Zhen Zhong

Research profile

Research summary

Zhen Zhong (Georgetown University, US) is profiled here as an early-stage researcher whose public record is name-resolved from OpenAlex bibliographic data, without ORCID linkage. The record is sparse: an h-index of 3 and an i10-index of 3, with per-year counts spanning 2002–2025 — 2 works in 2002, 2 in 2003, 1 in 2007, 3 in 2012, 2 in 2014, then single works in 2019, 2020 and 2025 — and only 1 work in the last three years. The Evolve catalog holds two 2026 first-authored arXiv preprints that define the current research identity: privacy-enhanced federated learning via asynchronous aggregation and local differential perturbation, and AI-assisted workflow optimization and automation for compliance processes (robotic process automation, rule engines, semantic recognition, KYC/AML). Evolve reviews find both technically coherent but rate reproducibility readiness low: the federated learning paper claims 82.6% accuracy and 21.3% communication savings without privacy proofs, ablation experiments, or rigorous statistical validation, while the automation framework's cases are illustrative and unvalidated. Collaboration is narrow — 2 distinct coauthors, with Shini Yang and Liesheng Wei the recurring partners. Notably, the OpenAlex topic footprint — scientometrics and bibliometrics, web visibility and informetrics, academic integrity — diverges from the recent machine-learning work, which may reflect record conflation or a change of research direction; output and impact remain unscored pending better data.

Research keywords

Highlights

• Asynchronous federated learning with layered differential-privacy defenses

  First-authored 2026 arXiv preprint combining differential privacy, homomorphic encryption, and local differential perturbation with asynchronous secure aggregation. Evolve's review finds the engineering approach feasible but flags the absence of privacy proofs, ablation experiments, and rigorous statistical validation behind the reported 82.6% accuracy and 21.3% communication savings.

• RPA-rule-engine-semantic framework for compliance automation

  First and corresponding author of a 2026 arXiv preprint proposing a compliance automation framework that couples robotic process automation, a rule engine, and semantic recognition with adaptive thresholds, aimed at workflows such as KYC/AML. Evolve judges the design coherent but notes all presented cases are illustrative and unvalidated.

• Long informetrics footprint in the public record

  The OpenAlex topics attributed to this name center on scientometrics and bibliometrics research (5), psychology and bibliometrics (3), and web visibility and informetrics (3), with further work on academic integrity and plagiarism (2) and physical education pedagogy (2) — a footprint that predates and diverges from the 2026 machine-learning preprints.

Timeline

1. 2002Earliest recorded works in the OpenAlex profile: 2 publications appear under this name.
2. 2012Most active recorded year, with 3 works — the annual maximum in the profile.
3. 2019Publishing resumes at a low rate: 1 recorded work, the first since 2014 (2 works).
4. 20251 recorded work — the most recent activity logged before the 2026 preprints.
5. 2026Two first-authored arXiv preprints appear — privacy-enhanced federated learning and AI-assisted compliance automation — both reviewed in the Evolve catalog.

Papers in the Evolve catalog

• 2026-09-14first author

  Privacy-enhanced federated learning via asynchronous aggregation and local differential perturbation

• 2026-09-13first authorcorresponding

  AI Assisted Workflow Optimization and Automation

Frequent co-authors