University of Maryland researcher bridging neural approximation theory and embodied-AI simulation infrastructure
2
papers here
2
reviewed by Evolve
1
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-08) 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 volume is not yet measurable: the underlying bibliometric record lacks a publication count for this name, so this axis cannot be scored. Once indexed, it would measure the number and cadence of peer-reviewed outputs over time.
Impact—
Impact cannot be scored yet: no citation counts, h-index, or download statistics are present in the data, so there is no basis for an estimate. Once available, this axis would measure citations and uptake of the two cataloged papers.
Focus0/100
Focus scores 0 because attention is evenly split rather than concentrated: the field distribution assigns 0.5 to statistical learning theory / approximation theory (nonparametric minimax regression) and 0.5 to robotics / embodied AI (world models and neural simulation). One paper sits on each side — a first-author theory paper and a middle-author systems paper — so no single field yet dominates the catalog profile. That split may consolidate as more outputs are indexed.
Collaboration48/100
Collaboration scores 48 from a base of 10 distinct co-authors in the record. The frequent co-author list — including Zuowei Shen, Haizhao Yang, Shijun Zhang, and Wenkang Qin alongside Yukun Zhou, Jisong Cai, Dongxiao Mao, Yue Zhang, and Wei Sui — spans both research lines, though the data does not map individuals to specific papers. A first-author theory paper plus a multi-author systems paper suggests work in both small-team and larger-group settings.
Activity—
Activity is not yet scored: the data provides no publication cadence, per-year output counts, or recent-venue records beyond the two 2026 arXiv entries. This axis would measure publishing rhythm and ongoing engagement once that record exists.
Reproducibility38/100
Reproducibility readiness averages 38 across the 2 cataloged papers. The first-author theory paper rates low — its contribution is mathematical rigor (a sharp width–radius law and log-free minimax regression) rather than an artifact others can rerun, and Evolve notes the radii involved are impractical. Uranus rates medium: a strong systems stack, but an unreconciled 24 FPS claim and closed-loop policy evaluation that fine-tunes on the test tasks limit independent replication today.
Research summary
Baicheng Li is affiliated with the University of Maryland, College Park, and — on the evidence of two 2026 arXiv papers in the Evolve catalog — works along two distinct technical lines. The first is statistical learning theory and approximation theory: a first-author paper deriving optimal tradeoffs between network size and parameter magnitude in neural approximation and minimax regression, establishing a sharp width–radius law of order [N²log(eNT)]^(−β/d) and a log-free minimax regression result built on a bespoke Lipschitz 'superexpressive' activation. Evolve's review judges the mathematics rigorous while noting that the radii involved are impractical. The second line is robotics and embodied AI: as a middle author on Uranus, Li contributes to a next-generation simulation infrastructure claiming a real-time causal diffusion robot simulator; the review credits a strong systems stack but flags an unreconciled 24 FPS claim and closed-loop policy evaluation that fine-tunes on the test tasks. Radar scores place collaboration at 48 (10 distinct co-authors), reproducibility readiness at a mean of 38 across the two papers, and focus at 0, reflecting an even 50/50 split between the two fields. Output, impact, and activity metrics are not yet available in the underlying data, so this profile should be read as early-stage; both papers are arXiv preprints.
Collaboration pattern
The record lists 10 distinct co-authors, spanning both research lines: Zuowei Shen, Haizhao Yang, Shijun Zhang, Wenkang Qin, Yukun Zhou, Noah Shen, Jisong Cai, Dongxiao Mao, Yue Zhang, and Wei Sui. The data does not map individuals to specific papers, so groupings are unconfirmed. A first-author theory paper alongside a multi-author systems paper suggests a mix of focused bilateral work and larger-team collaboration.
Research keywords
Highlights
Sharp width–radius law for neural approximation and log-free minimax regression
Uranus: a real-time causal diffusion simulator for embodied AI
Timeline
- 2026Publishes a first-author arXiv paper on optimal tradeoffs between network size and parameter magnitude in neural approximation and minimax regression.
- 2026Contributes as a middle author to Uranus, a next-generation simulation infrastructure for embodied AI.
- 2026Both lines enter the Evolve catalog the same year, each with an Evolve review verdict and a reproducibility readiness rating (low and medium).
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.
- Dimension scores are computed from OpenAlex-based public bibliometric data only.
- Name resolution relies on public bibliographic data; no ORCID is provided, so identity matching for a name like Baicheng Li carries some uncertainty.
- Both cataloged papers are 2026 arXiv preprints and may not yet reflect peer review.
- Output, impact, and activity metrics are absent from the underlying data: no citation counts or h-index are available, and the Evolve catalog holds only 2 papers for this researcher.