tools using machine languages (e.g., Python, APIs, MCP, Linux). We empirically validate MedicalOS on 214 patient cases across 22 specialties, demonstrating high diagnostic accuracy and confidence, clinically sound examination requests, and consistent generation of structured reports and medication recommendations. These results highlight MedicalOS as a trustworthy and scalable foundation for advancing workflow automation in clinical practice.

cs.AI↗

arXiv2025

Rational-Valued Affine Verifiers in Arthur--Merlin Proof Systems

Affine automata provide a finite-state computational model that preserves the linear-algebraic structure of quantum computation while operating entirely over the reals. Recent work has shown that affine automata can far surpass classical probabilistic finite-state verifiers. However, prior constructions relied on arbitrary real-valued transition matrices, leaving open whether the observed power stems from the affine mechanism itself or from computational resources implicitly encoded in irrational or infinite-precision parameters. This paper studies one-way and two-way automata with deterministic and affine states as verifiers in Arthur--Merlin proof systems under the restriction that every affine transition matrix has rational entries, and shows that the resulting rational model still supports the main verification advantages of affine finite-state verification. At the one-way level, we verify benchmark nonregular languages that are provably hard or impossible for classical two-way probabilistic verifiers. At the two-way level, we achieve weak verification of every Turing-recognizable language, strong bounded-error verification for every language in $\mathbf{ATIME}(2^{O(n)})$, and perfect-completeness strong verification for every language in $\mathbf{PSPACE}$. These results establish that the remarkable verification power of affine finite-state automata is structural.

cs.FL↗