ML researcher on graph structure learning and hidden-agent inference in partially observed multi-agent systems
1
papers here
1
reviewed by Evolve
1
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-09-20) 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 data is not yet available in the source record, so this axis is unscored. It would measure total and per-year publication counts, including preprints and peer-reviewed venues.
Impact—
Impact cannot be assessed: no citation counts or usage signals are present in the data. This axis would measure citations, h-index-style influence, and adoption of the methods by other groups.
Focus100/100
The focus score is 100, the maximum. It reflects a field distribution in which every reviewed paper — one in total — falls in a single field, machine learning (graph structure learning / dynamical systems). Keywords such as structural inference, neural relational inference and graph structure learning reinforce this coherent concentration; with only one record, focus should be read as current concentration rather than a long-term pattern.
Collaboration28/100
Collaboration scores 28, a low-breadth pattern consistent with the raw data: 3 distinct co-authors in total. The recurring names are Xiaoqun Wu, Mingyang Zhou and Hui Huang, and the catalog's paper is first-authored, suggesting a small, stable working group rather than a wide network. This is typical of an early-stage profile and may change quickly as more papers enter the record.
Activity—
Activity is unscored because the record contains a single 2026 paper, too little to judge cadence. It would measure ongoing output, publication regularity, and engagement across recent years.
Reproducibility55/100
Reproducibility is 55, the mean across 1 scored paper, and Evolve rates that paper's reproducibility readiness as medium. The review describes the method as sound and honest, which supports trust in reported results, while medium readiness means reimplementation would likely need extra effort or a code release. With a single data point, this score is indicative only.
Research summary
Zhongben Gong is a machine-learning researcher whose visible work centers on graph structure learning and dynamical systems, with a particular interest in inference under partial observability. The Evolve catalog currently holds one paper, the 2026 arXiv preprint 'Structural Inference under Hidden Agents', on which Gong is first author. It introduces SIHA, a method that jointly recovers hidden-agent trajectories and their interactions using an iterative, structure-guided form of neural relational inference (NRI). Evolve's review judges the work sound and honest, while noting that gains are modest once baselines receive matched supervision; the paper's reproducibility readiness is rated medium, with a reproducibility score of 55 averaged over one scored paper. Bibliometrically, the record is narrow and highly concentrated: all reviewed output falls in a single field, giving a focus score of 100, and collaboration is correspondingly compact, with 3 distinct co-authors (Xiaoqun Wu, Mingyang Zhou, Hui Huang) and a collaboration score of 28. Output, impact and activity are unscored because the underlying bibliometric data is not yet available. The profile is name-resolved from public bibliographic data (ORCID 0009-0007-3052-3186); no institution or country is recorded, so read this as an early-stage profile that will firm up as more papers are reviewed.
Collaboration pattern
The collaboration footprint is compact: 3 distinct co-authors in total, with Xiaoqun Wu, Mingyang Zhou and Hui Huang recurring most often. The single catalog paper is first-authored, consistent with a small, stable working group where Gong leads the research. The collaboration score of 28 reflects this low breadth; as the record covers only one 2026 paper, the pattern may broaden quickly.
Research keywords
Highlights
First-author paper on structural inference under hidden agents
Evolve verdict: sound and honest, with modest matched-supervision gains
Reproducibility rated medium
Timeline
- 2026Posts first-author arXiv preprint 'Structural Inference under Hidden Agents' (SIHA), introducing iterative structure-guided neural relational inference for recovering hidden agents and their interactions.
- 2026Evolve reviews the paper: verdict 'sound and honest', with supervision-matched gains described as modest; reproducibility readiness rated medium (score 55).
- 2026Profile compiled from bibliographic records: 3 distinct co-authors (Xiaoqun Wu, Mingyang Zhou, Hui Huang) and a single-field focus score of 100; output, impact and activity await more data.
Papers in the Evolve catalog
- first author
Structural Inference under Hidden Agents
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 derive from OpenAlex-based records plus the Evolve-reviewed paper set; output, impact and activity are unscored because the underlying data is insufficient.
- The name is likely not unique; this profile is name-resolved from public bibliographic data (ORCID 0009-0007-3052-3186).
- No institution or country is present in the source data, so affiliation context is unavailable.
- Reproducibility (55) and focus (100) rest on a single reviewed 2026 paper; treat them as early-stage indicators rather than established patterns.