Machine-learning researcher at USTC working on efficient long-context LLM inference and multilingual reward modeling.
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
10
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—
Output is not yet scored: validated venue and publication-type records are incomplete, so total volume and authorship roles cannot be computed. The snapshot records an h-index of 10 and i10-index of 12 as partial signals only.
Impact—
Impact is not yet scored: per-paper citation records linked to this identity are incomplete in the snapshot, so no citation or field-weighted influence figure can be computed. This axis would measure citations, citing works, and influence within the field.
Focus0/100
The focus score is 0, and the raw field distribution explains why: output splits exactly 0.5/0.5 between efficient long-context LLM inference via sparse attention and reward modeling within machine learning/NLP, a two-track pattern rather than a single concentration. OpenAlex topic labels add further breadth — intelligent tutoring systems and adaptive learning (23), online learning and analytics (22), topic modeling (17) — so no single topic dominates the record.
Collaboration56/100
The collaboration score is 56, built on 16 distinct coauthors in the recent window, with 12 recurring partners recorded — among them Xinwei Long, Weigao Sun, Pengkun Jiao, Biqing Qi, Feida Zhu, Yiran Zhong, Steven Hoi, and Bowen Zhou. On both Evolve-reviewed papers this researcher sits in a middle author position without corresponding-author status, consistent with contributing to larger machine-learning teams rather than leading solo projects.
Activity100/100
The activity score is the maximum, 100, and the yearly counts justify it: works rise from 1 in 2019 to 1 in 2020 and 2021, then 3 in 2022, 7 in 2023, 12 in 2024, 20 in 2025, and 22 in 2026 — a steep, uninterrupted climb. The record totals 54 works in the most recent three-year window, a high cadence for a profile whose h-index is still early.
Reproducibility55/100
The reproducibility score is 55, averaged over 2 Evolve-reviewed papers, each rated medium readiness. Both carry methodological caveats from the reviews: the sparse-attention paper reports strong speed (5.03x at 128K context) but unreported variance and thin gains over the Prism baseline, while UniRRM omits significance tests and relies on LLM-verified preference data that carries circularity risk. Replication should proceed with those caveats in mind.
Research summary
Weibo Gao is a machine-learning researcher at the University of Science and Technology of China whose OpenAlex-linked record shows recent work on two complementary tracks: efficient long-context LLM inference and reward modeling for large language models. Output has climbed steeply, from one recorded work in each of 2019-2021 to 22 in 2026, totaling 54 works in the most recent three-year window, with an h-index of 10 and an i10-index of 12 on record. Sixteen distinct coauthors appear in the recent window, with twelve recurring collaborators including Xinwei Long, Weigao Sun, Biqing Qi, Yiran Zhong, Steven Hoi, and Bowen Zhou. Evolve has reviewed two 2026 arXiv preprints listing Gao as a middle author. SGDR introduces training-free pre-RoPE block-sparse attention with offline geometric priors, reporting 5.03x speed at 128K context, though gains over the Prism baseline are thin and variance is unreported. UniRRM unifies pairwise, listwise, and pointwise reward modeling across 103 languages on a 64.5K-preference dataset, with reviewers noting missing significance tests and a circularity risk from LLM-verified data. OpenAlex topic labels — including intelligent tutoring systems, online learning analytics, and topic modeling — suggest broader recorded interests, while output and impact scoring awaits fuller citation coverage.
Collaboration pattern
With 16 distinct coauthors in the recent window and a collaboration score of 56, Gao works within sizeable teams. Twelve recurring partners are recorded, including Xinwei Long, Weigao Sun, Pengkun Jiao, Biqing Qi, Feida Zhu, Yiran Zhong, Steven Hoi, and Bowen Zhou. On both reviewed 2026 preprints Gao holds a middle author position without corresponding status, suggesting a contributor role inside multi-member machine-learning groups rather than solo or corresponding-author leadership.
Research keywords
Highlights
Training-free block-sparse attention for long-context inference
Unified reasoning reward models across 103 languages
Timeline
- 2019Earliest recorded output year in the OpenAlex-linked profile, with 1 work.
- 2023Output begins to accelerate, rising to 7 recorded works from 3 in 2022.
- 202412 works recorded, continuing the upward trajectory.
- 202520 works recorded in the steepest year-over-year climb to date.
- 202622 works recorded to date; Evolve reviews the SGDR sparse-attention and UniRRM reward-modeling preprints, both with Gao as a middle author.
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 counts are derived from an OpenAlex-linked record; citation coverage is incomplete, so output and impact scores are not yet computed.
- The profile is name-resolved from public bibliographic data; OpenAlex topic labels include education-technology areas (intelligent tutoring systems, online learning analytics) alongside the recent ML focus, so possible record merging with namesakes cannot be ruled out.
- Qualitative assessments rest on only two Evolve-reviewed papers, both 2026 arXiv preprints; peer-reviewed venue status is not recorded.
- The researcher appears as a middle author on both reviewed papers; independent leadership and corresponding-author roles are not established in the available data.