Li Zhang

University of California San Diego

Unclaimed · AI-structured profileOpenAlex

ML researcher at UC San Diego working across autonomous AI agents/AutoML and EEG computational neuroscience.

papers here

reviewed by Evolve

first-author

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 snapshot lacks the underlying publication-volume data (paper counts, venues, output per year). This axis would measure how much the researcher publishes and where; the two cataloged 2026 preprints alone cannot support a score.

Impact—

Not yet scored: citation-based data (citations, h-index, citing works) is unavailable in this snapshot. This axis would measure how widely the two reviewed papers and the broader record are cited and taken up by other researchers.

Focus0/100

The focus score is 0/100 because the cataloged output splits exactly evenly between two distinct areas: the field distribution assigns 50% to machine learning (autonomous AI agents / AutoML) and 50% to computational neuroscience / EEG machine learning. With neither line dominant, the record reads as two parallel agendas rather than one concentrated specialty; more field data would be needed to judge whether this breadth is deliberate.

Collaboration48/100

The collaboration score is 48/100, computed from 10 distinct coauthors in the underlying record. The frequent-coauthor list spans both research lines and includes Peijia Qin, Ruiyi Zhang, Qi Cao, Han Guo, Pengtao Xie, Weishan Ye, Yue Pan, Gan Huang, and Zhen Liang. On both cataloged 2026 papers Li Zhang sits in a middle author position without corresponding-author status, consistent with contribution roles inside team-based projects.

Activity—

Not yet scored: year-by-year publication cadence is unavailable beyond the two 2026 preprints. This axis would measure publishing frequency and recency — output per year and time elapsed since the latest paper — which two same-year entries cannot establish.

Reproducibility55/100

The reproducibility score is 55/100, a mean computed over 2 reviewed papers (reproPapers = 2), each rated medium reproducibility readiness. Evolve's verdicts identify the gaps that cap this score: the EEG tokenization paper lacks significance tests and modern baselines and contains a Table 2 contradiction, while AIBuildAI-2.5's claimed gains are plausible but un-ablated. Releasing code, statistical tests, and component ablations would most directly raise this axis.

Research summary

Li Zhang is a machine learning researcher at the University of California San Diego whose cataloged work splits evenly between two lines: autonomous AI agents and AutoML on one side, computational neuroscience and EEG machine learning on the other. The Evolve catalog holds two 2026 arXiv preprints, in both of which Li Zhang appears as a middle author without corresponding-author role. The first, Brain-Token Learning, introduces microstate-based tokenization with multi-scale interaction for long-horizon EEG sequence modeling; Evolve notes broad reported cross-dataset gains but flags missing significance tests, absent modern baselines, and a Table 2 contradiction, rating reproducibility readiness medium. The second, AIBuildAI-2.5, reports that LLM-guided tree search with scheduling and model routing achieves state-of-the-art autonomous model building, including 73.3% MLE-Bench medals; gains are judged plausible but un-ablated. The bibliometric picture is thin: collaboration scores 48/100 on 10 distinct coauthors, focus scores 0/100 on the exact 50/50 field split, and reproducibility averages 55/100 over 2 reviewed papers, while output, impact, and activity lack sufficient underlying data to score. Frequent coauthors include Peijia Qin, Ruiyi Zhang, Qi Cao, Han Guo, and Pengtao Xie. Because Li Zhang is a very common name and no ORCID is attached, this profile is name-resolved from public bibliographic data and should be read as an early-stage, evidence-limited record.

Collaboration pattern

Li Zhang's record shows 10 distinct coauthors, giving a collaboration score of 48/100. The frequent-coauthor list — Peijia Qin, Ruiyi Zhang, Qi Cao, Han Guo, Pengtao Xie, Weishan Ye, Yue Pan, Gan Huang, and Zhen Liang — cuts across both research lines, and on both cataloged 2026 papers Li Zhang appears in a middle author position without corresponding-author role, suggesting a contributor role within larger team projects rather than solo or senior-led work.

Research keywords

autonomous AI agentsautomated machine learning (AutoML)LLM-guided tree searchmodel routingMLE-Bench evaluationAIRS-Bench evaluationEEG machine learningEEG microstatessignal tokenizationlong-horizon sequence modelingcross-dataset generalizationcomputational neuroscience

Highlights

Microstate-based tokenization for long-horizon EEG sequence modeling
Brain-Token Learning proposes microstate-based tokenization with multi-scale interaction for long-horizon EEG sequence modeling. Evolve's verdict: reported cross-dataset gains are broad, but the paper omits significance tests and modern baselines and contains a Table 2 contradiction — promising, but not yet verified. Reproducibility readiness is rated medium.
LLM-guided tree search for autonomous AI model development
AIBuildAI-2.5 reports that LLM-guided tree search combined with scheduling and model routing enables state-of-the-art autonomous AI model building, citing 73.3% MLE-Bench medals. Evolve judges the gains plausible but un-ablated, so the contribution of each component remains open; reproducibility readiness is medium.

Timeline

2026Brain-Token Learning appears on arXiv, proposing microstate-based tokenization and multi-scale interaction for long-horizon EEG sequence modeling; Li Zhang is a middle author.
2026AIBuildAI-2.5 appears on arXiv, reporting LLM-guided tree search with scheduling and model routing for autonomous AI model development; Li Zhang is again a middle author.
2026Both preprints enter the Evolve catalog with medium reproducibility readiness, yielding a reproducibility mean of 55 over 2 reviewed papers; output, impact, and activity remain unscored for lack of data.

Papers in the Evolve catalog

Frequent co-authors

Peijia Qin×1 Ruiyi Zhang×1 Qi Cao×1 Han Guo×1 Pengtao Xie×1 Weishan Ye×1 Yue Pan×1 Li Zhang×1 Gan Huang×1 Zhen Liang×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.

Radar scores and raw metrics come from OpenAlex-based bibliometric data only; citation counts, h-index, and full publication volume are not part of this snapshot.
Li Zhang is a very common name, no ORCID is attached, the Chinese-character rendering of the name is unconfirmed, and the frequent-coauthor list itself contains a same-name Li Zhang entry (likely a collision artifact) — identity rests on public bibliographic resolution alone.
The evidence base is two 2026 arXiv preprints; no peer-reviewed venues, funding, or award data are in scope, and preprint content may change before peer review.
Output, impact, and activity dimensions are unscored because their underlying bibliometric data is insufficient in this snapshot.