Reinforcement learning for semantic-ID generative recommendation; middle author on two 2026 arXiv papers
2
papers here
0
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-09-27) 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—
Insufficient data: only two catalogued papers are available, so publication volume, venue mix, and productivity trends cannot be computed. This axis would measure total scholarly output across venues and years once a fuller bibliographic record is linked.
Impact—
Insufficient data: no citation counts, h-index, or influence metrics appear in the input, so scholarly impact cannot be scored. This axis would measure citations, h-index, and per-paper influence once citation records are available.
Focus—
Insufficient data: only two same-year papers in one topic area exist, so thematic breadth and long-term consistency cannot be assessed. This axis would measure how concentrated or diversified the research agenda is over a longer record.
Collaboration41/100
Scores 41 out of 100 from real bibliometric data. The snapshot records 7 distinct coauthors — Mengdan Zhu, Yufan Zhao, Sophie Di, Tao Di, Yulan Yan, Sridhar Iyer, and Liang Zhao — behind the two reviewed papers, on both of which Yao Zhao holds a middle, non-corresponding authorship position. The pattern suggests a compact, stable working group rather than a large or independently led collaboration network.
Activity—
Insufficient data: both catalogued papers are dated 2026, so publication cadence, gaps, and recent-productivity trends cannot be measured. This axis would track papers per year and recency of output as more records accumulate.
Reproducibility—
Insufficient data for a score: artifact availability evidence is not yet linked. This axis would measure code and data release and result verifiability; Evolve's per-paper reviews rate both papers' reproducibility readiness as medium.
Research summary
Yao Zhao is a researcher whose reviewed output centers on generative recommendation with semantic IDs, approached through reinforcement learning. The Evolve catalog holds two 2026 arXiv papers, on both of which Yao Zhao appears as a middle author without corresponding role. 'Learning Better Reasoning for Generative Recommendation with Semantic IDs' introduces Evo-Rec, pairing best-of-n reasoning supervised fine-tuning with NDCG-rewarded GRPO; Evolve reports gains across three Amazon datasets with solid ablations, but no seed variance and second-hand baselines. 'From Interests to Semantic IDs' adds retrieval-grounded credit assignment, improving SID recommendation on three datasets with good ablations, though significance is partial and reward leakage untested. Both papers carry a medium reproducibility-readiness rating. The collaboration dimension scores 41 out of 100 from real bibliometric data, built on 7 distinct coauthors — Mengdan Zhu, Yufan Zhao, Sophie Di, Tao Di, Yulan Yan, Sridhar Iyer, and Liang Zhao — indicating a compact, recurring team rather than an independently led network. Because 'Yao Zhao' is a common name and no institution, country, or ORCID is provided, this profile is name-resolved from public bibliographic data; output, impact, focus, activity, and reproducibility dimensions currently lack sufficient data to score.
Collaboration pattern
The two catalogued papers share a compact author team: 7 distinct coauthors appear in the snapshot, including Mengdan Zhu, Yufan Zhao, Sophie Di, Tao Di, Yulan Yan, Sridhar Iyer, and Liang Zhao. On both papers Yao Zhao is a middle author without corresponding role, consistent with contributing inside a stable, recurring group that spans reasoning and credit-assignment variants of semantic-ID generative recommendation, rather than leading an independent line or a large distributed network.
Research keywords
Highlights
Evo-Rec: Reasoning-Enhanced Generative Recommendation with Semantic IDs
Retrieval-Grounded Credit Assignment for Semantic IDs
Timeline
- 2026arXiv preprint 'Learning Better Reasoning for Generative Recommendation with Semantic IDs' appears, introducing Evo-Rec (BoN reasoning SFT plus NDCG-rewarded GRPO); it is later reviewed in the Evolve catalog with medium reproducibility readiness.
- 2026arXiv preprint 'From Interests to Semantic IDs: Retrieval-Grounded Credit Assignment for Generative Recommendation' appears, adding retrieval-grounded credit assignment; it is likewise reviewed by Evolve with medium reproducibility readiness.
- 2026Both papers enter the Evolve catalog in the same year, positioning Yao Zhao within reinforcement-learning methods for semantic-ID generative recommendation; the collaboration snapshot records 7 distinct coauthors and a collaboration score of 41/100.
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.
- Name resolution: 'Yao Zhao' is a common name; this profile is resolved from public bibliographic data and the two reviewed arXiv papers only.
- Bibliometrics come from a single snapshot; five of six radar dimensions (output, impact, focus, activity, reproducibility) lack sufficient data to score.
- No institution, country, or ORCID is provided, so affiliation history and identity verification are not possible.
- Only two 2026 arXiv papers are in the catalog; no citation counts, venue history, or career facts are available in the input data.