Zi Xu
—
Optimization theorist focused on single-loop extragradient methods for stochastic and structured minimax optimization
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-14) 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 metrics are available yet. This axis would measure publication volume, venue quality, and productivity over time; with only two catalogued 2026 preprints and no full publication record, any score would be speculative.
Impact—
Impact data is insufficient: the snapshot contains no citation counts or other uptake signals. This axis would measure citation-based influence and how widely the minimax optimization methods are adopted across the community.
Focus0/100
Focus scores 0, derived from a field distribution split exactly 50/50 between stochastic minimax optimization / machine learning theory and numerical minimax optimization theory. Read plainly, both labels describe one agenda — minimax optimization theory — but the even split leaves no dominant field. Both catalogued 2026 papers sit on this single topic, so the low score reflects labeling breadth rather than genuinely scattered interests.
Collaboration32/100
The collaboration score is 32, based on 4 distinct coauthors across the catalogued papers — a small, closed author network. Huiling Zhang and Minhao Zhang recur as collaborators, and Zi Xu occupies the last author position on both 2026 preprints without being the corresponding author. The pattern suggests a compact, theory-oriented working group rather than distributed or cross-institution collaboration; more papers would be needed to confirm the team structure.
Activity—
No activity metrics are available — recent publication cadence and ongoing output signals are absent from the data. This axis would measure research engagement and productivity across recent years.
Reproducibility55/100
Reproducibility scores 55, the mean over 2 papers with readiness assessments. Evolve's reviews note rigorous, verified proof chains but thin empirical grounding: the stochastic nonconvex paper reports no numerical experiments, and the strongly convex paper's experiments are small-scale. Both are rated medium reproducibility readiness — theory that can in principle be re-derived and re-implemented, but not yet backed by released, validated code or benchmark runs.
Research summary
Zi Xu is a researcher in optimization theory whose catalogued work centers on minimax optimization — saddle-point problems central to machine learning and numerical computation. The Evolve catalog holds two arXiv preprints, both dated 2026, with Zi Xu in the last author position: single-loop stochastic projected damped extragradient methods for stochastic nonconvex–(strongly) concave minimax optimization, and a near-optimal pure single-loop extragradient method for strongly convex–strongly concave minimax optimization. The field distribution splits exactly 50/50 between stochastic minimax optimization / machine learning theory and numerical minimax optimization theory, signaling a tightly focused agenda on algorithmic theory: single-loop algorithm design, extragradient-type updates, variance reduction, last-iterate convergence, and Lyapunov analysis. Evolve's reviews highlight substantive theory — strong stochastic first-order oracle (SFO) bounds in the nonconvex–(strongly) concave setting, and a near-optimal √(κxκy) last-iterate rate with a verified proof chain in the strongly convex–strongly concave setting — while noting limited empirical evaluation: no numerical experiments in the stochastic paper and only small-scale experiments in the other, both rated medium reproducibility readiness. Collaboration data show 4 distinct coauthors, anchored by Huiling Zhang and Minhao Zhang. Output, impact, and activity metrics are not yet available for this profile.
Collaboration pattern
The catalogued work is a small-team effort: 4 distinct coauthors appear across the two 2026 preprints, with Huiling Zhang and Minhao Zhang as recurring collaborators. Zi Xu sits in the last author position on both papers without being the corresponding author, a pattern consistent with a senior role within a compact theory group. A collaboration score of 32 reflects a narrow network, with no cross-institution or consortium patterns visible in the current data.
Research keywords
Highlights
Near-optimal single-loop extragradient for strongly convex–strongly concave minimax
Single-loop damped extragradient for stochastic nonconvex–(strongly) concave minimax
A consistent single-loop design philosophy across problem regimes
Timeline
- 2026Preprint released: Single-Loop Stochastic Projected Damped Extragradient Methods for Stochastic Nonconvex–(Strongly) Concave Minimax Optimization (arXiv; last author)
- 2026Preprint released: Near-Optimal Pure Single-Loop Extragradient Method for Strongly Convex–Strongly Concave Minimax Optimization (arXiv; last author)
- 2026Both preprints enter the Evolve catalog with medium reproducibility readiness; the year's catalogued output splits evenly between stochastic and numerical minimax optimization theory
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.
- Bibliometrics and radar scores are computed from OpenAlex-based public bibliographic data; no citation counts or h-index were available.
- "Zi Xu" is a common name; this profile is name-resolved from public bibliographic data and may not cover every publication by identically named researchers.
- Coverage is limited to the two 2026 arXiv preprints reviewed by Evolve; earlier or venue-published work falls outside this snapshot.
- No institution, country, or ORCID was provided, so affiliation and career stage are not described.