Chuhong Xu

—

Unclaimed · AI-structured profilename-matched identity

Works across RTL hardware-security verification and retrieval-augmented generation, auditing AI-generated outputs

papers here

reviewed by Evolve

first-author

corresponding

—

works · OpenAlex

—

h-index · OpenAlex

Bibliometric figures come from OpenAlex (2026-10-14) 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 yet scored: the underlying bibliometric source has no complete publication-volume record for this name. This axis would measure total papers, venue types, and the distribution of author positions across the full career record.

Impact—

Impact is not yet scored: citation-side data (citation counts, h-index, field-normalized percentiles) is unavailable for this profile. This axis would measure how much later work builds on these publications.

Focus0/100

Focus scores 0, the lowest value, because the field distribution is an even split: hardware security and AI for EDA (RTL verification) at 0.5 and retrieval-augmented generation (information retrieval / NLP) at 0.5. With only two catalog papers, this reads as a deliberate two-track portfolio rather than accumulated specialization; further output in either track would clarify where the center of gravity lies.

Collaboration44/100

Collaboration scores 44, computed from 8 distinct coauthors in the bibliometric record. Repeated list entries point to recurring partners — Kainan Zhou, Gangzhen Qian, and Lu Yi appear more than once, with Hang Xiao also present. On both catalog papers this researcher sits in a middle author position without corresponding-author status, consistent with contributing within a team rather than leading or working solo.

Activity—

Activity is not yet scored: the underlying record lacks a complete year-by-year publication series, so recent output velocity cannot be computed. This axis would measure publishing cadence and momentum across recent years.

Reproducibility38/100

Reproducibility averages 38 across the 2 scored papers. The RTL-verification audit reaches medium readiness: its preregistered design and authored security-regression proxies persist as a benchmark plus baseline even though the AI runs yielded no valid analysis. The RAG source-risk triage paper sits at low readiness, its evidence limited to a single-coded pilot plus a synthetic scenario. The gap between method transparency and empirical evidence underlies the moderate score.

Research summary

Chuhong Xu is an early-stage researcher whose visible record, as captured in Evolve's catalog, consists of two 2026 arXiv preprints in two distant fields: a preregistered audit of AI-generated RTL verification plans in hardware security and AI for EDA, and a query-aware source-risk triage scheme for retrieval-augmented generation in information retrieval / NLP. The bibliometric field distribution splits exactly 50/50 between the two, which drives the focus score of 0 and frames the work as a deliberate two-track portfolio rather than a settled specialization. Both papers list the researcher as a middle author without corresponding status, and the collaboration score of 44 rests on 8 distinct coauthors, with repeated entries suggesting recurring partners such as Kainan Zhou, Gangzhen Qian, and Lu Yi. Evolve's reviews point to a consistent methodological instinct: treat AI-generated artifacts as claims to be validated. The RTL paper is judged a valuable fail-closed benchmark and baseline even though its AI runs yielded no valid analysis (repro readiness medium); the RAG paper offers auditable scores but rests on a single-coded pilot plus a synthetic scenario (repro readiness low). Reproducibility averages 38 across the 2 scored papers. Output, impact, and activity remain unscored for lack of underlying data and are treated as open axes rather than estimates.

Collaboration pattern

Collaboration is team-based: the record shows 8 distinct coauthors, with repeated entries pointing to ongoing partnerships with Kainan Zhou, Gangzhen Qian, and Lu Yi (Hang Xiao appears once). On both reviewed papers Chuhong Xu sits in a middle author position without corresponding status, consistent with a contributor role in a group running hardware-security RTL verification and retrieval-augmented generation projects in parallel. The coauthor list itself contains 'Chuhong Xu', likely a self-matching artifact.

Research keywords

RTL verificationLLM-generated testbenchesHardware securityMutation testingCWE (Common Weakness Enumeration)PreregistrationRetrieval-augmented generationSource provenanceQuery-aware triageQuery clusteringRank-discounted aggregationSource-side bias

Highlights

Fail-closed benchmark from a preregistered LLM audit of RTL verification plans
Evolve's review notes that this 2026 preregistered audit of AI-generated RTL verification plans yielded no valid analysis from its AI runs and no model-performance evidence, yet the authored security-regression proxies survive as a reusable benchmark plus baseline; reproducibility readiness is rated medium.
Query-aware source-risk triage for RAG with auditable scores
This 2026 preprint proposes triaging retrieval sources by per-query risk, with query clustering and rank-discounted aggregation behind auditable scores and explicit attention to source-side bias; Evolve credits its honesty about limits while rating the evidence as a single-coded pilot plus synthetic scenario only, with low reproducibility readiness.
One methodological instinct across two distant fields: validate AI-generated artifacts
Both 2026 preprints treat outputs of AI systems — LLM-drafted RTL verification plans and RAG source choices — as claims to be audited rather than trusted; the 0.5/0.5 bibliometric field split quantifies this two-track spread, which the focus score of 0 reflects.

Timeline

2026Posts 'Trust, but Validate the Instrument' to arXiv — a preregistered audit of AI-generated RTL verification plans on authored security-regression proxies (middle author).
2026Posts 'Query-Aware Source-Risk Triage for Retrieval-Augmented Generation' to arXiv, introducing auditable per-query source-risk scores for RAG (middle author).
2026The visible record stands at two arXiv preprints in 2026, split evenly between RTL hardware-security verification and retrieval-augmented generation (0.5/0.5 field distribution).

Papers in the Evolve catalog

Frequent co-authors

Kainan Zhou×2 Gangzhen Qian×2 Hang Xiao×1 Chuhong Xu×1 Kainan Zhou×1 Gangzhen Qian×1 Lu Yi×1 Lu Yi×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 scores are computed from public bibliometric data (OpenAlex-based) only; no institutional, country, or ORCID-verified identity data was available for this profile.
'Chuhong Xu' is a relatively common Chinese name; this profile is name-resolved from public bibliographic data and may mix or omit records.
Only the two Evolve-reviewed 2026 arXiv preprints were assessed in depth; citation counts, h-index, and the full publication history are out of scope.
The frequent-coauthor list contains 'Chuhong Xu' itself, likely a self-matching artifact of the name-resolution step.