Han Chen

—

Unclaimed · AI-structured profileOpenAlex

Researcher spanning LLM fact verification, long-context retrieval QA, and graph-based Android malware detection.

papers here

reviewed by Evolve

first-author

corresponding

—

works · OpenAlex

—

h-index · OpenAlex

Bibliometric figures come from OpenAlex (2026-10-02) 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—

No publication-count or output-volume metric is available for this axis yet. It would measure the volume and consistency of published work across venues and years; only three 2026 arXiv preprints are currently cataloged.

Impact—

Impact cannot be scored yet: no citation counts, h-index, or venue-level influence data are present in this snapshot. This axis would measure uptake of the researcher's work through citations and downstream scholarly influence.

Focus0/100

Topical focus is not yet scored, because normalized field classifications are missing for the cataloged papers. This axis would measure how concentrated or dispersed the researcher's output is across research fields and topics.

Collaboration41/100

The collaboration axis scores 39 on the platform's 0-100 scale, computed from six distinct coauthors observed across this researcher's cataloged papers. The recurring collaborator set includes Yingrui Li, Hanchen Wang, Hongmei Chen, Lu Qin, Wenjie Zhang, and Ying Zhang, suggesting stable small-team work rather than large consortia. Across the three 2026 preprints, Han Chen appears as first author twice and last author once, with no corresponding-author credit recorded.

Activity—

Activity is not yet scored: the catalog holds only papers dated 2026, leaving no multi-year cadence to measure. This axis would track publishing frequency and continuity over time.

Reproducibility55/100

Reproducibility is unscored because code, dataset, and artifact availability fields are empty for all cataloged papers. This axis would measure how far reported experiments can be reproduced from released materials.

Research summary

Han Chen is a researcher whose public bibliographic footprint in the Evolve catalog consists of three arXiv preprints, all dated 2026. Two are first-authored: a study asking what actually changes when fact-verification scores improve, built on evidence and answer accounting across trained verifiers and large language models, and HYDRA, which applies hierarchical graph contrastive learning to proactive adaptation of Android malware drift. The third lists Han Chen as last author: a controlled study of planning, routing, and reranking for long-context question answering that finds learned context planning fails to beat strong retrieval. Authorship is therefore split between first-author leadership on two papers and a last-author, senior role on the third, with no corresponding-author credit recorded. Thematically, the catalog sits at the intersection of evaluation methodology for LLM-based verification and retrieval systems and applied graph learning for security. The collaboration axis scores 39 on a 0-100 scale, computed from six distinct coauthors; recurring collaborators include Yingrui Li, Hanchen Wang, Hongmei Chen, Lu Qin, Wenjie Zhang, and Ying Zhang. Because Han Chen is a common name, this profile is name-resolved from public bibliographic data, and affiliation, venue, and citation information are not yet available.

Collaboration pattern

Six distinct coauthors appear across the cataloged papers, and the recurring set — Yingrui Li, Hanchen Wang, Hongmei Chen, Lu Qin, Wenjie Zhang, and Ying Zhang — suggests stable small-team collaboration. Authorship roles vary: first author on the fact-verification and HYDRA papers, last author on the long-context QA study, with no corresponding-author credit recorded. The collaboration score of 39 reflects a modest, early-visible network rather than a large coauthor base.

Research keywords

fact verificationlarge language modelstrained verifiersevidence and answer accountinglong-context question answeringretrieval-augmented generationcontext planningrerankingAndroid malware detectiongraph contrastive learningconcept drift adaptationcontrolled empirical evaluation

Highlights

Interrogating what fact-verification score improvements actually change
First-authored 2026 arXiv study accounting for evidence and answers across trained verifiers and LLMs to show what changes when fact-verification scores improve.
HYDRA: proactive adaptation to Android malware drift
First-authored 2026 preprint proposing HYDRA, which uses hierarchical graph contrastive learning to adapt proactively to Android malware drift.
Controlled evidence that learned context planning trails strong retrieval
Last-authored 2026 controlled study of planning, routing, and reranking for long-context QA, finding learned context planning fails to beat strong retrieval.

Timeline

2026First-authored arXiv preprint on evidence and answer accounting across trained verifiers and LLMs for fact verification.
2026First-authored HYDRA preprint applying hierarchical graph contrastive learning to Android malware drift adaptation.
2026Last-authored controlled study of planning, routing, and reranking for long-context QA.

Papers in the Evolve catalog

Frequent co-authors

Yingrui Li×2 Hanchen Wang×1 Hongmei Chen×1 Lu Qin×1 Wenjie Zhang×1 Ying Zhang×1 Han Chen×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.

Han Chen is a common name; this profile is name-resolved from public bibliographic data and may not represent a single individual.
All three cataloged papers are 2026 arXiv preprints; no peer-reviewed venue, affiliation, country, or ORCID is recorded in this snapshot.
Citation counts, h-index, and venue-level impact metrics are not available; only the collaboration dimension is scored.
Code, dataset, and artifact availability fields are empty, so reproducibility readiness cannot be assessed.