Robot learning researcher working on world-model-based manipulation and cross-embodiment generalization
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
0
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-02) 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 output score yet: this axis measures publication volume, venue depth, and authorship share, but the snapshot holds only 2 OpenAlex-indexed arXiv works (h-index 0, i10-index 0) — too thin to score, so no estimate is offered.
Impact—
No impact score yet: this axis would aggregate citations, influence on follow-up work, and community recognition, but the snapshot reports no citation data at all (h-index and i10-index both 0), so influence cannot be honestly estimated from the available evidence.
Focus0/100
Scores 0/100 on focus, indicating no concentrated line in this thin snapshot. The field distribution splits evenly (0.5/0.5) between "robotics (robot learning and embodied intelligence)" and "robot learning / robotic manipulation", and OpenAlex topics scatter across five clusters — Robot Manipulation and Learning (2), plus Soft Robotics, Action Observation and Synchronization, Embodied and Extended Cognition, and Modular Robots and Swarm Intelligence (1 each).
Collaboration51/100
Collaboration scores 51/100, mid-range. The snapshot records 12 distinct coauthors, and the frequent-coauthor list includes Nicklas Hansen, Xiaolong Wang, Dieter Fox, Henrik I. Christensen, Abhishek Gupta, Yashraj Narang, Jie Xu, Iretiayo Akinola, Bingjie Tang, Yijie Guo and Bo Ai. Both cataloged papers are multi-author efforts — a middle position on the insertion paper and the last position on the embodiment position paper — with neither flagged as corresponding author.
Activity32/100
Activity scores 32/100, reflecting a thin recent record: the OpenAlex counts-by-year histogram shows indexed works only in 2026 (2 works), and recent-3-year output is just 2 papers. This looks like an early or newly-disambiguated profile whose visible activity concentrates in a single year, so the score measures indexed coverage as much as actual research pace.
Reproducibility38/100
Reproducibility scores 38/100, the mean across the 2 papers Evolve has assessed (reproMean 38, reproPapers 2). The world-model insertion paper rates medium readiness — strong simulated evidence (56% vs 7% zero-shot) but no real-robot validation — while the cross-embodiment position paper rates low, being purely conceptual with no formalization or experiments.
Research summary
This profile is name-resolved from public OpenAlex bibliographic data. It covers a researcher working at the intersection of robot learning, manipulation, and embodied intelligence. The Evolve catalog contains two 2026 arXiv papers. The first, "Generalizable Robotic Insertion with World Models," develops a sim-only world-model approach to peg-in-hole insertion using TD-MPC2-style model-based reinforcement learning, reporting 56% zero-shot success on unseen objects versus a 7% baseline and showing clear scaling behavior, though without real-robot validation. The second, "Intelligence Across Embodiments," is a position paper with the researcher in the last author position, arguing that embodiment diversity should be treated as a scaling axis for robot intelligence and proposing a cross-embodiment transfer taxonomy; it is purely conceptual. The OpenAlex snapshot lists 2 works in 2026, 12 distinct coauthors, and a topical spread across robot manipulation and learning, soft robotics, action observation, embodied cognition, and modular robots. Radar scores are mid-range for collaboration (51), low for activity (32) and reproducibility (38), and 0 for topical focus, while output and impact cannot be scored because h-index and i10-index are both 0 in this snapshot — a signal of thin bibliometric coverage rather than measured performance. Institution, country, and ORCID are not attached, and both papers are preprints of unknown peer-review status.
Collaboration pattern
The profile shows mid-range collaboration (51/100) with 12 distinct coauthors recorded. Both cataloged papers are team efforts — a middle authorship on the world-model insertion paper and a last authorship on the embodiment position paper, neither corresponding. The frequent-coauthor list spans Nicklas Hansen, Xiaolong Wang, Dieter Fox, Henrik I. Christensen, Abhishek Gupta, Yashraj Narang, Jie Xu, Iretiayo Akinola, Bingjie Tang, Yijie Guo and Bo Ai, suggesting work within established robot-learning groups.
Research keywords
Highlights
World-model-based zero-shot insertion generalist
Embodiment diversity framed as a scaling axis
A consistent sim-to-real generalization thread
Across both 2026 papers the shared thread is generalization in robot learning: one delivers a quantitative world-model result for contact-rich insertion, the other argues conceptually that embodiment diversity is the next scaling axis. Together they sketch an agenda around zero-shot transfer and sim-to-real, though both remain preprints lacking real-robot or experimental validation.
Timeline
- 2026Posts "Generalizable Robotic Insertion with World Models" to arXiv, reporting 56% zero-shot insertion success on unseen objects versus a 7% baseline; listed in a middle author position.
- 2026Co-authors the position paper "Intelligence Across Embodiments" in the last author position, arguing that embodiment diversity is a scaling axis for robot intelligence.
- 2026OpenAlex snapshot for this year records 2 indexed works and 12 distinct coauthors, with h-index and i10-index both 0, indicating early or thin bibliometric coverage.
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.
- Identity is name-resolved from public OpenAlex bibliographic data; "Hao Su" is a common name and no ORCID, institution, or country is attached to this profile.
- All bibliometrics come from OpenAlex only; the h-index (0), i10-index (0), and the unscored output/impact axes reflect thin indexing coverage rather than measured performance.
- Both papers are 2026 arXiv preprints reviewed by Evolve; peer-review status and venue acceptance are unknown, and the reproducibility score rests on only 2 assessed papers.
- The frequent-coauthor list includes "Hao Su" itself, likely a name-disambiguation artifact; co-author counts should be read with that uncertainty.