Jiaojiao Zhang

Great Bay University

Unclaimed · AI-structured profileOpenAlex

Great Bay University researcher in differentially private decentralized optimization and robust federated learning

papers here

reviewed by Evolve

first-author

corresponding

—

works · OpenAlex

—

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—

The output axis is unscored: this snapshot contains no publication-count, venue, or productivity metrics beyond the two cataloged 2026 arXiv preprints. It would measure total papers, venue quality, and annual output volume.

Impact—

No citation, h-index, or influence data is available, so impact is not scored; it would measure citations, citation velocity, and uptake of the two preprints. Neither paper yet has recorded citation evidence in this dataset.

Focus0/100

The focus score is 0, computed from an even 50/50 field distribution between machine learning (decentralized optimization and differential privacy) and machine learning (distributed/federated learning) across the two cataloged papers. The split reads as a consolidating agenda rather than scattered work: both papers address distributed learning under adversarial or privacy constraints (nonconvex convergence analysis under DP; fairness under label poisoning), and deeper publication history would likely concentrate the distribution.

Collaboration36/100

The collaboration score is 36, drawn from 5 distinct coauthors in the snapshot. Both cataloged 2026 arXiv papers are multi-author efforts in which Jiaojiao Zhang holds no corresponding-author role — last author on the private decentralized optimization paper and a middle author on FairMean — a pattern consistent with contributing within a small, stable team rather than leading large collaborations.

Activity—

Recent-activity metrics (publication cadence, preprint posting frequency over time) are unavailable, so activity is unscored. Both cataloged papers are dated 2026, but the dataset lacks the multi-year history needed to compute an activity trajectory.

Reproducibility78/100

The reproducibility score is 78, the mean readiness across 2 papers assessed by Evolve. FairMean is rated high readiness (bounded loss-reweighting with clean theory and small-scale, self-contained experiments); the private decentralized optimization paper is medium — careful theory resting on an inactive clipping assumption, with CIFAR-10 gains of 0.2–0.4 points reported without error bars.

Research summary

Jiaojiao Zhang is a researcher at Great Bay University whose cataloged work sits in distributed machine learning, split evenly (50/50) between decentralized optimization with differential privacy and distributed/federated learning. Two 2026 arXiv preprints define the current evidence base. The first, 'Private Decentralized Optimization with Noise Reduction and Bias Correction' (last author), combines the STORM recursive-momentum recursion with Exact Diffusion to achieve variance reduction and bias correction for differentially private decentralized training, backed by nonconvex convergence analysis; Evolve credits careful theory but notes the analysis relies on an inactive clipping assumption and that CIFAR-10 gains of 0.2–0.4 points are reported without error bars. The second, 'FairMean' (middle author), introduces bounded loss reweighting to preserve fairness in federated learning under label poisoning attacks on non-IID data — judged a modest but sound contribution with clean theory and small-scale experiments. Reproducibility is a relative strength: the mean readiness across 2 assessed papers is 78 (one high, one medium). Collaboration is compact — 5 distinct coauthors (score 36) — with Zhang in non-corresponding positions on both papers, consistent with a contributing role in a small team. With no citation, output-volume, or multi-year activity data available, this is an early-stage profile built from a limited bibliometric snapshot and two reviewed papers.

Collaboration pattern

Collaboration is compact: 5 distinct coauthors (score 36), including frequent collaborators Yizhao Fan, Wenjian Luo, Huigan Zheng, and Yongxiang Liu. Both 2026 arXiv papers are multi-author but place Jiaojiao Zhang in non-corresponding positions — last author on the privacy paper, middle author on FairMean — suggesting a contributing member of a small, stable team rather than an independent group leader.

Research keywords

differential privacydecentralized optimizationExact Diffusionvariance reductionrecursive momentum (STORM)nonconvex convergence analysisfederated learningfairnesslabel poisoning attackloss reweightingrobust aggregationnon-IID data

Highlights

STORM-style momentum meets Exact Diffusion under differential privacy
Last-author 2026 arXiv preprint combining the STORM recursion with Exact Diffusion for differentially private decentralized optimization, with variance reduction, bias correction, and nonconvex convergence analysis. Evolve's verdict: careful theory, but it rests on an inactive clipping assumption and the CIFAR-10 gains of 0.2–0.4 points lack error bars; reproducibility readiness is medium.
FairMean: bounded loss reweighting against label poisoning in federated learning
Middle-author 2026 arXiv preprint proposing FairMean, which uses bounded loss reweighting to preserve fairness in distributed learning under label poisoning attacks on non-IID data. Evolve judges it a modest but sound contribution — clean theory and small-scale experiments — and rates its reproducibility readiness high.

Timeline

2026Posted 'Private Decentralized Optimization with Noise Reduction and Bias Correction' on arXiv as last author — STORM recursion combined with Exact Diffusion under differential privacy.
2026Co-authored 'FairMean: Promoting Fairness in Distributed Learning under Label Poisoning Attacks' on arXiv as a middle author — bounded loss reweighting for fairness under label poisoning.
2026Both cataloged papers fall in 2026, spanning the two declared fields (decentralized optimization with DP; distributed/federated learning) at a 50/50 split, with a mean reproducibility readiness of 78.

Papers in the Evolve catalog

Frequent co-authors

Yizhao Fan×1 Wenjian Luo×1 Huigan Zheng×1 Yongxiang Liu×1 Jiaojiao Zhang×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.

Bibliometric scores come only from the provided snapshot; output, impact, and activity metrics were unavailable, so those radar axes are unscored.
No ORCID or country is provided and the name may be shared by multiple researchers; this profile is name-resolved from public bibliographic data only.
Evidence covers 2 arXiv preprints from 2026 reviewed by Evolve; peer-review venue status and any earlier publications are out of scope, so treat this as an early-stage profile.
The frequent-coauthor list includes an entry identical to the subject's own name, so the distinct-coauthor count (5) may be inflated by a name-resolution artifact.