Wanli Ouyang
Chinese University of Hong Kong · Hong Kong SAR China
CUHK researcher spanning AI-for-science omics, immune world models, and self-supervised representation learning
5
papers here
4
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
0
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—
Output metrics are unavailable: h-index and i10-index are recorded as 0 with no reliable publication totals for this OpenAlex identity, so this axis is left unscored. It would measure cumulative productivity once author disambiguation improves.
Impact—
Impact cannot be assessed: citation counts, h-index, and i10-index are absent or zero in the underlying record, so no influence score is computed. This axis would measure citations and scholarly uptake once coverage for this identity matures.
Focus0/100
The focus score is 0 because the record is maximally dispersed: the field distribution assigns exactly one third each to computational biology (AI for science), machine learning for immunology, and self-supervised representation learning and world models. OpenAlex topic counts reinforce this breadth — Multimodal Machine Learning Applications (16), Topic Modeling (14), Domain Adaptation and Few-Shot Learning (6), Advanced Neural Network Applications (5), and Natural Language Processing Techniques (5) — with no single dominant line.
Collaboration81/100
Collaboration scores 73 on the strength of a wide network: 37 distinct co-authors appear alongside 38 works in the last three years, roughly one new collaborator per output, and 12 recurring collaborators head the list, including Taoyong Cui, Pheng Ann Heng, Yingcheng Wu, and Jie Ying. All three Evolve-reviewed papers place Ouyang in a middle author position without corresponding status, consistent with specialist roles inside larger team-led projects.
Activity100/100
Activity is the standout axis at 100: the record lists 38 works over the recent three-year window, and counts by year show an abrupt acceleration — one work in 2024, one in 2025, then 36 in 2026. Nearly all indexed output lands in a single year, and all three Evolve-reviewed papers are 2026 arXiv preprints, confirming that the current surge is concentrated in the present period.
Reproducibility49/100
Reproducibility averages 58 over the three papers Evolve has reviewed, and the spread is wide. The omics tool-augmented distillation paper rates high readiness with strong releases; JEPA-Anything is medium, with matched gains but unverifiable wet-lab and clinical detail; the immune world model is low, with disciplined gating but unreleased code. The mean of 58 thus reflects one openly reproducible effort offset by two partially or poorly reproducible ones.
Research summary
Wanli Ouyang is a researcher at the Chinese University of Hong Kong whose recent work sits at the intersection of machine learning methodology and AI-for-science applications. On OpenAlex-recorded evidence, the profile shows an exceptionally active recent period: 38 works in the last three years, with 36 recorded in 2026 alone, up from one per year in 2024 and 2025. The catalogued output splits evenly across three fields — computational biology (AI for science), machine learning for immunology, and self-supervised representation learning and world models — each holding about a third of the record, which is why the focus score sits at 0. Three 2026 arXiv preprints reviewed by Evolve illustrate the range: a tool-augmented on-policy distillation approach for LLM domain adaptation in omics tasks, released with high reproducibility readiness; JEPA-Anything, which adds orthogonal predictive factors to joint-embedding predictive architectures with matched gains across seven domains; and an AI-Scientist-built immune world model for multiscale immune forecasting and therapeutic hypothesis generation. Across these, Ouyang appears as a middle author without corresponding-author roles, consistent with a collaborative style: 37 distinct co-authors and 12 frequent collaborators, including Taoyong Cui and Pheng Ann Heng. Reproducibility averages 58 across the three reviewed papers, spanning high to low readiness. Citation-based output and impact metrics are not yet reliably attributable to this identity.
Collaboration pattern
The collaboration footprint is broad and team-based: 37 distinct co-authors across 38 recent works, with a stable core of 12 frequent collaborators including Taoyong Cui, Pheng Ann Heng, Yingcheng Wu, Jie Ying, Zhefan Wang, and Jinzhe Li. On all three Evolve-reviewed papers Ouyang is a middle author and never the corresponding author, suggesting a contributor role within larger group efforts rather than solo-led projects. The near one-to-one ratio of co-authors to works points to many short-lived collaborations.
Research keywords
Highlights
First omics reasoning benchmark with tool-augmented distillation
JEPA-Anything: orthogonal predictive factors across seven domains
An immune world model that forecasts immune states and proposes targets
Timeline
- 2024OpenAlex counts a single indexed work this year — the earliest activity in the available record.
- 2025Indexed output remains at one work, before the surge that follows.
- 2026Output jumps to 36 indexed works in a single year, driving the activity score to 100.
- 2026Three arXiv preprints enter the Evolve catalog — tool-augmented distillation for omics, JEPA-Anything, and an immune world model — each with Ouyang as a middle author and no corresponding role.
Papers in the Evolve catalog
Tool-Augmented On-Policy Distillation for LLM Domain Adaptation in Sequence-Based Omics Tasks
JEPA-Anything: Learning Predictive Models across Different Worlds
An immune world model for multiscale forecasting and therapeutic hypothesis generation
UniGIO: Unified Generative Global In-situ Weather Modeling from Spatiotemporal Incomplete Observations
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 are computed from OpenAlex only; h-index and i10-index are recorded as 0, suggesting an incomplete or newly split author record.
- Output and impact dimensions are unscored because citation data cannot be reliably attributed to this identity.
- Qualitative assessments rest on three Evolve-reviewed 2026 arXiv preprints, all with middle-author, non-corresponding positions.
- No ORCID is available; the profile is name-resolved from public bibliographic data, and 'Wanli Ouyang' is a name shared by multiple researchers, so identity matching carries ambiguity.