Yinan Liu

Northeastern University · China

Unclaimed · AI-structured profileOpenAlex

Northeastern University researcher working on LLM-based knowledge graph completion and event linking in NLP

papers here

reviewed by Evolve

first-author

corresponding

—

works · OpenAlex

h-index · OpenAlex

Bibliometric figures come from OpenAlex (2026-10-17) 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—

Output is not scored yet: publication-level volume data for this identity is insufficient in the OpenAlex record. The axis would measure total works, yearly productivity, and authorship roles across venues once the paper-level record is verified.

Impact—

Impact is not scored yet: the OpenAlex record lacks reliable citation-linked data for this identity. The axis would measure citation counts, h-index trajectory, and uptake of reviewed papers once a complete citation record exists.

Focus100/100

The focus dimension scores 100, reflecting a tightly clustered topic profile. OpenAlex topic counts show Topic Modeling (9) and Advanced Graph Neural Networks (9) tied at the top, followed by Semantic Web and Ontologies (6), Recommender Systems (3), and Natural Language Processing Techniques (2) — nearly all effort goes to methods that join language and graph structure rather than scattered topics.

Collaboration46/100

Collaboration scores 46, a mid-range value. The recent three-year record shows 9 distinct coauthors, with repeated partners including Xiaochun Yang, Ziyang Zhang, Boyi Xue, Yingxuan Huang, Bin Wang, Jiaqi Wang, and Dongying Lin. Both Evolve-catalog papers list Yinan Liu in a middle author position without corresponding-author credit, indicating contribution within an established team rather than independent group leadership so far.

Activity77/100

Activity scores 77 on the strength of a dense recent record: OpenAlex lists 13 works in the trailing three-year window, with the year-level counts showing all 13 recorded in 2026. That concentration signals a rapid ramp-up of output rather than a long publication history, consistent with an early-stage profile.

Reproducibility55/100

Reproducibility scores 55, computed from a single reviewed paper (reproPapers 1, mean 55). That paper carries a medium repro-readiness rating: the MACE plug-in's reported gains are credible, yet the review notes partly floor-effect benchmark results and high online inference cost, so independent verification budget should be planned before reuse.

Research summary

Yinan Liu is a researcher at Northeastern University (China), profiled here from public OpenAlex bibliographic data. The indexed record is early-stage but tightly focused: 13 works appear in the trailing three-year window, all recorded under 2026, and the topic distribution concentrates on Topic Modeling (9) and Advanced Graph Neural Networks (9), followed by Semantic Web and Ontologies (6), Recommender Systems (3), and Natural Language Processing Techniques (2). Two 2026 arXiv papers sit in the Evolve catalog. The first, on enhancing event candidate acquisition for event linking, was reviewed by Evolve: it presents MACE, an LLM multi-agent evidence-refinement plug-in that lifts three event linkers on two benchmarks, with gains judged credible but partly attributable to floor effects and with high online inference cost. The second applies LLMs to knowledge graph completion by combining discrete structural coding with similar-entity information; it has not yet been reviewed. On both papers Yinan Liu holds a middle author position without corresponding-author credit, consistent with a mid-range collaboration score (46) built on 9 distinct coauthors, including frequent partners such as Xiaochun Yang and Ziyang Zhang. Radar dimensions read focus 100, activity 77, and reproducibility 55, while output and impact are not yet scoreable from the available record.

Collaboration pattern

Collaboration scores 46 (mid-range): 9 distinct coauthors in the recent three-year window. The frequent-partner list is headed by Xiaochun Yang and also includes Ziyang Zhang, Boyi Xue, Yingxuan Huang, Bin Wang, Jiaqi Wang, Dongying Lin, and Yang Yang. In both Evolve-catalog papers Yinan Liu appears as a middle author without corresponding credit, suggesting steady contribution inside an established NLP/knowledge-graph team rather than independent group leadership so far.

Research keywords

event linkingcandidate event acquisitionknowledge graph completionmulti-agent LLM systemsknowledge base groundingevent argument extractionretrieve-and-ranktopic modelinggraph neural networkssemantic web and ontologiesrecommender systems

Highlights

MACE: an LLM multi-agent evidence-refinement plug-in for event linking
In the 2026 arXiv study on enhancing event candidate acquisition, MACE refines candidate evidence with multiple LLM agents and lifts three event linkers on two benchmarks. Evolve judges the gains credible but partly due to floor effects, notes high online inference cost, and rates reproducibility as medium.
Discrete structural coding meets similar-entity information in LLM-based KG completion
A second 2026 arXiv paper combines discrete structural coding of graph structure with similar-entity information to guide large language models on knowledge graph completion. Yinan Liu is a middle author on this paper, which Evolve has not yet reviewed, so its evidence base remains unassessed here.
A text-plus-graph research program in its first indexed year
OpenAlex topic counts — Topic Modeling 9, Advanced Graph Neural Networks 9, Semantic Web and Ontologies 6, Recommender Systems 3, Natural Language Processing Techniques 2 — show near-exclusive concentration on methods joining language and graph structure, yielding a focus dimension score of 100 across 13 recent works.

Timeline

2026OpenAlex records 13 works for this identity in the most recent year, with all indexed bibliographic activity clustered in 2026.
2026The arXiv preprint 'Enhancing Event Candidate Acquisition for Event Linking' appears; Evolve's review finds the MACE plug-in lifts three event linkers on two benchmarks, with partly floor-effect gains and high online cost.
2026A second arXiv preprint on LLM-based knowledge graph completion combining discrete structural coding with similar entity information is indexed; Evolve has not yet reviewed it.
2026The first Evolve-reviewed paper sets this profile's reproducibility baseline: a medium readiness rating, feeding a reproducibility dimension score of 55 from one paper.

Papers in the Evolve catalog

2026-09-12
Enhancing Event Candidate Acquisition for Event Linking

Frequent co-authors

Xiaochun Yang×2 Ziyang Zhang×1 Boyi Xue×1 Yingxuan Huang×1 Bin Wang×1 Jiaqi Wang×1 Dongying Lin×1 Yang Yang×1 Bin Wang×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.

All bibliometric figures come from OpenAlex; the identity was resolved by name matching against public bibliographic data, and 'Yinan Liu' is a common name, so author attribution may merge or split works incorrectly.
Output and impact axes are flagged as missing: citation counts and the h-index/i10-index fields in the source record read as 0 and are not treated as reliable indicators.
Only two papers from this researcher appear in the Evolve catalog, and only one carries a full Evolve review; research output beyond these preprints is not covered here.
Both catalog papers are 2026 arXiv preprints; peer-reviewed venue status, funding, and career details are not captured in the input data.