arXiv subjects

Jian Wang

Publications and source records attributed to Jian Wang.

At least 271 records · Page 15Linked to original sources

arXiv2025

OptPO: Optimal Rollout Allocation for Test-time Policy Optimization

Test-time policy optimization enables large language models (LLMs) to adapt to distribution shifts by leveraging feedback from self-generated rollouts. However, existing methods rely on fixed-budget majority voting to estimate rewards, incurring substantial computational redundancy. We propose Optimal Rollout Allocation for Test-time Policy Optimization (OptPO), a principled framework that adaptively allocates inference budgets. By formulating the voting process as a Bayesian sequential probability ratio test, OptPO dynamically halts sampling once the posterior confidence in a consensus answer exceeds a specified threshold. Crucially, it utilizes the retained rollouts for on-policy updates, seamlessly integrating with algorithms like PPO or GRPO without requiring ground-truth labels. Across diverse reasoning benchmarks, OptPO significantly reduces rollout overhead compared to fixed-sample baselines while preserving or improving accuracy. By unifying statistically optimal stopping with test-time learning, OptPO offers a computationally efficient paradigm for test-time adaptation. The source code will be open upon acceptance at https://open-upon-acceptance.

cs.LG↗

arXiv2025

Fisher-KPP waves and the minimal speed on hexagonal lattice

The hexagonal structure is ubiquitous in nature. The propagation phenomena occurring in a media with a hexagonal structure remain to be explored. One way of exploring this question is to formulate lattice dynamical systems and analyze the propagation dynamics. In this paper, we propose a lattice differential equation model featuring a discrete diffusion operator with the hexagonal structure, and a monostable nonlinear term known as the Fisher-KPP mechanism in modeling population growth. A rigorous analysis is conducted on the traveling waves, thoroughly establishing the existence and uniqueness (up to translation) of the traveling waves. Moreover, the periodicity and monotonicity of the minimal wave speed concerning an angle are demonstrated, which is different from the existing results of the minimal wave speed in $\mathbb{R}^2$ and $\mathbb{Z}^2$. Numerical results validate theoretical analysis and further suggest that the minimal wave speed is also the spreading speed of solutions with compactly supported initial values.

math.DS↗

arXiv2025

Hierarchical Feature Integration for Multi-Signal Automatic Modulation Recognition

Automatic modulation recognition (AMR) is a crucial step in wireless communication systems, which identifies the modulation scheme from detected signals to provide key information for further processing. However, previous work has mainly focused on the identification of a single signal, overlooking the phenomenon of multiple signal superposition in practical channels and the signal detection procedures that must be conducted beforehand. Considering the susceptibility of radio frequency (RF) signals to noise interference and significant spectral variations, we propose a novel Hierarchical Feature Integration (HIFI)-YOLO framework for multi-signal joint detection and modulation recognition. Our HIFI-YOLO framework, with its unique design of hierarchical feature integration, effectively enhances the representation capability of features in different modules, thereby improving detection performance. We construct a large-scale AMR dataset specifically tailored for scenarios of the coexistence or overlapping of multiple signals transmitted through channels with realistic propagation conditions, consisting of diverse digital and analog modulation schemes. Extensive experiments on our dataset demonstrate the excellent performance of HIFI-YOLO in multi-signal detection and modulation recognition as a joint approach.

eess.SP↗

arXiv2025

Quantitative homogenization on time-dependent random conductance models with stable-like jumps

We establish quantitative homogenization results for time-dependent random conductance models with stable-like long range jumps on $\Z^d$, where the transition probability from $x$ to $y$ is given by $w_{t, x,y}|x-y|^{-d-α}$ with $α\in (0,2)$. In particular, time-dependent random coefficients $\{w_{t,x,y}: t\in \R_+, (x,y)\in E\}$ are uniformly bounded from above (but may be degenerate), and satisfy the Kolmogorov continuous condition, where $E=\{(x, y): x \not= y \in \Z^d\}$ is the set of all unordered pairs on $\Z^d$. The proofs are based on $L^2$-estimates and energy estimates for solutions to regionalparabolic equations and multi-scale Poincaré inequalities associated with time-dependent symmetric stable-like random walks with random coefficients.

math.PR↗

arXiv2025

EHR-R1: A Reasoning-Enhanced Foundational Language Model for Electronic Health Record Analysis

Electronic Health Records (EHRs) contain rich yet complex information, and their automated analysis is critical for clinical decision-making. Despite recent advances of large language models (LLMs) in clinical workflows, their ability to analyze EHRs remains limited due to narrow task coverage and lack of EHR-oriented reasoning capabilities. This paper aims to bridge the gap, specifically, we present EHR-Ins, a large-scale, comprehensive EHR reasoning instruction dataset, comprising 300k high-quality reasoning cases and 4M non-reasoning cases across 42 distinct EHR tasks. Its core innovation is a thinking-graph-driven framework that enables to generate high-quality reasoning data at scale. Based on it, we develop EHR-R1, a series of reasoning-enhanced LLMs with up to 72B parameters tailored for EHR analysis. Through a multi-stage training paradigm, including domain adaptation, reasoning enhancement, and reinforcement learning, EHR-R1 systematically acquires domain knowledge and diverse reasoning capabilities, enabling accurate and robust EHR analysis. Lastly, we introduce EHR-Bench, a new benchmark curated from MIMIC-IV, spanning 42 tasks, to comprehensively assess reasoning and prediction across EHR scenarios. In experiments, we show that the resulting EHR-R1 consistently outperforms state-of-the-art commercial and open-source LLMs (including DeepSeek-V3 and GPT-4o), surpassing GPT-4o by over 30 points on MIMIC-Bench and achieving a 10\% higher zero-shot AUROC on EHRSHOT. Collectively, EHR-Ins, EHR-R1, and EHR-Bench have significantly advanced the development for more reliable and clinically relevant EHR analysis.

cs.CL↗

arXiv2025

Enhancing Automated Program Repair via Faulty Token Localization and Quality-Aware Patch Refinement

Large language models (LLMs) have recently demonstrated strong potential for automated program repair (APR). However, existing LLM-based techniques primarily rely on coarse-grained external feedback (e.g.,test results) to guide iterative patch generation, while lacking fine-grained internal signals that reveal why a patch fails or which parts of the generated code are likely incorrect. This limitation often leads to inefficient refinement, error propagation, and suboptimal repair performance. In this work, we propose TokenRepair, a novel two-level refinement framework that enhances APR by integrating internal reflection for localizing potentially faulty tokens with external feedback for quality-aware patch refinement. Specifically, TokenRepair first performs internal reflection by analyzing context-aware token-level uncertainty fluctuations to identify suspicious or low-confidence tokens within a patch. It then applies Chain-of-Thought guided rewriting to refine only these localized tokens, enabling targeted and fine-grained correction. To further stabilize the iterative repair loop, TokenRepair incorporates a quality-aware external feedback mechanism that evaluates patch quality and filters out low-quality candidates before refinement. Experimental results show that TokenRepair achieves new state-of-the-art repair performance, correctly fixing 88 bugs on Defects4J 1.2 and 139 bugs on HumanEval-Java, demonstrating substantial improvements ranging from 8.2% to 34.9% across all models on Defects4J 1.2 and from 3.3% to 16.1% on HumanEval-Java.

cs.SE↗

arXiv2025

Bell Inequality Violation with Vacuum-One-Photon Number Superposition States

Entanglement is a central resource in quantum technologies, and the realization of photonic entanglement necessarily relies on interaction with matter. Resonance fluorescence (RF), originating from the coherent interaction between a driving field and a two-level system, plays a pivotal role in quantum optics. Here, we demonstrate a novel route to entanglement generation based on RF from a single quantum dot. Rather than relying on generation of multiphoton states, our approach directly exploits vacuum-one-photon number superposition states created under resonant excitation. By delocalizing this superposition via a beam splitter, we realize time-bin entanglement and observe a clear violation of the Clauser-Horn-Shimony-Holt Bell inequality using Franson-type interferometry. Our scheme removes the need for multiphoton generation, simplifies the experimental requirements, and establishes a scalable pathway toward solid-state entangled photon sources.

quant-ph↗

arXiv2025

On the largest degrees in intersecting hypergraphs

Let $\binom{[n]}{k}$ denote the collection of all $k$-subsets of the standard $n$-set $[n]=\{1,2,\ldots,n\}$. Let $n>2k$ and let $\mathcal{F}\subset \binom{[n]}{k}$ be an {\it intersecting} $k$-graph, i.e., $F\cap F'\neq \emptyset$ for all $F,F'\in \mathcal{F}$. The number of edges $F\in \mathcal{F}$ containing $x\in [n]$ is called the {\it degree} of $x$. Assume that $d_1\geq d_2\geq \ldots\geq d_n$ are the degrees of $\mathcal{F}$ in decreasing order. An important result of Huang and Zhao states that for $n>2k$ the minimum degree $d_n$ is at most $\binom{n-2}{k-2}$. For $n\geq 6k-9$ we strengthen this result by showing $d_{2k+1}\leq \binom{n-2}{k-2}$. As to the second and third largest degrees we prove the best possible bound $d_3\leq d_2\leq \binom{n-2}{k-2}+\binom{n-3}{k-2}$ for $n>2k$. Several more best possible results of a similar nature are established.

math.CO↗

arXiv2025

Rényi Differential Privacy for Heavy-Tailed SDEs via Fractional Poincaré Inequalities

Characterizing the differential privacy (DP) of learning algorithms has become a major challenge in recent years. In parallel, many studies suggested investigating the behavior of stochastic gradient descent (SGD) with heavy-tailed noise, both as a model for modern deep learning models and to improve their performance. However, most DP bounds focus on light-tailed noise, where satisfactory guarantees have been obtained but the proposed techniques do not directly extend to the heavy-tailed setting. Recently, the first DP guarantees for heavy-tailed SGD were obtained. These results provide $(0,δ)$-DP guarantees without requiring gradient clipping. Despite casting new light on the link between DP and heavy-tailed algorithms, these results have a strong dependence on the number of parameters and cannot be extended to other DP notions like the well-established Rényi differential privacy (RDP). In this work, we propose to address these limitations by deriving the first RDP guarantees for heavy-tailed SDEs, as well as their discretized counterparts. Our framework is based on new Rényi flow computations and the use of well-established fractional Poincaré inequalities. Under the assumption that such inequalities are satisfied, we obtain DP guarantees that have a much weaker dependence on the dimension compared to prior art.

stat.ML↗

arXiv2025

Multiwavelength Observations of the Apparently Non-repeating FRB 20250316A

The physical origin of fast radio bursts (FRBs) remains uncertain. Although multiwavelength observations have been widely conducted, only Galactic FRB~20200428D is associated with an X-ray burst from the magnetar SGR J1935+2154. Here, we present multiwavelength follow-up observations of the nearby bright FRB~20250316A, including the Five-hundred-meter Aperture Spherical radio Telescope (FAST), Einstein Probe (EP) X-ray mission, Chandra X-ray Observatory, Wide Field Survey Telescope (WFST) and Space Variable Object Monitor/Visible Telescope (SVOM/VT). The 13.08-hour FAST follow-up campaign without pulse detection requires an energy distribution flatter than those of well-known repeating FRBs, suggesting that this burst is likely a one-off event. A prompt EP follow-up and multi-epoch observational campaign totaling $>$ 100 ks led to the detection of an X-ray source within the angular resolution of its Follow-up X-ray Telescope (FXT, $10^{\prime\prime}$). A subsequent Chandra observation revealed this source to be offset by $7^{\prime\prime}$ from the FRB position, and established a 0.5-10 keV flux upper limit of $7.6\times 10^{-15}$ $\rm erg\,cm^{-2}\,s^{-1}$ at the FRB position, corresponding to $\sim 10^{39}$ $\rm erg\,s^{-1}$ at the 40 Mpc distance of the host galaxy NGC~4141. These results set one of the most stringent limits on X-ray emission from a non-repeating FRB, disfavoring ultra-luminous X-ray sources (ULXs) as counterparts of apparently one-off FRBs and offering critical insights into afterglow models. Our study suggests that an arcsecond localization of both the FRB and its potential X-ray counterpart is essential for exploring the X-ray counterpart of an FRB.

astro-ph.HE↗

arXiv2025

Efficient Reinforcement Learning for Zero-Shot Coordination in Evolving Games

Zero-shot coordination(ZSC), a key challenge in multi-agent game theory, has become a hot topic in reinforcement learning (RL) research recently, especially in complex evolving games. It focuses on the generalization ability of agents, requiring them to coordinate well with collaborators from a diverse, potentially evolving, pool of partners that are not seen before without any fine-tuning. Population-based training, which approximates such an evolving partner pool, has been proven to provide good zero-shot coordination performance; nevertheless, existing methods are limited by computational resources, mainly focusing on optimizing diversity in small populations while neglecting the potential performance gains from scaling population size. To address this issue, this paper proposes the Scalable Population Training (ScaPT), an efficient RL training framework comprising two key components: a meta-agent that efficiently realizes a population by selectively sharing parameters across agents, and a mutual information regularizer that guarantees population diversity. To empirically validate the effectiveness of ScaPT, this paper evaluates it along with representational frameworks in Hanabi cooperative game and confirms its superiority.

cs.LG↗

arXiv2025

Root Cause Analysis for Microservice Systems via Cascaded Conditional Learning with Hypergraphs

Root cause analysis in microservice systems typically involves two core tasks: root cause localization (RCL) and failure type identification (FTI). Despite substantial research efforts, conventional diagnostic approaches still face two key challenges. First, these methods predominantly adopt a joint learning paradigm for RCL and FTI to exploit shared information and reduce training time. However, this simplistic integration neglects the causal dependencies between tasks, thereby impeding inter-task collaboration and information transfer. Second, these existing methods primarily focus on point-to-point relationships between instances, overlooking the group nature of inter-instance influences induced by deployment configurations and load balancing. To overcome these limitations, we propose CCLH, a novel root cause analysis framework that orchestrates diagnostic tasks based on cascaded conditional learning. CCLH provides a three-level taxonomy for group influences between instances and incorporates a heterogeneous hypergraph to model these relationships, facilitating the simulation of failure propagation. Extensive experiments conducted on datasets from three microservice benchmarks demonstrate that CCLH outperforms state-of-the-art methods in both RCL and FTI.

cs.LG↗

arXiv2025

Invisible Triggers, Visible Threats! Road-Style Adversarial Creation Attack for Visual 3D Detection in Autonomous Driving

Modern autonomous driving (AD) systems leverage 3D object detection to perceive foreground objects in 3D environments for subsequent prediction and planning. Visual 3D detection based on RGB cameras provides a cost-effective solution compared to the LiDAR paradigm. While achieving promising detection accuracy, current deep neural network-based models remain highly susceptible to adversarial examples. The underlying safety concerns motivate us to investigate realistic adversarial attacks in AD scenarios. Previous work has demonstrated the feasibility of placing adversarial posters on the road surface to induce hallucinations in the detector. However, the unnatural appearance of the posters makes them easily noticeable by humans, and their fixed content can be readily targeted and defended. To address these limitations, we propose the AdvRoad to generate diverse road-style adversarial posters. The adversaries have naturalistic appearances resembling the road surface while compromising the detector to perceive non-existent objects at the attack locations. We employ a two-stage approach, termed Road-Style Adversary Generation and Scenario-Associated Adaptation, to maximize the attack effectiveness on the input scene while ensuring the natural appearance of the poster, allowing the attack to be carried out stealthily without drawing human attention. Extensive experiments show that AdvRoad generalizes well to different detectors, scenes, and spoofing locations. Moreover, physical attacks further demonstrate the practical threats in real-world environments.

cs.CV↗

arXiv2025

Compton Scattering Total Cross Section at Next-to-Next-to-Leading Order and Resummation of Leading Logarithms

Compton scattering is a fundamental process in QED with broad applications, yet its theoretical description at high energies is challenged by substantial next-to-leading order (NLO) corrections arising from double-logarithmic enhancements. To address this, we report the first calculation of the next-to-next-to-leading order (NNLO) total cross section with full electron mass dependence. Our analysis reveals that the NNLO correction, albeit still containing double logarithms, is numerically small due to a suppressing prefactor. By identifying the origin of these logarithms in a kinematic regime featuring a Glauber electron exchange, we successfully resum the leading logarithmic series to all orders, obtaining a compact result in terms of a modified Bessel function. The all-order structure reveals a suppression mechanism, with double factorial terms in the denominator, which explains the negligible nature of higher-order contributions. The combination of our NNLO calculation and all-orders resummation delivers a reliable and precise prediction, poised to serve the needs of high-precision experiments in the foreseeable future.

hep-ph↗

arXiv2025

Overview of CHIP 2025 Shared Task 2: Discharge Medication Recommendation for Metabolic Diseases Based on Chinese Electronic Health Records

Discharge medication recommendation plays a critical role in ensuring treatment continuity, preventing readmission, and improving long-term management for patients with chronic metabolic diseases. This paper present an overview of the CHIP 2025 Shared Task 2 competition, which aimed to develop state-of-the-art approaches for automatically recommending appro-priate discharge medications using real-world Chinese EHR data. For this task, we constructed CDrugRed, a high-quality dataset consisting of 5,894 de-identified hospitalization records from 3,190 patients in China. This task is challenging due to multi-label nature of medication recommendation, het-erogeneous clinical text, and patient-specific variability in treatment plans. A total of 526 teams registered, with 167 and 95 teams submitting valid results to the Phase A and Phase B leaderboards, respectively. The top-performing team achieved the highest overall performance on the final test set, with a Jaccard score of 0.5102, F1 score of 0.6267, demonstrating the potential of advanced large language model (LLM)-based ensemble systems. These re-sults highlight both the promise and remaining challenges of applying LLMs to medication recommendation in Chinese EHRs. The post-evaluation phase remains open at https://tianchi.aliyun.com/competition/entrance/532411/.

cs.CL↗

arXiv2025

Learning Fourier shapes to probe the geometric world of deep neural networks

While both shape and texture are fundamental to visual recognition, research on deep neural networks (DNNs) has predominantly focused on the latter, leaving their geometric understanding poorly probed. Here, we show: first, that optimized shapes can act as potent semantic carriers, generating high-confidence classifications from inputs defined purely by their geometry; second, that they are high-fidelity interpretability tools that precisely isolate a model's salient regions; and third, that they constitute a new, generalizable adversarial paradigm capable of deceiving downstream visual tasks. This is achieved through an end-to-end differentiable framework that unifies a powerful Fourier series to parameterize arbitrary shapes, a winding number-based mapping to translate them into the pixel grid required by DNNs, and signal energy constraints that enhance optimization efficiency while ensuring physically plausible shapes. Our work provides a versatile framework for probing the geometric world of DNNs and opens new frontiers for challenging and understanding machine perception.

cs.CV↗

arXiv2025

Two-photon interference between mutually-detuned resonance fluorescence signals scattered off a semiconductor quantum dot

The radiative linewidth of a two-level emitter (TLE) fundamentally limits the bandwidth available for quantum information processing. Despite its importance, no prior experiment has systematically examined how driving detuning affects the indistinguishability of photons scattered from a TLE - a parameter critical for photonic quantum computing. Here, we perform post-selective two-photon interference measurements between mutually detuned resonance fluorescence signals from an InAs quantum dot embedded in a micropillar cavity. At small mutual laser detunings (<=0.5GHz), the results are accurately described by the pure-state model [Nat. Commun. 16, 6453 (2025)], which treats all resonance-fluorescence photons as spontaneous emission. At larger detunings, we uncover an anomalous feature in the two-photon interference, where the normalised second-order correlation function under orthogonal polarisations yields g2_vert(0) < 0.5.

quant-ph↗

arXiv2025

Enhancing Diffusion-based Restoration Models via Difficulty-Adaptive Reinforcement Learning with IQA Reward

Reinforcement Learning (RL) has recently been incorporated into diffusion models, e.g., tasks such as text-to-image. However, directly applying existing RL methods to diffusion-based image restoration models is suboptimal, as the objective of restoration fundamentally differs from that of pure generation: it places greater emphasis on fidelity. In this paper, we investigate how to effectively integrate RL into diffusion-based restoration models. First, through extensive experiments with various reward functions, we find that an effective reward can be derived from an Image Quality Assessment (IQA) model, instead of intuitive ground-truth-based supervision, which has already been optimized during the Supervised Fine-Tuning (SFT) stage prior to RL. Moreover, our strategy focuses on using RL for challenging samples that are significantly distant from the ground truth, and our RL approach is innovatively implemented using MLLM-based IQA models to align distributions with high-quality images initially. As the samples approach the ground truth's distribution, RL is adaptively combined with SFT for more fine-grained alignment. This dynamic process is facilitated through an automatic weighting strategy that adjusts based on the relative difficulty of the training samples. Our strategy is plug-and-play that can be seamlessly applied to diffusion-based restoration models, boosting its performance across various restoration tasks. Extensive experiments across multiple benchmarks demonstrate the effectiveness of our proposed RL framework.

cs.CV↗