Zhao Wei

Agency for Science, Technology and Research

Unclaimed · AI-structured profileOpenAlex

A*STAR researcher in few-shot multitask and transfer optimization, blending evolutionary methods with foundation models

papers here

reviewed by Evolve

first-author

corresponding

—

works · OpenAlex

h-index · OpenAlex

Bibliometric figures come from OpenAlex (2026-10-16) 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—

The output axis measures publication volume and research productivity over time. It is not scored here because the available OpenAlex record is too thin — yearly counts cover only 2026 — to support a reliable output estimate.

Impact—

The impact axis measures citation-based influence such as citations, h-index, and i10-index. The current record (h-index 0, i10-index 0) is too early-stage to score, so this axis is left unscored.

Focus0/100

Focus scores 0, the lowest concentration reading in this profile. OpenAlex topic tags spread across five areas — natural language processing techniques (4), topic modeling (2), advanced neural network applications (2), stochastic gradient optimization (2), and evolutionary algorithms (2) — even though the field distribution splits evenly (0.5/0.5) between machine learning (optimization and evolutionary computation) and evolutionary computation and transfer optimization, and both catalog papers converge on few-shot multitask optimization.

Collaboration39/100

Collaboration scores 39. The OpenAlex record shows 6 distinct coauthors across 6 recent works, and Evolve's co-author network lists six recurring collaborators, including Tingyang Wei, Haofeng Wu, Jiao Liu, Yew-Soon Ong, Puay Siew Tan, and Ananda Phan Iman. Both catalog papers place Zhao Wei in a middle authorship position without the corresponding-author role, consistent with a contributing member of larger optimization research groups.

Activity57/100

Activity scores 57. OpenAlex records 6 works in the most recent three-year window, with the counts-by-year data showing all 6 appearing in 2026; both Evolve-reviewed papers are 2026 arXiv preprints. This indicates an actively publishing but recent and compact record — output is concentrated in a single year rather than distributed across a longer career, which tempers the activity reading.

Reproducibility78/100

Reproducibility scores 78, the strongest dimension in this profile. Evolve assessed 2 papers with a mean reproducibility readiness of 78: the iterative sequential transfer paper rates high, while the in-context guidance paper rates medium, with the review noting that no ablation isolates the informed guidance component. The pattern suggests methods built on reproducible Bayesian-optimization and Gaussian-process machinery.

Research summary

Zhao Wei is a researcher at the Agency for Science, Technology and Research (A*STAR) whose visible work centers on optimization, particularly multitask, few-shot, and transfer optimization. The profile is name-resolved from public OpenAlex bibliographic data, which records 6 works in the most recent three-year window — all logged in 2026 — with 6 distinct coauthors. Evolve has reviewed two 2026 arXiv papers. 'In-Context Guidance: Learning Inter-Task Synergies via Numerical Foundational Models for Few-Shot Multitask Optimization' uses a frozen TabPFN to guide inter-task coupling through in-context priors, with strong reported results; 'Solving Few-Shot Multiobjective Multitask Optimization via Iterative Sequential Transfer' proposes IST, framing multitask optimization as iterative sequential transfer with Gaussian-process-kernel task prioritization. On both papers Zhao Wei is a middle author, not the corresponding author. Bibliometric signals are early-stage: h-index and i10-index are both 0, so output and impact are left unscored. Reproducibility is the strongest dimension at 78, drawn from Evolve's assessment of both papers (one rated high, one medium). Focus scores 0: topic tags scatter across five areas even though the field distribution splits evenly between machine learning (optimization and evolutionary computation) and evolutionary computation and transfer optimization.

Collaboration pattern

Collaboration scores 39. The OpenAlex record shows 6 distinct coauthors over 6 recent works, and six recurring collaborators appear in Evolve's network: Tingyang Wei, Haofeng Wu, Jiao Liu, Yew-Soon Ong, Puay Siew Tan, and Ananda Phan Iman. On both reviewed 2026 papers Zhao Wei sits in a middle authorship position without the corresponding role, suggesting a contributing member of a multi-person optimization group rather than a solo or lead-PI working mode.

Research keywords

multitask optimizationfew-shot optimizationsequential transfer optimizationmultiobjective optimizationBayesian optimizationin-context learningfoundation modelsTabPFNGaussian processesevolutionary computation

Highlights

Frozen TabPFN as in-context guidance for few-shot multitask optimization
In this 2026 arXiv paper, a frozen numerical foundational model (TabPFN) supplies in-context priors that guide inter-task coupling in few-shot multitask optimization, with strong reported results; Evolve's review flags that no ablation isolates the informed guidance component, and reproducibility readiness is rated medium.
IST: multitask optimization as iterative sequential transfer
This 2026 arXiv paper proposes IST, which treats multitask optimization as iterative sequential transfer and prioritizes source tasks with a Gaussian-process kernel; Evolve notes promising few-shot gains while observing that win counts slightly outrun the tables, and rates reproducibility readiness high.
Reproducible-by-design optimization methods
Across the two Evolve-reviewed papers the mean reproducibility readiness is 78 — this profile's strongest dimension — with the sequential-transfer paper rated high; both build on standard Bayesian-optimization and Gaussian-process machinery, which supports replication.

Timeline

2026Published 'In-Context Guidance: Learning Inter-Task Synergies via Numerical Foundational Models for Few-Shot Multitask Optimization' on arXiv, in a middle authorship position.
2026Published 'Solving Few-Shot Multiobjective Multitask Optimization via Iterative Sequential Transfer' on arXiv, also as a middle author.
2026OpenAlex logged 6 works in the recent three-year window, all in 2026, with 6 distinct coauthors — the entire visible record to date.

Papers in the Evolve catalog

Frequent co-authors

Tingyang Wei×2 Haofeng Wu×2 Jiao Liu×2 Yew-Soon Ong×2 Puay Siew Tan×1 Ananda Phan Iman×1

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.

Bibliometrics are drawn from OpenAlex only (identity: openalex); 'Zhao Wei' is a common name and no ORCID is provided, so this profile is name-resolved from public bibliographic data.
Output and impact are not scored: h-index and i10-index are 0 and yearly counts cover only 2026, which is too little to estimate productivity or citation impact.
The reproducibility score (78) rests on Evolve's assessment of only 2 papers; the researcher's wider output has not been assessed.
Only the 2 catalog papers reviewed by Evolve inform the qualitative reading; country and ORCID fields are absent from the source data.