Yi Chen

—

Unclaimed · AI-structured profileOpenAlex

Yi Chen — early-stage AI researcher spanning multimodal physical-reasoning benchmarks and generative video AI

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: no aggregate publication count or venue record is available for this researcher in the current data scope. This axis would measure total research output volume, venue quality, and author position on recent papers.

Impact—

Not yet scored: citation counts, h-index and downstream uptake are unavailable in the current snapshot. This axis would measure scholarly influence through citations to the two reviewed arXiv works and their reception.

Focus0/100

The focus score is 0 because the observed field distribution is exactly even: 50% artificial intelligence (multimodal physical-reasoning benchmarks) and 50% multimedia computing and generative AI, with the two 2026 arXiv papers each anchoring a different cluster. Within the current two-paper sample there is no dominant specialization, so concentration cannot be established; the split may reflect deliberate breadth or early-stage direction-setting rather than lack of focus.

Collaboration55/100

The collaboration score is 55, based on 15 distinct coauthors across the reviewed records and a recurring group of 12 frequent collaborators, including Qiang Chen, Hao Guo, Huatai Zhu and Haifeng Li. On both cataloged 2026 papers Yi Chen holds a middle author position without corresponding-author status, indicating a contributing role inside team-based projects rather than sole leadership; the moderate score reflects a real but not yet extensive network in the current sample.

Activity—

Not yet scored: the available scope contains only two 2026 arXiv records, which is too thin to establish a publication cadence. This axis would measure yearly output rhythm and sustained research engagement over time.

Reproducibility60/100

The reproducibility score is 60, the mean across 2 papers assessed by Evolve. The split is wide: FireWorldBench rates high because its 520 fire worlds and 9,074 QA pairs are grounded in reproducible FDS simulation, while CogenPVG rates low because its human evaluation is weak — inter-rater κ=0.33 with no significance tests. Overall readiness is thus moderate and benchmark-heavy rather than uniform.

Research summary

Yi Chen is an early-stage AI researcher whose cataloged work, name-resolved from public bibliographic data, spans two adjacent areas: multimodal physical-reasoning benchmarks for large models, and multimedia computing with generative AI. The Evolve-reviewed sample consists of two 2026 arXiv preprints on which Yi Chen is a middle author and not corresponding author. FireWorldBench contributes an FDS-simulation-grounded benchmark of 520 coupled-field fire worlds with 9,074 QA pairs that probes where multimodal LLMs fail at physical reasoning; Evolve rates its reproducibility readiness high while noting the gold answers are simulation-derived rather than measured. CogenPVG proposes the first ELM-guided multi-agent framework for persuasive video generation, combining generator-critic reflection with storyboard planning; Evolve credits the novel task framing but flags weak human-study evidence (κ=0.33, no significance tests) and low reproducibility readiness. The field distribution is evenly split at 50/50 between the two areas, which yields a focus score of 0 in the current sample. Collaboration is moderate (score 55), with 15 distinct coauthors and a recurring core of 12 frequent collaborators. Aggregate output, citation impact and publication cadence are not yet scored because the underlying data is out of scope for this snapshot; both records are recent preprints, so this profile should be read as a provisional, evolving picture.

Collaboration pattern

Yi Chen's reviewed records involve 15 distinct coauthors, with a recurring core of 12 frequent collaborators including Qiang Chen, Hao Guo, Huatai Zhu, Tairan Huang, Haifeng Li and Wenfeng Song. On both 2026 arXiv papers Yi Chen appears as a middle author without corresponding status, consistent with a contributing role in team-driven AI projects. Given the very common name, coauthor overlap should be verified before assuming identity with other 'Yi Chen' entries.

Research keywords

persuasive video generationmulti-agent LLM frameworksElaboration Likelihood Modelgenerator-critic reflectionstoryboard planningpersuasion evaluationmultimodal LLM benchmarksphysical world understandingfire dynamics simulation (FDS)coupled physical fieldscausal attribution

Highlights

FDS-grounded benchmark of complex physical world intelligence
FireWorldBench couples 520 fire worlds with 9,074 QA pairs built on fire dynamics simulation, exposing limits in multimodal LLM physical reasoning; Evolve rates its reproducibility readiness high but notes the gold answers are simulation-derived rather than measured.
First ELM-guided multi-agent framework for persuasive video generation
CogenPVG frames persuasion as a generative video task and orchestrates generator-critic reflection with storyboard planning; Evolve credits the novel task framing but flags weak human-study evidence (inter-rater κ=0.33, no significance tests) and low reproducibility readiness.
Benchmark-heavy reproducibility profile with a wide internal split
Across the 2 Evolve-assessed papers, reproducibility readiness averages 60: high for the simulation-grounded benchmark versus low for the evaluation-weak framework, making reproducibility the clearest differentiator in this early portfolio.

Timeline

2026Published on arXiv: FireWorldBench, an FDS-grounded benchmark of 520 coupled-field fire worlds and 9,074 QA pairs for physical world intelligence (middle author, not corresponding).
2026Published on arXiv: CogenPVG, the first ELM-guided multi-agent framework for persuasive video generation (middle author, not corresponding).
2026Both cataloged works land in the same year across two distinct clusters, producing an even 50/50 field split and a two-paper reproducibility mean of 60.

Papers in the Evolve catalog

Frequent co-authors

Qiang Chen×1 Hao Guo×1 Huatai Zhu×1 Tairan Huang×1 Yichao Cao×1 Hongyan Xu×1 Keke Huang×1 Haifeng Li×1 Xiu Su×1 Yuntian Xiao×1 Shoulong Zhang×1 Wenfeng Song×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.

Bibliometric dimensions are derived from OpenAlex-based data; no ORCID, institution or country record was available to disambiguate this very common name.
Only 2 Evolve-reviewed arXiv preprints are in scope; citation counts, journal output and h-index are not represented.
Output, impact and activity dimensions are unscored (dimMissing) because the underlying data is insufficient in this snapshot.
Both works are 2026 preprints with unknown peer-review status; scores may change substantially as the record matures.