arXiv subjects

Ling Xu

Publications and source records attributed to Ling Xu.

At least 19 recordsLinked to original sources

arXiv2026

Multi-Perspective Agentic Program Repair via Code Property Graphs and Temporal Execution Graphs

Large language models (LLMs) have improved automated program repair (APR), but two limitations remain. First, raw execution traces are often too large and repetitive to serve as effective model context. Second, repeated patch sampling may produce different implementations without yielding distinct root-cause hypotheses or repair strategies. We present CT-Repair, an agentic APR framework representing static and dynamic evidence as queryable Code Property Graph (CPG) and Temporal Execution Graph (TEG). CT-Repair applies a three-stage filtering pipeline to construct compact TEGs. Three finite-state-machine-guided agents analyze each bug from static, dynamic, and hybrid perspectives and independently produce evidence-grounded repair strategies. A strategy-guided generation procedure instantiates these strategies as candidate patches and uses validation feedback to refine the most promising strategy. We evaluate CT-Repair on 854 Java bugs from Defects4J v3.0. In the mixed-model configuration, CT-Repair correctly repairs 489 bugs. Under a controlled GPT-5.4-mini configuration, it repairs 388 bugs, 19 and 30 more than ReinFix and RepairAgent, respectively. The union of the three evidence perspectives repairs 99 more bugs than the strongest individual perspective. The filtering pipeline also compacts runtime evidence, with execution filtering narrowing the candidate method scope by 94.85% on average and behavior filtering further reducing retained runtime records by 55.97%. These results show that structured runtime evidence and multi-perspective reasoning can improve repair effectiveness without relying solely on a larger patch-generation budget.

cs.SE↗

arXiv2026

Embodied.cpp: A Portable Inference Runtime of Embodied AI Models on Heterogeneous Robots

Embodied AI models now span vision-language-action (VLA) models and world-action models (WAMs), but practical deployment remains fragmented across model-specific Python stacks, backend assumptions, and robot-side glue code, especially on heterogeneous edge devices. Existing inference runtimes are designed mainly for request-response serving and therefore do not satisfy the runtime contract of embodied deployment: multi-rate execution inside closed-loop control, latency-first batch-1 inference on heterogeneous hardware, and extensible embodied interfaces beyond fixed token I/O. We present Embodied$.$cpp, a portable C++ inference runtime for embodied models. Based on an architectural analysis of representative VLA models and WAMs, Embodied$.$cpp captures a shared execution path and organizes it into five layers: input adapters, sequence builders, backbone execution, head plugins, and deployment adapters. The runtime provides modular multi-rate execution, latency-first fused inference, and extensible operator and I/O support, enabling deployment across heterogeneous devices, robots, and simulators through one backend abstraction. We evaluate Embodied$.$cpp on three VLA and two WAM models, using normalized comparisons across Python and C++ quantization configurations. Overall, Embodied$.$cpp achieves 1.05x-2.70x inference speedups and 7\%-77\% lower VRAM relative to Python baselines, while maintaining near-baseline success for most configurations. These results show that Embodied$.$cpp improves deployment efficiency while preserving high control quality across diverse embodied model architectures. Project Link: https://github.com/SEU-PAISys/Embodied.cpp

cs.RO↗

arXiv2026

An automated method of identifying incorrectly labelled images based on the sequences of loss functions of deep learning networks

Deep learning is widely applied in medical image analysis, but up to 10% of manually labelled images may be incorrect, degrading model performance. This paper proposes an automated method to identify incorrectly labelled medical images by analyzing sequences of loss functions from deep learning classification networks over multiple training epochs. Identified images can be reviewed and relabelled by experts, improving dataset quality and model performance. Two experiments validate the method on a fundus image dataset for referable diabetic retinopathy screening. In the first, 6% (648) of 10,788 gold-standard labels were intentionally flipped. The method identified 75.31% (488) of the flipped samples, with only 4.85% (492) false positives among correctly labelled samples. In the second, reviewing and correcting the 980 identified samples (9.1% of the dataset) and retraining the model improved best accuracy on an independent test set from 95.93% (with 6% label noise) to 96.50% (with 1.5% noise), approaching the ideal 96.57% (with 0% noise). The results demonstrate the method's effectiveness in improving model performance through automated label quality control.

cs.CV↗

arXiv2025

DynaFix: Iterative Automated Program Repair Driven by Execution-Level Dynamic Information

Automated Program Repair (APR) aims to automatically generate correct patches for buggy programs. Recent approaches leveraging large language models (LLMs) have shown promise but face limitations. Most rely solely on static analysis, ignoring runtime behaviors. Some attempt to incorporate dynamic signals, but these are often restricted to training or fine-tuning, or injected only once into the repair prompt, without iterative use. This fails to fully capture program execution. Current iterative repair frameworks typically rely on coarse-grained feedback, such as pass/fail results or exception types, and do not leverage fine-grained execution-level information effectively. As a result, models struggle to simulate human stepwise debugging, limiting their effectiveness in multi-step reasoning and complex bug repair. To address these challenges, we propose DynaFix, an execution-level dynamic information-driven APR method that iteratively leverages runtime information to refine the repair process. In each repair round, DynaFix captures execution-level dynamic information such as variable states, control-flow paths, and call stacks, transforming them into structured prompts to guide LLMs in generating candidate patches. If a patch fails validation, DynaFix re-executes the modified program to collect new execution information for the next attempt. This iterative loop incrementally improves patches based on updated feedback, similar to the stepwise debugging practices of human developers. We evaluate DynaFix on the Defects4J v1.2 and v2.0 benchmarks. DynaFix repairs 186 single-function bugs, a 10% improvement over state-of-the-art baselines, including 38 bugs previously unrepaired. It achieves correct patches within at most 35 attempts, reducing the patch search space by 70% compared with existing methods, thereby demonstrating both effectiveness and efficiency in repairing complex bugs.

cs.SE↗

arXiv2025

Entanglement-driven responses through multiscale 3D-printed knits

For their resilience and toughness, filamentous entanglements are ubiquitous in both natural and engineered systems across length scales, from polymer-chain- to collagen-networks and from cable-net structures to forest canopies. Textiles are an everyday manifestation of filamentous entanglement: the remarkable resilience and toughness in knitted fabrics arise predominately from the topology of interlooped yarns. Yet most architected materials do not exploit entanglement as a design primitive, and industrial knitting fixes a narrow set of patterns for manufacturability. Additive manufacturing has recently enabled interlocking structures such as chainmail, knot and woven assemblies, hinting at broader possibilities for entangled architectures. The general challenge is to treat knitting itself as a three-dimensional architected material with predictable and tunable mechanics across scales. Here, we show that knitted architectures fabricated additively can be recast as periodic entangled solids whose responses are both fabric-like and programmable. We reproduce the characteristic behavior of conventional planar knits and extend knitting into the third dimension by interlooping along three orthogonal directions, yielding volumetric knits whose stiffness and dissipation are tuned by prescribed pre-strain. We propose a simple scaling that unifies the responses across stitch geometries and constituent materials. Further, we realize the same topology from centimeter to micrometer scales, culminating in the fabrication of what is, to our knowledge, the smallest knitted structure ever made. By demonstrating 3D-printed knits can be interpreted both as a traditional fabric, as well as a novel architected material with defined periodicity, this work establishes the dual nature of entangled filaments and paves the way towards a new form of material architectures with high degrees of entanglement.

physics.app-ph↗

arXiv2025

Evaluate-and-Purify: Fortifying Code Language Models Against Adversarial Attacks Using LLM-as-a-Judge

The widespread adoption of code language models in software engineering tasks has exposed vulnerabilities to adversarial attacks, especially the identifier substitution attacks. Although existing identifier substitution attackers demonstrate high success rates, they often produce adversarial examples with unnatural code patterns. In this paper, we systematically assess the quality of adversarial examples using LLM-as-a-Judge. Our analysis reveals that over 80% of adversarial examples generated by state-of-the-art identifier substitution attackers (e.g., ALERT) are actually detectable. Based on this insight, we propose EP-Shield, a unified framework for evaluating and purifying identifier substitution attacks via naturalness-aware reasoning. Specifically, we first evaluate the naturalness of code and identify the perturbed adversarial code, then purify it so that the victim model can restore correct prediction. Extensive experiments demonstrate the superiority of EP-Shield over adversarial fine-tuning (up to 83.36% improvement) and its lightweight design 7B parameters) with GPT-4-level performance.

cs.SE↗

arXiv2025

Corrected Trapezoidal Rules for Near-Singular Surface Integrals Applied to 3D Interfacial Stokes Flow

Interfacial Stokes flow can be efficiently computed using the Boundary Integral Equation method. In 3D, the fluid velocity at a target point is given by a 2D surface integral over all interfaces, thus reducing the dimension of the problem. A core challenge is that for target points near, but not on, an interface, the surface integral is near-singular and standard quadratures lose accuracy. This paper presents a method to accurately compute the near-singular integrals arising in elliptic boundary value problems in 3D. It is based on a local series approximation of the integrand about a base point on the surface, obtained by orthogonal projection of the target point onto the surface. The elementary functions in the resulting series approximation can be integrated to high accuracy in a neighborhood of the base point using a recursive algorithm. The remaining integral is evaluated numerically using a standard quadrature rule, chosen here to be the 4th order Trapezoidal rule. The method is reduced to the standard quadrature plus a correction, and is uniformly of 4th order. The method is applied to resolve Stokes flow past several ellipsoidal rigid bodies. We compare the error in the velocity near the bodies, and in the time and displacement of particles traveling around the bodies, computed with and without the corrections.

math.NA↗

arXiv2025

Reparametrization of 3D CSC Dubins Paths Enabling 2D Search

This paper addresses the Dubins path planning problem for vehicles in 3D space. In particular, we consider the problem of computing CSC paths -- paths that consist of a circular arc (C) followed by a straight segment (S) followed by a circular arc (C). These paths are useful for vehicles such as fixed-wing aircraft and underwater submersibles that are subject to lower bounds on turn radius. We present a new parameterization that reduces the 3D CSC planning problem to a search over 2 variables, thus lowering search complexity, while also providing gradients that assist that search. We use these equations with a numerical solver to explore numbers and types of solutions computed for a variety of planar and 3D scenarios. Our method successfully computes CSC paths for the large majority of test cases, indicating that it could be useful for future generation of robust, efficient curvature-constrained trajectories.

cs.RO↗

arXiv2025

Numerical study of vortices within a background vortex

This study aims to exploit the analogy of vortex dynamics in a 2D ideal fluid and 2D non-neutral plasma. Numerical simulations using contour dynamics with adaptive refinement are conducted to study the dynamics of one or more vortices within a background vortex, and results are compared with experiments using magnetized non-neutral electron plasmas confined in a Penning-Malmberg trap. This comparison is based on the isomorphism between the 2D Euler equations for ideal fluid and the 2D drift-Poisson equations for non-neutral plasma, where the the fluid vorticity and stream function correspond to the plasma charge density and electric potential. Agreement and discrepancy between simulation and experiment are observed and discussed.

physics.flu-dyn↗

arXiv2024

Enhancing Visual Representation for Text-based Person Searching

Text-based person search aims to retrieve the matched pedestrians from a large-scale image database according to the text description. The core difficulty of this task is how to extract effective details from pedestrian images and texts, and achieve cross-modal alignment in a common latent space. Prior works adopt image and text encoders pre-trained on unimodal data to extract global and local features from image and text respectively, and then global-local alignment is achieved explicitly. However, these approaches still lack the ability of understanding visual details, and the retrieval accuracy is still limited by identity confusion. In order to alleviate the above problems, we rethink the importance of visual features for text-based person search, and propose VFE-TPS, a Visual Feature Enhanced Text-based Person Search model. It introduces a pre-trained multimodal backbone CLIP to learn basic multimodal features and constructs Text Guided Masked Image Modeling task to enhance the model's ability of learning local visual details without explicit annotation. In addition, we design Identity Supervised Global Visual Feature Calibration task to guide the model learn identity-aware global visual features. The key finding of our study is that, with the help of our proposed auxiliary tasks, the knowledge embedded in the pre-trained CLIP model can be successfully adapted to text-based person search task, and the model's visual understanding ability is significantly enhanced. Experimental results on three benchmarks demonstrate that our proposed model exceeds the existing approaches, and the Rank-1 accuracy is significantly improved with a notable margin of about $1\%\sim9\%$. Our code can be found at https://github.com/zhangweifeng1218/VFE_TPS.

cs.CV↗

arXiv2024

Seed-Music: A Unified Framework for High Quality and Controlled Music Generation

We introduce Seed-Music, a suite of music generation systems capable of producing high-quality music with fine-grained style control. Our unified framework leverages both auto-regressive language modeling and diffusion approaches to support two key music creation workflows: controlled music generation and post-production editing. For controlled music generation, our system enables vocal music generation with performance controls from multi-modal inputs, including style descriptions, audio references, musical scores, and voice prompts. For post-production editing, it offers interactive tools for editing lyrics and vocal melodies directly in the generated audio. We encourage readers to listen to demo audio examples at https://team.doubao.com/seed-music "https://team.doubao.com/seed-music".

cs.SD↗

arXiv2024

A dataset of primary nasopharyngeal carcinoma MRI with multi-modalities segmentation

Multi-modality magnetic resonance imaging(MRI) data facilitate the early diagnosis, tumor segmentation, and disease staging in the management of nasopharyngeal carcinoma (NPC). The lack of publicly available, comprehensive datasets limits advancements in diagnosis, treatment planning, and the development of machine learning algorithms for NPC. Addressing this critical need, we introduce the first comprehensive NPC MRI dataset, encompassing MR axial imaging of 277 primary NPC patients. This dataset includes T1-weighted, T2-weighted, and contrast-enhanced T1-weighted sequences, totaling 831 scans. In addition to the corresponding clinical data, manually annotated and labeled segmentations by experienced radiologists offer high-quality data resources from untreated primary NPC.

eess.IV↗

arXiv2024

Sensing environmental physical interaction to traverse cluttered obstacles

The long-standing, dominant approach to robotic obstacle negotiation relies on mapping environmental geometry to avoid obstacles. However, this approach does not allow for traversal of cluttered obstacles, hindering applications such as search and rescue operations through earthquake rubble and exploration across lunar and Martian rocks. To overcome this challenge, robots must further sense and utilize environmental physical interactions to control themselves to traverse obstacles. Recently, a physics-based approach has been established towards this vision. Self-propelled robots interacting with obstacles results in a potential energy landscape. On this landscape, to traverse obstacles, a robot must escape from certain landscape basins that attract it into failure modes, to reach other basins that lead to successful modes. Thus, sensing the potential energy landscape is crucial. Here, we developed new methods and performed systematic experiments to demonstrate that the potential energy landscape can be estimated by sensing environmental physical interaction. We developed a minimalistic robot capable of sensing obstacle contact forces and torques for systematic experiments over a wide range of parameter space. Surprisingly, although these forces and torques are not fully conservative, they match the potential energy landscape gradients that are conservative forces and torques, enabling an accurate estimation of the potential energy landscape. Additionally, a bio-inspired strategy further enhanced estimation accuracy. Our results provided a foundation for further refining these methods for use in free-locomoting robots. Our study is a key step in establishing a new physics-based approach for robots to traverse clustered obstacles to advance their mobility in complex, real-world environments.

cs.RO↗

arXiv2023

Aggregation Phenomena in Lyotropic Chromonic Liquid Crystals

We study the aggregation phenomenon in lyotropic chromonic liquid crystals as the molecular concentration changes and condensing agents are added into the system. Using properties of the critical points of the Oseen-Frank energy of a nematic liquid crystal, combined with the geometric constraints of the hexagonal columnar chromonic phases, we show that the minimizers of the total energy are topologically equivalent to tori, in agreement with available experimental evidence on chromonic liquid crystals and DNA condensates, in viral capsids as well as in free solution. We model the system as bi-phasic, consisting of liquid crystal molecules and water, and postulate the total energy as the sum of the Flory-Huggins energy of mixing together with the bending and surface tension contributions of the liquid crystal. Two types of problems are considered, one related to finding the optimal shape of a torus, once the phase separation has occurred, and the second one that models the conditions leading to molecular aggregation. This work follows recent experimental investigations, but without addressing the topological properties of the toroidal nuclei observed and focusing on how the liquid crystal order competes with the aggregation phenomenon.

cond-mat.soft↗

arXiv2021

Meromorphic solutions of delay differential equations related to logistic type and generalizations

Let $\{b_{j}\}_{j=1}^{k}$ be meromorphic functions, and let $w$ be admissible meromorphic solutions of delay differential equation $$w'(z)=w(z)\left[\frac{P(z, w(z))}{Q(z,w(z))}+\sum_{j=1}^{k}b_{j}(z)w(z-c_{j})\right]$$ with distinct delays $c_{1}, \ldots, c_{k}\in\mathbb{C}\setminus\{0\},$ where the two nonzero polynomials $P(z, w(z))$ and $Q(z, w(z))$ in $w$ with meromorphic coefficients are prime each other. We obtain that if $\limsup_{r\rightarrow\infty}\frac{\log T(r, w)}{r}=0,$ then $$deg_{w}(P/Q)\leq k+2.$$ Furthermore, if $Q(z, w(z))$ has at least one nonzero root, then $deg_{w}(P)=deg_{w}(Q)+1\leq k+2;$ if all roots of $Q(z, w(z))$ are nonzero, then $deg_{w}(P)=deg_{w}(Q)+1\leq k+1;$ if $deg_{w}(Q)=0,$ then $deg_{w}(P)\leq 1.$\par In particular, whenever $deg_{w}(Q)=0$ and $deg_{w}(P)\leq 1$ and without the growth condition, any admissible meromorphic solution of the above delay differential equation (called Lenhart-Travis' type logistic delay differential equation) with reduced form can not be an entire function $w$ satisfying $\overline{N}(r, \frac{1}{w})=O(N(r, \frac{1}{w}));$ while if all coefficients are rational functions, then the condition $\overline{N}(r, \frac{1}{w})=O(N(r, \frac{1}{w}))$ can be omitted. Furthermore, any admissible meromorphic solution of the logistic delay differential equation (that is, for the simplest special case where $k=1$ and $deg_{w}(P/Q)=0$ ) satisfies that $N(r,w)$ and $T(r, w)$ have the same growth category. Some examples support our results.

math.CV↗

arXiv2020

A Probability Distribution and Location-aware ResNet Approach for QoS Prediction

In recent years, the number of online services has grown rapidly, invoke the required services through the cloud platform has become the primary trend. How to help users choose and recommend high-quality services among huge amounts of unused services has become a hot issue in research. Among the existing QoS prediction methods, the collaborative filtering(CF) method can only learn low-dimensional linear characteristics, and its effect is limited by sparse data. Although existing deep learning methods could capture high-dimensional nonlinear features better, most of them only use the single feature of identity, and the problem of network deepening gradient disappearance is serious, so the effect of QoS prediction is unsatisfactory. To address these problems, we propose an advanced probability distribution and location-aware ResNet approach for QoS Prediction(PLRes). This approach considers the historical invocations probability distribution and location characteristics of users and services, and first use the ResNet in QoS prediction to reuses the features, which alleviates the problems of gradient disappearance and model degradation. A series of experiments are conducted on a real-world web service dataset WS-DREAM. The results indicate that PLRes model is effective for QoS prediction and at the density of 5%-30%, which means the data is sparse, it significantly outperforms a state-of-the-art approach LDCF by 12.35%-15.37% in terms of MAE.

cs.SE↗

arXiv2020

PPG-based singing voice conversion with adversarial representation learning

Singing voice conversion (SVC) aims to convert the voice of one singer to that of other singers while keeping the singing content and melody. On top of recent voice conversion works, we propose a novel model to steadily convert songs while keeping their naturalness and intonation. We build an end-to-end architecture, taking phonetic posteriorgrams (PPGs) as inputs and generating mel spectrograms. Specifically, we implement two separate encoders: one encodes PPGs as content, and the other compresses mel spectrograms to supply acoustic and musical information. To improve the performance on timbre and melody, an adversarial singer confusion module and a mel-regressive representation learning module are designed for the model. Objective and subjective experiments are conducted on our private Chinese singing corpus. Comparing with the baselines, our methods can significantly improve the conversion performance in terms of naturalness, melody, and voice similarity. Moreover, our PPG-based method is proved to be robust for noisy sources.

cs.SD↗

arXiv2020

Precoding and Transmit Antenna Subarray Selection for Secure Hybrid Spatial Modulation

Spatial modulation (SM) is a particularly important form of multiple-input-multiple-output (MIMO). Unlike traditional MIMO, it uses both modulation symbols and antenna indices to carry information. In this paper, to avoid the high cost and circuit complexity of fully-digital SM, we mainly consider the hybrid SM system with a hybrid precoding transmitter architecture, combining a digital precoder and an analog precoder. Here, the partially-connected structure is adopted with each radio frequency chain (RF) being connected to a transmit antenna subarray (TAS). In such a system, we made an investigation of secure hybrid precoding and transmit antenna subarray selection (TASS) methods. Two hybrid precoding methods, called maximizing the approximate secrecy rate (SR) via gradient ascent (Max-ASR-GA) and maximizing the approximate SR via alternating direction method of multipliers (Max-ASR-ADMM), are proposed to improve the SR performance. As for TASS, a high-performance method of maximizing the approximate SR (Max-ASR) TASS method is first presented. To reduce its high complexity, two low-complexity TASS methods, namely maximizing the eigenvalue (Max-EV) and maximizing the product of signal-to-interference-plus-noise ratio and artificial noise-to-signal-plus-noise ratio (Max-P-SINR-ANSNR), are proposed. Simulation results will demonstrate that the proposed Max-ASR-GA and Max-ASR-ADMM hybrid precoders harvest substantial SR performance gains over existing method. For TASS, the proposed three methods Max-ASR, Max-EV, and Max-P-SINR-ANSNR perform better than existing leakage method. Particularly, the proposed Max-EV and Max-P-SINR-ANSNR is low-complexity at the expense of a little performance loss compared with Max-ASR.

eess.SP↗