University of Liverpool researcher working on LLM multi-agent debate efficiency and multimodal evaluation
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-02) 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—
Publication-output data is not yet available for this profile, so the axis is unscored. It would measure total scholarly output, venue spread, and productivity over time across indexed sources, rather than the two reviewed arXiv preprints alone.
Impact—
Citation and influence data is insufficient, leaving impact unscored. This axis would measure citations, citing works, and downstream uptake of the researcher's papers; with only two 2026 preprints in scope, no reliable impact signal exists yet.
Focus0/100
Focus scores 0 on the 0-100 scale, and the raw field distribution explains why: reviewed work splits evenly, 0.5 to multimodal large language model evaluation and 0.5 to artificial intelligence (LLM multi-agent systems). The two 2026 preprints each anchor one side, so output spans two areas rather than concentrating in a single field.
Collaboration32/100
Collaboration scores 32 on the 0-100 scale. Only 4 distinct coauthors appear across the reviewed record, with Zhuoyun Li, Boxuan Wang, and Yi Dong recurring; both 2026 papers place Huang in a middle author position without corresponding-author role. The pattern suggests small-team, contributory collaboration rather than a large network or senior-led groups.
Activity—
Recent-activity data is not available, so this axis is unscored. It would measure publication cadence and research pulse over recent years; the only visible signal is two arXiv submissions in 2026, too thin to infer an activity trend.
Reproducibility38/100
Reproducibility scores 38, a mean over 2 scored papers. Repro readiness differs by paper: the sparse multi-agent debate preprint is rated medium but released no code, while the evidence-order study is rated low. Together they suggest partially documented methods with room to improve artifact availability.
Research summary
Xiaowei Huang is a researcher at the University of Liverpool whose reviewed work sits at the intersection of large language model multi-agent systems and multimodal large language model evaluation, with the Evolve field distribution split evenly, 0.5 each. The profile currently rests on two 2026 arXiv preprints, in both of which Huang appears as a middle-list author without corresponding-author role. The first, on sparse multi-agent debate, argues that distinct-peer random routing plus a round-three stopping rule rivals complex topology control, with a solid token-cost story while reported accuracy gains fall within seed noise; no code was released. The second offers a controlled empirical account of order sensitivity: presenting the same evidence in swapped order shifts vision/speech-text conflict judgments toward the image or speech modality. Bibliometric scores frame an early-stage profile: collaboration is 32 on the 0-100 scale with only 4 distinct coauthors observed; focus is 0, consistent with the even two-field split; reproducibility averages 38 across 2 scored papers whose repro readiness runs from medium (the debate paper) to low (the evidence-order paper). Output, impact, and activity axes remain unscored because their underlying data is not yet available. Frequent coauthors include Zhuoyun Li, Boxuan Wang, and Yi Dong.
Collaboration pattern
Collaboration is small-scale and team-based: 4 distinct coauthors appear across the reviewed record, with Zhuoyun Li, Boxuan Wang, and Yi Dong recurring alongside Huang. Both 2026 preprints list Huang in a middle author position without corresponding-author credit, consistent with a contributing member of a compact working group rather than the hub of a wide network; the collaboration dimension scores 32 out of 100. Note: the raw frequent-coauthor list also contains Huang's own name, likely a disambiguation artifact.
Research keywords
Highlights
Simpler routing may be enough for sparse multi-agent debate
Evidence order sways multimodal judgments
Timeline
- 2026arXiv preprint on sparse multi-agent debate: distinct-peer random routing with round-three stopping is offered as a cost-efficient alternative to complex topology control.
- 2026arXiv preprint on evidence order effects: same-evidence order swaps shift vision/speech-text conflict judgments toward image or speech.
- 2026Both reviewed works land in the same year, splitting evenly across the profile's two fields — LLM multi-agent systems and multimodal LLM evaluation — marking an early-stage, two-track record.
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.
- Profile is name-resolved from public bibliographic data; 'Xiaowei Huang' is a common Chinese name and no ORCID is provided, so identity confidence is limited.
- Bibliometric dimensions derive from public bibliometric aggregations only; citation counts, venue tiers, and funding data are out of scope.
- Output, impact, and activity axes are unscored because the underlying data is insufficient; the evidence base is only the two 2026 arXiv preprints reviewed by Evolve.
- The frequent-coauthor list includes the researcher's own name, a likely author-disambiguation artifact; collaborator counts should be read with caution.