Early-stage AI researcher at the Chinese Academy of Social Sciences: LLM compliance evaluation and GUI agents
2
papers here
2
reviewed by Evolve
2
first-author
0
corresponding
—
works · OpenAlex
0
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-08) 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: no reliable publication-output totals are available — OpenAlex indexes only 2 works (both 2026) with no aggregated venue totals, so cumulative output volume and productivity cannot yet be measured.
Impact—
Not yet scored: citation-based impact cannot be assessed — the recorded h-index and i10-index are both 0 and no per-paper citation counts are provided, so this axis would measure citation influence and uptake of the two 2026 papers.
Focus0/100
The focus score of 0 reflects a spread portfolio rather than a consolidated specialty. OpenAlex tags this profile with five distinct topics, each appearing once — topic modeling, robotic path planning, AI ethics and social impact, adversarial robustness, and multimodal applications — while the field distribution splits exactly 50/50 between LLM safety/regulatory-compliance evaluation and GUI agents/vision-language models. With only 2 indexed works, no dominant theme has emerged yet.
Collaboration61/100
The collaboration score of 61 rests on 20 distinct co-authors recorded in the three-year window — a broad network for a profile with only 2 indexed works. Evolve's catalog lists recurring partners including Ruien Li, Yaoyao Peng, and Wanxin Ren. Both 2026 arXiv papers carry Zhang as first author but not corresponding author, consistent with team-based authorship.
Activity32/100
The activity score of 32 is grounded in thin but current output: OpenAlex records 2 works in the last three years, and the counts-by-year data places both in 2026. All indexed activity is concentrated in a single recent year with no earlier record in this snapshot — a profile whose publication record is still accumulating.
Reproducibility20/100
Reproducibility scores 20, computed as the mean readiness across 2 reviewed papers, both rated low. Evolve flags that the EADC benchmark leaves its judge LLM unspecified and reports no error bars, while the GUI-agent paper claims 2-3 point gains on 5 benchmarks without significance tests — both gaps limit independent verification.
Research summary
Yan Zhang, affiliated with the Chinese Academy of Social Sciences (China), works across two complementary branches of applied artificial intelligence. The Evolve catalog records two 2026 arXiv papers, both with Zhang as first author. The first, EADC, introduces a law-grounded benchmark of 4,435 question–answer items spanning 24 large language models, probing advanced and deep-level regulatory compliance with reference to the EU AI Act; Evolve's review credits the benchmark design but flags an unspecified judge LLM and missing error bars. The second proposes masked trajectory prediction (MaP) for omni-dimensional GUI agent navigation, unifying navigation tasks and delivering 2–3 point gains over mixture training across five benchmarks, though significance tests are absent. OpenAlex bibliometrics show 2 works in the last three years, both dated 2026, with 20 distinct co-authors and a field distribution split evenly (0.5/0.5) between LLM safety/regulatory-compliance evaluation and GUI agents/vision-language models. The recorded h-index and i10-index are both 0, so citation impact is not yet measurable, and reproducibility readiness averages 20 across the two papers. The profile is name-resolved from public bibliographic data.
Collaboration pattern
Bibliometrics record 20 distinct co-authors within the three-year window — a wide network relative to only 2 indexed works — yielding a collaboration score of 61. Recurring names in the Evolve catalog include Ruien Li, Yaoyao Peng, Wanxin Ren, Yijia Zhang, and Wusheng Zhang. Both papers are first-authored but not corresponding-author works, suggesting Zhang leads drafting and experiments within a larger team spanning both the compliance-evaluation and GUI-agent projects.
Research keywords
Highlights
EADC: a law-grounded compliance benchmark for 24 LLMs
Masked trajectory prediction unifies GUI agent navigation
Timeline
- 2026First Evolve-catalog paper appears: EADC, a law-grounded LLM compliance benchmark (4,435 QA items, 24 LLMs), with Zhang as first author.
- 2026Second first-authored arXiv paper: GUI agent navigation unified via masked trajectory prediction, reporting gains on 5 benchmarks.
- 2026OpenAlex snapshot places both works in a single year with 20 distinct co-authors — the entire indexed record to date.
Papers in the Evolve catalog
Frequent co-authors
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.
- Bibliometrics come from OpenAlex only; no ORCID is linked and 'Yan Zhang' is a common name, so this profile is name-resolved from public bibliographic data.
- The recorded h-index and i10-index are 0 with no citation counts, so the output and impact dimensions are unscored rather than estimated.
- Both catalog papers are 2026 arXiv preprints; no peer-reviewed venue record appears in this snapshot.
- The reproducibility score reflects Evolve's review of only 2 papers, both rated low readiness.