AI researcher at Hochschule Aalen: LLM agent evaluation, enterprise deployment readiness, steerable speech dialogue.
3
papers here
3
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
5
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-17) 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—
Not yet scored: the underlying output data is insufficient for a reliable measure. This axis would normally capture total works, authorship position, and venue-normalized productivity; only partial year-by-year counts are currently available.
Impact—
Not yet scored: citation impact data is judged insufficient despite early indicators (h-index 5, i10-index 3). This axis would measure citation uptake and scholarly influence, particularly for the 2026 arXiv work once citations accumulate.
Focus0/100
The focus score is 0, reflecting low topical concentration in the OpenAlex topic data: Topic Modeling and Reinforcement Learning in Robotics tie at 9 works each, followed by Recommender Systems and Techniques (5), Natural Language Processing Techniques (4), and Information Retrieval and Search Behavior (4). The 2026 work recenters on three fields — agent evaluation and enterprise deployment, conversational AI, and LLM agent reliability — each exactly one third of the field distribution.
Collaboration74/100
Collaboration scores 74, supported by 41 distinct coauthors across roughly 20 works in the last three years — a large, rotating team pattern. Recurring names include Veronica Chatrath, Bryan Zhu, Jingxuan Fan, George Pu, and Soham Dan. The cataloged 2026 papers place him in middle or last author positions, with no corresponding-author roles recorded.
Activity89/100
Activity scores 89, the strongest axis: OpenAlex records 16 works already in 2026 and 20 works over the last three years, against single-digit annual output from 2016 to 2023 (3 works in 2023, 4 in 2025). The recent surge concentrates on arXiv preprints in LLM agent evaluation and speech dialogue, indicating a very high current publication cadence.
Reproducibility55/100
Reproducibility scores 55 (the mean over 3 reviewed papers). All three 2026 catalog entries carry a medium reproducibility-readiness rating: Evolve's reviews credit solid quantitative gains — F1 rising from 0.349 to 0.498 for Continual Search and a 44.5pp steerability gain for SteerDuplex — while flagging thin baselines, single-judge scoring, and single-workflow validation as limits.
Research summary
Yunzhong He is a researcher affiliated with Hochschule Aalen (Germany), profiled from OpenAlex bibliographic records (ORCID 0000-0002-5429-5372). The publication record is small but accelerating sharply: OpenAlex year counts show 1 work each in 2016, 2017, 2020 and 2021, 3 in 2023, 4 in 2025, and 16 works already recorded for 2026, with 20 works in the last three years. Cumulative citation indicators remain early-stage (h-index 5, i10-index 3). Topic assignments are broad — Topic Modeling (9 works) and Reinforcement Learning in Robotics (9) lead, followed by Recommender Systems (5), NLP Techniques (4), and Information Retrieval and Search Behavior (4) — yielding a low focus score. The current center of gravity is clearly applied LLM-agent research: the three 2026 arXiv papers in the Evolve catalog span agent failure attribution (Continual Search), steerable full-duplex speech dialogue (SteerDuplex), and statistically grounded enterprise agent deployment (READY), each field accounting for exactly one third of the field distribution. He works in sizeable, rotating teams — 41 distinct coauthors in the recent window — and appears in middle or last author positions, with no corresponding-author roles recorded. Evolve reviews rate all three papers at medium reproducibility readiness, crediting solid quantitative gains while flagging thin baselines, single-judge scoring, and single-workflow validation as limits on certainty.
Collaboration pattern
He publishes in sizeable, rotating teams: 41 distinct coauthors appear across roughly 20 works in the last three years, and no cataloged paper records a corresponding-author role — positions are middle (Continual Search, READY) and last (SteerDuplex). Recurring collaborators include Veronica Chatrath, Bryan Zhu, Jingxuan Fan, George Pu, Soham Dan, Ryan Young, and Yuang Yao, suggesting stable clusters around the LLM-agent and speech-dialogue projects.
Research keywords
Highlights
Continual Search reframes agent root-cause attribution as iterative search
SteerDuplex improves steerability of duplex speech dialogue models
READY statistically qualifies enterprise agent deployment under oversight
Timeline
- 2016First works recorded in OpenAlex; output stays sparse, with 1 work each in 2016 and 2017.
- 2020Output remains isolated, with 1 recorded work in 2020 and 1 in 2021.
- 2023Publication activity resumes with 3 recorded works.
- 2025Output accelerates to 4 recorded works, preceding the 2026 surge.
- 2026Peak activity: 16 works recorded this year, including three cataloged arXiv papers — Continual Search on agent failure attribution, SteerDuplex on steerable duplex speech dialogue, and READY on enterprise agent deployment.
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.
- Bibliometric indicators (h-index, i10-index, per-year counts, topics) are derived from OpenAlex only; coverage of very recent 2026 arXiv work may be incomplete.
- The output and impact radar axes are unscored because the underlying data is insufficient; no citation-impact estimate is provided.
- Identity is name-resolved from public bibliographic data (OpenAlex, ORCID 0000-0002-5429-5372); biographical details beyond the affiliation are not verified.
- The reproducibility score (55) is the mean over the 3 Evolve-reviewed papers, all rated medium readiness; it is not a general-purpose reproducibility metric.