Yu Ma

Output—

Output cannot be scored yet: the OpenAlex record attributes only 2 works (both 2026) with an h-index of 0, likely incomplete coverage. This axis would measure publication volume and productivity over time once fuller attribution is available.

Impact—

Impact cannot be scored yet: citation signals are absent (h-index 0, i10-index 0, no citation history). This axis would measure citation influence and uptake of the two 2026 papers once citation data accumulates.

Focus0/100

The focus score of 0 reflects a record split exactly 50/50 across two distinct fields — natural language processing / enterprise information systems and marketing science / consumer behavior (AI agents) — with all five indexed topics (topic modeling, AI ethics, AI in healthcare and education, data quality, service interactions) carrying a single work each. The profile currently reads as exploratory breadth around LLM agents rather than one concentrated line, so specialization should be expected to emerge as more works accrue.

Collaboration46/100

Collaboration scores 46, a moderate level. The record lists 9 distinct coauthors across 2 recent works, indicating team-based research on both papers; Yu Ma appears as last author on the shopping study and a middle author on UniDataAgent, neither as corresponding author. Frequent coauthors include Yutai Duan, Yahui Zhao, Zhangti Li, Zhenfeng Qi, Shaoyang Yuan, Jing Fan, Jie Liu, and Davood Wadi, though affiliations are unverified.

Activity32/100

Activity scores 32, modest. The OpenAlex counts-by-year data contains a single year, 2026, with 2 works, and the recent three-year window shows 2 works total — both arXiv preprints. This sparse, front-loaded record suggests a very recent or recently attributed research presence rather than an established multi-year publishing cadence; activity should be re-read as more years accumulate.

Reproducibility60/100

Reproducibility scores 60, the mean across 2 reviewed papers. The two are sharply split: the shopping study carries high reproducibility readiness, consistent with its credible large-scale process-tracing experimental design, while UniDataAgent is rated low — its reported 95.0% vs 72.5% (RAG) gains rest on a small, single-site, baseline-thin evaluation. Averaging an experimental paper with a deployed-system paper tempers the score; per-paper readiness differs widely.