Jing Xu

Nankai University

Unclaimed · AI-structured profileOpenAlex

Jing Xu — Nankai researcher building LLM-driven frameworks and benchmarks for visual analytics and scientific ML

papers here

reviewed by Evolve

first-author

corresponding

—

works · OpenAlex

—

h-index · OpenAlex

Bibliometric figures come from OpenAlex (2026-10-09) and cover the author's entire public record, not just the papers listed here. Reproducibility signal comes from Evolve's own reviews.

Research profile

Output—

Not yet scored: the underlying productivity data for this axis — total publication count and output volume over time — is unavailable in the current snapshot, which covers only the two reviewed 2026 preprints.

Impact—

Not yet scored: citation counts, h-index-style measures, and downstream uptake are absent from the available data, so this axis — which would measure scholarly influence and adoption of the work — cannot be estimated here.

Focus0/100

The focus score is 0, the low end of the scale. The raw field distribution is consistent with it: the two reviewed papers split evenly, 0.5 scientific machine learning and 0.5 human-computer interaction (visual analytics), so no single-field concentration is visible yet. That even split reflects a deliberately broad early agenda — LLM-driven PINN design on one side, agentic visual analytics on the other — rather than scattered work; a larger paper sample could move this score.

Collaboration50/100

The collaboration score is 50, mid-range. The raw record shows 11 distinct coauthors, with frequent collaborators including Mingyang Yu, Xu Yang, Jun Zhang, Xiaolong Wang, Keqian Li, and Weikai Yang. On both reviewed 2026 preprints Jing Xu holds a middle authorship position and is not the corresponding author, a pattern consistent with contributing inside larger teams rather than leading them at this stage.

Activity—

Not yet scored: a year-by-year publication cadence cannot be computed from a snapshot containing only two 2026 papers; this axis would measure sustained research output and recent publishing tempo.

Reproducibility55/100

The reproducibility score is 55, computed as a mean of 55 across 2 scored papers; both carry medium reproducibility readiness in Evolve's review. Concretely, WidgetVA ships a benchmark and diagnostic tooling but rests on an n=2 human baseline with flagged workflow circularity, while PINNForge reports solid statistics across 24 of 25 PDE tasks yet leaves novelty comparisons and uncounted LLM cost open. Both sit in the reproducible-with-effort band rather than turnkey.

Research summary

Jing Xu is a researcher at Nankai University whose visible record, as captured in Evolve's catalog, sits at the intersection of two fields in near-equal measure: scientific machine learning and human-computer interaction (visual analytics), each accounting for 0.5 of the field distribution. Both reviewed works are 2026 arXiv preprints on which Jing Xu holds a middle authorship position without corresponding-author role. The first, PINNForge, applies large language models to the automated design of physics-informed neural networks for PDE solving via execution-grounded evolutionary search; Evolve's review credits it with winning 24 of 25 PDE tasks with solid statistics while flagging novelty questions against Lang-PINN and uncounted LLM cost. The second, WidgetVA, proposes a widget-centric abstraction and benchmark for agentic visual analytics driven by vision-language models; reviewers praise the diagnostic design but note an n=2 human baseline and possible workflow circularity. Both papers carry medium reproducibility readiness, yielding a reproducibility score of 55 (a reproMean of 55 across 2 papers). With 11 distinct coauthors and a collaboration score of 50, this is an early-stage, preprint-phase profile: citation-based output, impact, and activity metrics are not yet available and remain unscored.

Collaboration pattern

Jing Xu's reviewed work is team-produced: 11 distinct coauthors appear in the record, with a recurring circle including Mingyang Yu, Xu Yang, Jun Zhang, Xiaolong Wang, Keqian Li, Yutong Chen, Zhike Tang, Zhihao Mai, Zhihao Shuai, Danli Luo, and Weikai Yang. Across both 2026 preprints Jing Xu holds a middle authorship position without corresponding-author status, suggesting a contributor role within larger groups spanning scientific machine learning and visual analytics at Nankai University.

Research keywords

visual analyticsagentic AIvision-language modelsGUI agentsbenchmark evaluationwidget abstractionphysics-informed neural networksPDE solvinglarge language modelsevolutionary searchautomated algorithm designexecution feedback

Highlights

WidgetVA: widget-centric framework and benchmark for agentic visual analytics
The 2026 arXiv preprint proposes a widget abstraction so vision-language model agents can drive visual analytics; Evolve's review finds the diagnostic design solid but flags an n=2 human baseline and workflow circularity, with medium reproducibility readiness.
PINNForge: execution-grounded LLM evolution of physics-informed networks
The 2026 arXiv preprint uses large language models with execution feedback to evolve PINN designs for PDE solving; Evolve credits it with winning 24 of 25 PDE tasks and solid statistics, tempering the claim over novelty versus Lang-PINN and uncounted LLM cost.

Timeline

2026PINNForge appears on arXiv — LLM-driven, execution-grounded evolutionary design of physics-informed neural networks for PDE solving — later reviewed into the Evolve catalog with medium reproducibility readiness.
2026WidgetVA appears on arXiv — a widget-centric framework and benchmark for agentic, vision-language-model-driven visual analytics — also reviewed with medium reproducibility readiness.
2026Both preprints enter the Evolve catalog the same year, splitting evenly between scientific machine learning and human-computer interaction (visual analytics) and anchoring this early-stage profile.

Papers in the Evolve catalog

Frequent co-authors

Mingyang Yu×1 Xu Yang×1 Jun Zhang×1 Xiaolong Wang×1 Keqian Li×1 Yutong Chen×1 Zhike Tang×1 Zhihao Mai×1 Zhihao Shuai×1 Danli Luo×1 Weikai Yang×1

About this profile

This profile is AI-structured by Evolve.Pub from public bibliographic data (OpenAlex) and Evolve's own paper reviews. It is not written or endorsed by the researcher unless the profile is claimed. Radar dimensions are computed from real bibliometrics with fixed formulas — raw figures are shown alongside; missing dimensions mean insufficient data, never zero. Errors? Claim the profile to correct or hide it.

Jing Xu is a common name; this profile is name-resolved from public bibliographic data and the two reviewed papers, with no ORCID provided for disambiguation.
Bibliometrics come from the provided snapshot only; citation counts, venue history, h-index, and total output are unavailable, so output, impact, and activity remain unscored.
Both catalog entries are 2026 arXiv preprints; peer-reviewed venue acceptance is not reflected in the current data.
The reproducibility score is a mean over only 2 papers (reproMean 55) and should be read as indicative rather than settled.