Language AI researcher spanning autonomous LLM research agents and human-aware dialogue systems
2
papers here
2
reviewed by Evolve
1
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-09-20) 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—
Output is unscored: no publication counts or venue-volume data are available yet. This axis would measure total and per-year output across venues, which a two-paper arXiv snapshot cannot support.
Impact—
Impact is unscored: the input contains no citation counts, h-index, or venue-influence data. This axis would measure how widely the work is cited and taken up, and is deliberately left unestimated here.
Focus0/100
Focus scores 0 because the field distribution is an even 50/50 split: half natural language processing, half artificial intelligence (autonomous LLM research agents). Current work bridges dialogue modeling and agent planning rather than concentrating in one specialty; with only two reviewed papers, this breadth reading remains tentative and could narrow as the record grows.
Collaboration66/100
Collaboration scores 66, grounded in 26 distinct coauthors with a recurring core of twelve, including Fan Bai, Zhen Wang, Zixuan Wang, and Yufan Zhou. Roles span first-author leadership on PrimeScientist and mid-list contribution on Mind2Dialogue, indicating comfort in both leading and supporting positions within LLM-agent and language-modeling teams.
Activity—
Activity is unscored: the record spans only 2026 arXiv preprints. This axis would measure publication cadence and sustained productivity across multiple years, which a single-year snapshot cannot establish.
Reproducibility38/100
Reproducibility scores 38, the mean over two rated papers. PrimeScientist is rated medium — attempt-efficiency gains look solid, though headline reward gains rest on single runs. Mind2Dialogue is rated low — personalization gains hold across three backbones but depend on a closed teacher model. Both pipelines lean on LLM-based setups whose artifacts are not fully open, tempering replication confidence.
Research summary
Xinle Yu works across two connected areas of language AI: natural language processing and artificial intelligence for autonomous LLM research agents, with the reviewed record splitting evenly (50/50) between the two fields. The catalog holds two 2026 arXiv preprints. The first, PrimeScientist, is a first-authored paper on strategic allocation of research effort in autonomous research: budget-adaptive Monte Carlo tree search over executable plan trees decides where autonomous research effort should go. Evolve's review finds the attempt-efficiency gains solid, while noting headline reward gains rest on single runs; reproducibility readiness is medium. The second, Mind2Dialogue, is a middle-authored NLP paper that trains human-aware assistants from an Oracle simulating user mental states; personalization gains hold across three backbones but rest on a closed teacher model, rated low for reproducibility. Collaboration is the strongest measured axis (66), built on 26 distinct coauthors and a recurring core of twelve. The reproducibility score of 38 averages the two paper ratings. Output, impact, and activity metrics are not yet available, and no institution or ORCID is attached to this record.
Collaboration pattern
With 26 distinct coauthors and a recurring core of twelve — Fan Bai, Zhen Wang, Zixuan Wang, Yufan Zhou, Jinzhou Tang, Chengjun Wu, Lyumanshan Ye, Zhaoxiang Feng, Letian Peng, Adyasha Patra, Enze Ma, and Zhengding Hu — Xinle Yu works within sizeable LLM research teams. The collaboration score is 66. Roles alternate between first-author leadership (PrimeScientist) and mid-list contribution (Mind2Dialogue), suggesting both leading and supporting positions across agent and dialogue projects.
Research keywords
Highlights
PrimeScientist: budget-adaptive effort allocation for autonomous research
Mind2Dialogue: teaching assistants to sense simulated user mental states
Timeline
- 2026First-authors PrimeScientist on arXiv, proposing budget-adaptive Monte Carlo tree search over executable plan trees for allocating autonomous research effort.
- 2026Co-authors Mind2Dialogue (middle position), training human-aware language models from an Oracle that simulates user mental states.
- 2026The year's two catalogued papers split evenly between the profile's fields — autonomous LLM research agents and natural language processing.
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.
- Radar scores and coauthor counts come from public bibliometric data; coverage beyond the two reviewed arXiv preprints is limited.
- No institution, country, or ORCID is attached; this profile is name-resolved from public bibliographic records and identity may be ambiguous.
- Output, impact, and activity are unscored because the catalog holds only two 2026 arXiv preprints, with no citation, venue, or multi-year data.
- The reproducibility score (38) averages just two paper ratings, one medium and one low.