Rui Wang — bridging LLM agent evaluation and variational Monte Carlo for neural-network quantum states (2026)
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-15) 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-volume metrics are available in this snapshot, so output is unscored. This axis would measure the number of papers, yearly output, and authorship roles across the full publication record, rather than only the 2 reviewed arXiv papers in the catalog.
Impact—
No citation counts, h-index or venue-influence data are provided, so impact is unscored. This axis would measure how often the reviewed papers are cited and taken up downstream; neither paper's citation record appears in the input data.
Focus0/100
The focus score is 0, and the raw field distribution explains it: the reviewed sample splits exactly 50/50 between machine learning (LLM agent evaluation) and computational quantum many-body physics and quantum chemistry. Each of the two cataloged 2026 papers sits in a different field, so no dominant direction has emerged yet. This is a two-field profile rather than a concentrated one, based on a small two-paper sample.
Collaboration48/100
Collaboration scores 48, derived from 10 distinct coauthors in the recorded network. Both cataloged 2026 papers place Rui Wang in a middle author position without corresponding-author status, consistent with contributing within multi-author teams. The input does not map individual coauthors to specific papers, so per-project team structures and leadership patterns cannot be confirmed from this snapshot alone.
Activity—
No temporal activity data beyond the two 2026 entries is available, so activity is unscored. This axis would measure publishing cadence, year-over-year continuity, and whether output is accelerating or declining; the input only shows both papers dated 2026.
Reproducibility60/100
Reproducibility scores 60, the mean across the 2 papers with readiness ratings. The spread is wide: the quantum many-body paper on gradient-estimator design is rated high readiness, while the agentic evaluation paper is rated low, reflecting its self-judged protocol with assumed perfect recall. An average over two papers is fragile, so this score should be read as indicative rather than settled.
Research summary
Rui Wang is an early-career researcher whose reviewed work spans two distinct areas: machine learning for LLM agent evaluation, and computational quantum many-body physics and quantum chemistry. Each area accounts for exactly half (0.5) of the reviewed field distribution, drawn from two 2026 arXiv papers in the Evolve catalog. The first proposes direct and adaptive-minimum-variance gradient estimators for neural-network quantum states in variational Monte Carlo, tackling the weak-gradient trainability failures that arise with complex phase structure and spin-orbit coupling; Evolve judged the matched-state evidence strong but baselines thin, and rated reproducibility readiness high. The second introduces a skill-based framework for evaluating agents on real-time data science tasks, combining executable ground truth defined as code with factoid-level scoring and LLM-as-a-judge; Evolve flagged its 53-case, single-skill, self-judged design as thin, with low reproducibility readiness. Both papers list Rui Wang as a middle author rather than corresponding author. The recorded network holds 10 distinct coauthors, and the reproducibility mean across the two papers is 60.
Research keywords
Highlights
Gradient estimators that unblock neural-network variational Monte Carlo
Skill-based evaluation for real-time data science agents
Timeline
- 2026First cataloged paper: gradient-estimator design overcoming trainability barriers in neural-network-based variational optimization (arXiv, middle author).
- 2026Second cataloged paper: skill-based agentic evaluation for real-time data science tasks with executable ground truth and factoid scoring (arXiv, middle author).
- 2026Both 2026 papers are middle-author, non-corresponding contributions, splitting the reviewed field distribution evenly (0.5/0.5) across the two fields.
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.