Audit-first researcher working on trustworthy generative search evaluation and AI-assisted RTL verification auditing
4
papers here
4
reviewed by Evolve
2
first-author
0
corresponding
—
works · OpenAlex
0
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-16) 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—
Insufficient data: this axis would measure total publication volume, output rate, and venue spread across the career. The cataloged record here covers only two 2026 arXiv preprints, so no volume or trajectory estimate is computed.
Impact—
Not yet scorable: impact would track citations, references to the work by others, and uptake over time. Neither preprint in scope carries citation data in the provided snapshot, so no influence level is inferred.
Focus0/100
Scores 100/100, the maximum. The raw field distribution behind this axis lists a single entry — 'hardware security and AI for EDA (RTL verification)' — with no spread across other fields, indicating a tightly concentrated record. The two cataloged preprints sit in RTL verification and in trustworthy generative search/RAG evaluation, linked by a shared audit-and-verification methodology rather than scattered across unrelated areas.
Collaboration50/100
Scores 36/100 on a record of 5 distinct coauthors, with repeated collaborators Chuhong Xu, Gangzhen Qian, Zhaoyi Li, Hang Xiao, and Lu Yi. Both cataloged papers are multi-authored and neither lists Zhou as corresponding author — one as first author, one as middle author — consistent with a contributor role inside small teams rather than an independent hub.
Activity72/100
Insufficient data: this axis would measure publishing cadence — papers per year, gaps, and recency of output. The visible window here contains only two preprints, both dated 2026, which cannot establish a multi-year activity pattern, so no score is assigned.
Reproducibility66/100
Scores 55/100, a mean over exactly 1 reproducibility-scored paper. The first-authored auditing paper carries a 'high' readiness rating from Evolve's review, while the preregistered RTL audit is rated 'medium' after its AI runs yielded no valid analysis. One scored paper is a thin base; treat the mean as provisional.
Research summary
Kainan Zhou is an early-stage researcher whose public bibliographic footprint, as indexed here, consists of two 2026 arXiv preprints linked through the Evolve review catalog; the profile is name-resolved from public bibliographic data. The work spans two fields: trustworthy generative search and RAG evaluation in information retrieval/NLP, and hardware security with AI for EDA, specifically RTL verification. On the first, Zhou is first author of 'Claim-Gated Source-Risk Auditing for Generative Search,' which specifies a four-predicate omission audit for generative search; Evolve's review notes rigorous but self-referential conformance tests, leaving semantic validity to future work. On the second, Zhou contributes as a middle author to 'Trust, but Validate the Instrument,' a preregistered audit of AI-generated RTL verification plans; its AI runs produced no valid analysis, and the review frames it as a valuable fail-closed benchmark and baseline rather than model-performance evidence. The methodological through-line is audit-first: explicit specifications, conformance testing, preregistration, and honest reporting of null outcomes. Radar scores computed from the underlying bibliometric data give focus 100/100, collaboration 36/100 on 5 distinct coauthors, and reproducibility 55/100 averaged over 1 scored paper. Output, impact, and activity dimensions lack sufficient data and are left unscored. No institutional affiliation, country, or ORCID is attached to this profile.
Collaboration pattern
Collaboration is small-team and repeated: 5 distinct coauthors appear in the record, with Chuhong Xu, Gangzhen Qian, Zhaoyi Li, Hang Xiao, and Lu Yi as frequent collaborators, yielding a radar collaboration score of 36/100. Zhou held author position 1 on the generative-search audit and a middle position on the RTL verification audit, and is corresponding author on neither — a pattern consistent with a contributing specialist within compact project teams.
Research keywords
Highlights
A four-predicate omission audit for generative search
Preregistered audit of AI-generated RTL verification plans
High reproducibility readiness on the first-authored audit
Timeline
- 2026First-authored preprint 'Claim-Gated Source-Risk Auditing for Generative Search' appears on arXiv, specifying a four-predicate omission audit for generative search (information retrieval / NLP).
- 2026Middle-authored preprint 'Trust, but Validate the Instrument' appears on arXiv, a preregistered audit of AI-generated RTL verification plans (hardware security and AI for EDA).
- 2026Both preprints enter the Evolve review catalog with reproducibility-readiness ratings of high and medium respectively, and a coauthor pool of 5 distinct collaborators is recorded.
Papers in the Evolve catalog
- first author
Claim-Gated Source-Risk Auditing for Generative Search
Trust, but Validate the Instrument: Auditing AI-Generated RTL Verification Plans on Authored Security-Regression Proxies
- first author
IBBench-Light: A Paired Evaluation of Task-Conditioned Responses to External Directives
Protocol effects on feature-based hardware-Trojan detection across Trust-Hub families
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 derive from a small public bibliometric snapshot covering only the two cataloged 2026 arXiv preprints.
- No institution, country, or ORCID is attached; the profile is name-resolved from public bibliographic data and may not disambiguate same-name researchers.
- Output, impact, and activity axes are unscored because citation counts and multi-year output data are absent from the input.
- The reproducibility mean (55) rests on a single scored paper and should be read as provisional.