Researcher in LLM-driven multimodal document generation and agentic visual RAG, active on 2026 arXiv preprints
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—
Publication-volume data is not yet available for this axis. It would measure total paper output, venue distribution, and first/corresponding-author share over time; the snapshot currently covers only two Evolve-reviewed 2026 preprints, so no score is computed.
Impact—
No citation or impact metrics are available yet. This axis would measure citations to the reviewed papers, citation velocity, and downstream uptake by other work; because the underlying data is missing, it is left unscored rather than estimated.
Focus0/100
Focus scores 0, consistent with an even field split: the raw distribution assigns 0.5 to natural language processing (multimodal document generation) and 0.5 to multimodal retrieval-augmented generation / vision-language agents. Neither direction dominates, so no concentrated specialization is detectable in this snapshot. The two cataloged papers map one-to-one onto these two fields, reading as a deliberately dual-track agenda rather than a measurable core area.
Collaboration53/100
Collaboration scores 53, based on 13 distinct coauthors in the bibliometric snapshot. The recurring pool is compact: Yucheng Shen, Shuaiqiang Wang, Jianmin Wu, and Lingyong Yan appear repeatedly in the frequent-coauthor list, alongside Runze Li, Yukun Zhao, Can Xu, Jiulong Wu, and Min Cao. A moderate score with a stable core suggests team-based projects rather than a wide, loose network; both 2026 papers were multi-authored group efforts.
Activity—
Year-by-year publication records are insufficient to compute an activity score. This axis would measure recent output cadence and research continuity over time; the snapshot only confirms two 2026 preprints, so no trend can be established.
Reproducibility38/100
Reproducibility averages 38 across the 2 papers Evolve has assessed (reproPapers = 2). The poster-generation paper, PosterVisor, is rated medium readiness, while SCORE is rated low, with Evolve flagging reliance on an LLM-judge metric without variance estimates and partly non-comparable baselines. A below-midpoint mean indicates that methods, prompts, contracts, or training details are not yet fully specified for independent replication without further author input.
Research summary
Dawei Yin is a researcher working at the intersection of natural language processing and multimodal retrieval-augmented generation, with an Evolve catalog spanning two 2026 arXiv preprints. The first, SCORE ("Navigating Sparse Evidence"), makes evidence selection and consolidation an explicit, trainable stage of agentic visual RAG, using GRPO reinforcement learning and reporting state-of-the-art results on three benchmarks; Evolve's review notes the gains rest on an LLM-judge metric without variance estimates and partly non-comparable baselines. The second, PosterVisor, introduces frozen rubric contracts and repair-aware rechecking for LLM-based scientific poster generation, with gains described as real but modest and some ablation trade-offs. Yin is a middle author on the former and last author (non-corresponding) on the latter. Bibliometric radar data records 13 distinct coauthors, with recurring collaborators including Yucheng Shen, Shuaiqiang Wang, Jianmin Wu, and Lingyong Yan. The field distribution splits evenly (0.5/0.5) between multimodal document generation and multimodal RAG / vision-language agents, yielding a focus score of 0 — no dominant specialization is yet measurable. Reproducibility averages 38 across the two assessed papers (medium and low readiness). This is a full-tier but early-stage profile: the output, impact, and activity axes lack underlying data and are left unscored.
Collaboration pattern
Work is team-based: 13 distinct coauthors appear in the snapshot, with Yucheng Shen, Shuaiqiang Wang, Jianmin Wu, and Lingyong Yan recurring most often. On the two reviewed papers Yin is neither first nor corresponding author — middle author on SCORE and last (non-corresponding) author on PosterVisor — consistent with a senior but supporting role in group projects spanning visual RAG and document generation.
Research keywords
Highlights
Persistent contract-based control for LLM poster generation
SCORE: explicit evidence selection and consolidation for agentic visual RAG
Timeline
- 2026Co-authored SCORE on agentic visual RAG with explicit context selection and consolidation (arXiv), as a middle author.
- 2026Last-authored (non-corresponding) PosterVisor, on scientific poster generation via recursive semantic-geometric contracts (arXiv).
- 2026Both Evolve-cataloged papers fall in 2026; they are the only entries in the current snapshot, so no multi-year timeline can be drawn.
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.
- Name resolved from public bibliographic data only; no ORCID, affiliation, or country was provided, so identity with same-name researchers is not confirmed.
- Radar scores and statistics come from a limited snapshot covering only the two Evolve-reviewed 2026 preprints; citation, output, and activity data are unavailable.
- The reproducibility mean (38) is computed from just 2 assessed papers and should be read as indicative rather than stable.
- Collaboration counts reflect co-authorship within the snapshot (arXiv-sourced papers) and may omit work outside that coverage.