arXiv subjects

Ningyuan Xu

Publications and source records attributed to Ningyuan Xu.

4 recordsLinked to original sources

arXiv2026

SWE-Explore: Benchmarking How Coding Agents Explore Repositories

Repository-level coding benchmarks such as SWE-bench have driven a rapid surge in the capabilities of coding agents. Yet they usually treat coding tasks as a holistic, binary prediction problem (e.g., resolved or unresolved), neglecting fine-grained agent capabilities such as repository understanding, context retrieval, code localization, and bug diagnosis. In this paper, we introduce SWE-Explore, a benchmark that isolates the evaluation of repository exploration, a critical capability of coding agents. Given a repository and an issue, SWE-Explore asks an explorer to return a ranked list of relevant code regions under a fixed line budget. SWE-Explore covers 848 issues across 10 programming languages and 203 open-source repositories. For each instance, we derive line-level ground truth from independent agent trajectories that successfully solved the same issue, distilling the specific code regions their solution paths actually consulted. We evaluate exploration along coverage, ranking, and context-efficiency dimensions, showing that these metrics strongly track downstream repair behavior. Across a broad set of retrieval methods, general coding agents, and specialized localizers, we find that agentic explorers form a clear tier above classical retrieval. While file-level localization is already strong for modern methods, line-level coverage and efficient ranking remain the key axes differentiating state-of-the-art explorers.

cs.SE↗

arXiv2023

Stellar Populations in the Central 0.5 pc of Our Galaxy III: The Dynamical Sub-structures

We measure the 3D kinematic structures of the young stars within the central 0.5 parsec of our Galactic Center using the 10 m telescopes of the W.~M.~Keck Observatory over a time span of 25 years. Using high-precision measurements of positions on the sky, and proper motions and radial velocities from new observations and the literature, we constrain the orbital parameters for each young star. Our results show two statistically significant sub-structures: a clockwise stellar disk with 18 candidate stars, as has been proposed before, but with an improved disk membership; a second, almost edge-on plane of 10 candidate stars oriented East-West on the sky that includes at least one IRS 13 star. We estimate the eccentricity distribution of each sub-structure and find that the clockwise disk has <$e$> = 0.39 and the edge-on plane has <$e$> = 0.68. We also perform simulations of each disk/plane with incompleteness and spatially-variable extinction to search for asymmetry. Our results show that the clockwise stellar disk is consistent with a uniform azimuthal distribution within the disk. The edge-on plane has an asymmetry that cannot be explained by variable extinction or incompleteness in the field. The orientation, asymmetric stellar distribution, and high eccentricity of the edge-on plane members suggest that this structure may be a stream associated with the IRS 13 group. The complex dynamical structure of the young nuclear cluster indicates that the star formation process involved complex gas structures and dynamics and is inconsistent with a single massive gaseous disk.

astro-ph.GA↗

arXiv2022

Image restoration quality assessment based on regional differential information entropy

With the development of image recovery models,especially those based on adversarial and perceptual losses,the detailed texture portions of images are being recovered more naturally.However,these restored images are similar but not identical in detail texture to their reference images.With traditional image quality assessment methods,results with better subjective perceived quality often score lower in objective scoring.Assessment methods suffer from subjective and objective inconsistencies.This paper proposes a regional differential information entropy (RDIE) method for image quality assessment to address this problem.This approach allows better assessment of similar but not identical textural details and achieves good agreement with perceived quality.Neural networks are used to reshape the process of calculating information entropy,improving the speed and efficiency of the operation. Experiments conducted with this study image quality assessment dataset and the PIPAL dataset show that the proposed RDIE method yields a high degree of agreement with people average opinion scores compared to other image quality assessment metrics,proving that RDIE can better quantify the perceived quality of images.

eess.IV↗

arXiv2021

A Dataset and Method for Hallux Valgus Angle Estimation Based on Deep Learing

Angular measurements is essential to make a resonable treatment for Hallux valgus (HV), a common forefoot deformity. However, it still depends on manual labeling and measurement, which is time-consuming and sometimes unreliable. Automating this process is a thing of concern. However, it lack of dataset and the keypoints based method which made a great success in pose estimation is not suitable for this field.To solve the problems, we made a dataset and developed an algorithm based on deep learning and linear regression. It shows great fitting ability to the ground truth.

cs.CV↗