LLM-agent control and agentic visual RAG researcher; co-author of two 2026 arXiv papers reviewed in the Evolve catalog
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-09-20) 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: beyond the two reviewed catalog papers, no verified publication-volume data exists in this snapshot. This axis would measure total scholarly output and consistency over time; the underlying data is insufficient, so no estimate is offered.
Impact—
Not yet scored: citation counts and downstream-influence indicators are absent from the current bibliometric snapshot. This axis would measure citations, uptake, and real-world adoption of the work; the data is unavailable, so no score or estimate is given.
Focus0/100
The focus score is 0, reflecting an even split (0.5 / 0.5) between two fields: natural language processing for multimodal document generation, and multimodal retrieval-augmented generation / vision-language agents. Each reviewed 2026 paper sits in one field, so the profile reads as a deliberate two-track portfolio — LLM-agent control of generated documents on one side, agentic visual evidence use on the other — rather than concentration in a single specialty.
Collaboration53/100
Collaboration scores 53, with 13 distinct co-authors recorded. On both reviewed 2026 papers the researcher sits in a middle author position rather than as corresponding author, consistent with specialist roles inside larger team efforts. Recurring names such as Shuaiqiang Wang, Dawei Yin, Jianmin Wu, and Yucheng Shen point to a stable core group behind the catalog entries.
Activity—
Not yet scored: the snapshot holds only two 2026 arXiv entries and no multi-year publication series, so research tempo cannot be computed. This axis would measure publishing cadence and recent engagement over time; the data is insufficient, and no estimate is made.
Reproducibility38/100
Reproducibility scores 38, a mean over the 2 catalog papers with assessable artifacts (reproMean 38, reproPapers 2). The poster-generation paper rates medium readiness: frozen rubric contracts and rechecking loops aid replication. The agentic visual RAG paper rates low: reported SOTA rests on an LLM-judge metric without variance estimates and partly non-comparable baselines, complicating independent verification.
Research summary
Lingyong Yan is a researcher working at the intersection of large language model agents and multimodal information access, profiled here from public bibliographic data held in the Evolve catalog. The current snapshot contains two Evolve-reviewed 2026 arXiv papers, and Yan is listed in a middle author position on both. The first, "Navigating Sparse Evidence: Agentic Visual RAG via Explicit Context Selection and Consolidation" (SCORE), makes visual-evidence selection and consolidation an explicit, trainable stage of agentic visual retrieval-augmented generation, using GRPO reinforcement learning and reporting state-of-the-art results on three benchmarks — gains that Evolve notes rest on an LLM-judge metric without variance estimates and partly non-comparable baselines. The second, "From Transient Prompts to Persistent Control" (PosterVisor), targets scientific poster generation with frozen rubric contracts and repair-aware rechecking, delivering real but modest improvements with some ablation trade-offs. The field distribution is an even 0.5/0.5 split between natural language processing for multimodal document generation and multimodal retrieval-augmented generation / vision-language agents, giving a two-track research profile. Thirteen distinct co-authors appear in the record, with recurring names such as Shuaiqiang Wang, Dawei Yin, Jianmin Wu, and Yucheng Shen pointing to a stable core team. No institutional affiliation or citation-level impact data is available, so the impact, output, and activity axes remain unscored.
Collaboration pattern
Records show 13 distinct co-authors, with repeated names including Shuaiqiang Wang, Dawei Yin, Jianmin Wu, Yucheng Shen, Runze Li, Yukun Zhao, Can Xu, Jiulong Wu, and Min Cao — suggesting a stable core team. On both reviewed 2026 papers Lingyong Yan appears in a middle author position, not as corresponding author, consistent with specialist contributions inside larger group efforts rather than sole project leadership.
Research keywords
Highlights
Explicit, trainable evidence selection for agentic visual RAG
Contract-based persistent control for scientific poster generation
Timeline
- 2026Enters the Evolve catalog as a middle-position co-author of "Navigating Sparse Evidence" (SCORE), an arXiv paper on agentic visual retrieval-augmented generation with explicit context selection and consolidation.
- 2026Co-authors "From Transient Prompts to Persistent Control" (PosterVisor) on arXiv, again in a middle author position, applying recursive semantic-geometric contracts to scientific poster generation.
- 2026The catalog footprint spans two fields at parity (0.5 / 0.5): multimodal document generation in NLP and multimodal retrieval-augmented generation / vision-language agents.
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.
- Radar scores and collaboration statistics are computed from a public bibliometric snapshot only; no citation, h-index, or venue-level metrics were available, which is why output, impact, and activity are unscored.
- No institutional affiliation, country, or ORCID is present in the data; this profile is name-resolved from public bibliographic records and may conflate same-name researchers.
- Qualitative verdicts cover only the two Evolve-reviewed 2026 arXiv preprints; peer-review status and any wider publication record are unknown.
- The frequent-coauthor list contains the researcher's own name and duplicate spellings (e.g. two forms of Jianmin Wu), indicating some record noise in the underlying data.