Google researcher auditing LLM-generated RTL verification and data leakage in ML hardware-Trojan detection
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
—
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—
Not yet scored: this axis measures publication volume and output consistency over time. The current snapshot lacks the underlying productivity metrics for this researcher, so the dimension is left unscored rather than estimated.
Impact—
Not yet scored: this axis would capture citation-based influence, such as citation counts and h-index. The input data contains no citation or impact metrics, so the dimension is honestly left unscored.
Focus0/100
The focus score is 0/100, reflecting an even 50/50 field distribution between hardware security and AI for EDA (RTL verification) and hardware security (machine learning for hardware Trojan detection) across the two reviewed papers. Because the two topics are methodologically adjacent — both audit and harden security evaluation — the split reads as a coherent dual track rather than scattered work.
Collaboration44/100
Collaboration scores 44/100, based on 8 distinct coauthors in the reviewed data. A recurring core of four — Chuhong Xu, Kainan Zhou, Gangzhen Qian, and Hang Xiao — suggests a small, stable team. Lu Yi appears as last author (non-corresponding) on both 2026 papers, a pattern consistent with a senior or supervisory position within a compact group.
Activity—
Not yet scored: this axis measures research activity over time, such as publication frequency and recency. The snapshot provides only paper years (both 2026) without the volume or cadence data needed, so the dimension is left unscored.
Reproducibility55/100
Reproducibility averages 55/100 across the two reviewed papers (reproPapers: 2, both rated medium readiness). The preregistered RTL-verification audit is explicitly fail-closed — reporting its null outcome rather than discarding it — and the Trust-Hub study documents its evaluation protocol in enough detail to expose leakage, supporting the mid-range rating.
Research summary
Lu Yi is a hardware-security researcher at Google (United States) whose reviewed work concentrates on making evaluation itself more trustworthy in two adjacent areas: AI-assisted RTL verification and machine-learning-based hardware Trojan detection. The Evolve catalog holds two 2026 arXiv papers, both with Lu Yi in the last author position (non-corresponding), a placement typically associated with supervisory involvement. The first is a preregistered audit of LLM-generated RTL verification plans against authored security-regression proxies; its AI runs yielded no valid analysis, so the contribution is a fail-closed benchmark and baseline rather than model-performance evidence. The second quantifies how Trust-Hub's sibling-family structure inflates feature-based Trojan detectors, with random-forest F1 falling from 0.914 pooled to 0.460 under family-level holdout across five host families. Bibliometrically, the field distribution is an even 50/50 split between the two areas, yielding a low focus score (0/100); collaboration scores 44/100 from eight distinct coauthors, and reproducibility averages 55/100 over two papers, each rated medium readiness. Output, impact, and activity axes remain unscored for lack of data. Because Lu Yi is a common name, this profile is name-resolved from public bibliographic data.
Collaboration pattern
Lu Yi's reviewed work shows a compact collaboration pattern: 8 distinct coauthors overall, with a recurring core of four — Chuhong Xu, Kainan Zhou, Gangzhen Qian, and Hang Xiao — appearing across papers. On both 2026 arXiv papers Lu Yi is the last author but not the corresponding author, a placement consistent with a supervisory role within a small, stable hardware-security team rather than a wide networking pattern.
Research keywords
Highlights
Detector F1 drops from 0.914 to 0.460 when Trust-Hub leakage is removed
A fail-closed preregistered audit of AI-generated RTL verification plans
Timeline
- 2026Last-authored preregistered audit of AI-generated RTL verification plans posted to arXiv; its AI runs yielded no valid analysis, released as a fail-closed benchmark and baseline.
- 2026Last-authored Trust-Hub protocol study quantified sibling-family leakage, with random-forest F1 falling from 0.914 (pooled) to 0.460 (family holdout) across five host families.
- 2026Both reviewed outputs landed evenly across two adjacent hardware-security fields, producing the current 50/50 field distribution behind the 0/100 focus score.
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.
- Bibliometric scores are computed from public bibliometric data only; output, impact, and activity are unscored because underlying metrics such as citation counts and publication volume were not available in this snapshot.
- The profile rests on only two Evolve-reviewed arXiv papers (both 2026); it is an early-stage, name-resolved profile — Lu Yi is a common name and the identity match is not guaranteed.
- No citation counts, h-index, venue list, or career history were provided; nothing beyond the supplied data is asserted.
- ORCID is not linked and the affiliation (Google, United States) is taken from the input data without independent verification.