Yubo Wang — Waterloo researcher across autonomous AI research agents and large-scale retrieval systems
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-15) 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—
No publication-volume data is available yet: this axis would measure total output, venue distribution and first/corresponding-author share across the career. Only the two reviewed 2026 preprints inform this profile, so no score is given.
Impact—
Citation-based impact cannot be assessed yet: this axis would measure citations and uptake of the reviewed papers. The catalog holds two 2026 arXiv preprints without citation metadata, so no score is reported.
Focus0/100
The focus score is 0, reflecting a perfectly even field distribution: the reviewed papers split 50/50 between autonomous research agents for AI-driven science and large-scale information retrieval. No dominant field has emerged yet, so this axis reads as breadth across two distinct research communities at an early stage rather than established concentration.
Collaboration65/100
Collaboration scores 65, grounded in 25 distinct coauthors across the reviewed work. Both catalog papers are middle-author, many-hands efforts: Evolve's notes tie ScientistTwo to Google and the retrieval paper to a deployed Meta production system. The pattern suggests embedded participation inside large industrial research teams rather than small-group or solo leadership.
Activity—
Recent-activity data is insufficient: this axis would measure annual publication cadence and continuity of output over time. With only two papers, both dated 2026, no trend or productivity rate can be computed, so no score is given.
Reproducibility38/100
Reproducibility averages 38 across the 2 scored papers. ScientistTwo is rated medium: its 80.4% success rate and +25.2% average gain rest partly on AI reviewers inside its own evaluation loop. The GPU-CPU retrieval paper is rated low, since its 4.51% DCG@20 gain is a bundled production improvement without ablations.
Research summary
Yubo Wang is a researcher affiliated with the University of Waterloo whose visible record — resolved from public bibliographic data — centers on two 2026 arXiv preprints at opposite ends of the AI systems spectrum. The first, ScientistTwo, belongs to autonomous research agents for AI-driven science: Evolve's review records an 80.4% success rate improving 107 top-venue papers with a +25.2% average gain, while flagging that headline evidence leans on AI reviewers inside the system's own evaluation loop (reproducibility: medium). The second, a hybrid GPU-CPU retrieval system for personalized search at ultra-large scale, reports a deployed 4.51% DCG@20 improvement over the legacy CPU-only stack — credible production evidence, but bundled without ablations (reproducibility: low). The field distribution is an even 50/50 split between these two areas, yielding a focus score of 0: breadth rather than concentration at this stage. Collaboration is the strongest measured dimension at 65, grounded in 25 distinct coauthors, with recurring names including Jinsung Yoon, Tomas Pfister, Parthasarathy Ranganathan, Rui Meng and Yanzhou Pan; both papers are middle-author, non-corresponding team efforts. The output, impact and activity axes lack sufficient data and are deliberately left unscored. Overall this is an early-stage profile: industrially embedded, heavily collaborative, and spanning two demanding but distinct research communities.
Collaboration pattern
Collaboration is the strongest measured axis (65), grounded in 25 distinct coauthors. Both catalog papers are middle-author, non-corresponding, many-hands efforts — Evolve's notes tie one to Google (ScientistTwo) and the other to a deployed Meta production system. Recurring names in the coauthor pool include Jinsung Yoon, Tomas Pfister, Parthasarathy Ranganathan, Rui Meng and Yanzhou Pan, suggesting embedded roles inside large industrial research teams rather than small academic groups.
Research keywords
Highlights
ScientistTwo: autonomous agent improves 107 top-venue papers
Deployed GPU-CPU retrieval lifts DCG@20 by 4.51%
Timeline
- 2026Co-authored (middle position) ScientistTwo, an autonomous AI research agent; Evolve records an 80.4% success rate improving 107 top-venue papers with a +25.2% average gain.
- 2026Co-authored (middle position) the hybrid GPU-CPU retrieval paper for ultra-large-scale personalized search, reporting a deployed 4.51% DCG@20 gain over the legacy CPU-only stack.
- 2026Profile snapshot closes 2026 with two arXiv preprints split evenly (50/50) across autonomous research agents and large-scale information retrieval, with 25 distinct coauthors recorded.
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 scores are computed from OpenAlex-sourced metadata only; output, impact and activity axes currently lack sufficient data.
- The catalog covers two 2026 arXiv preprints; no peer-reviewed venue, citation or award data is included in this profile.
- ORCID and country are not provided, and 'Yubo Wang' is a common name — this profile is name-resolved from public bibliographic data.
- The reproducibility score (mean 38) rests on only 2 reviewed papers, so it is a thin signal.