ML researcher at UQ: topic modeling, graph neural networks, and domain-adaptive NLP for scientific literature mining
2
papers here
1
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
13
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—
Not scored: this snapshot lacks venue-tier, per-venue volume and authorship-share data needed for the output axis. Partial counts exist (45 works in 3 years; 18 in 2026), but the composite output measure cannot be computed from them.
Impact—
Not scored: citation distributions, per-paper citation counts and field-normalized percentiles are absent, so impact cannot be computed. The snapshot's h-index (13) and i10-index (18) exist, but this axis measures normalized citation influence, which the data does not yet provide.
Focus100/100
Score 100/100 reflects strong topical concentration. OpenAlex topic counts are led by Topic Modeling (22 works), Advanced Graph Neural Networks (17), Recommender Systems and Techniques (14), Domain Adaptation and Few-Shot Learning (14), and Multimodal Machine Learning Applications (12) — a coherent machine-learning and data-mining core, extended by one cataloged work in NLP for environmental science (field distribution count 1).
Collaboration46/100
Score 46/100 with 9 distinct co-authors in the recent 3-year window — a moderate, project-driven collaboration footprint. The nine recorded frequent co-authors (Zi Huang, Yanran Tang, Mudi Zhai, Haoran Duan, Boyu Luo, Yilun Liu, Qingyun Zeng, T. David Waite, Bing-Jie Ni) span the machine-learning core and the water-treatment NLP work visible in the catalog. Both reviewed papers place Qiu in a middle author position, not corresponding author.
Activity100/100
Score 100/100. The OpenAlex counts-by-year series rises from 1 (2020) through 10 (2021), 7 (2022), 10 (2023), 13 (2024) and 14 (2025) to 18 in 2026 — the highest annual count in the snapshot — with 45 works over the recent 3-year window. Both cataloged papers, including two arXiv preprints, are dated 2026, indicating a sustained and currently accelerating output pace.
Reproducibility20/100
Score 20/100, computed as reproMean 20 over 1 assessed paper (reproPapers 1). The only Evolve-reviewed artifact, WaterBERT, is rated low reproducibility readiness; the verdict flags LLM-judge circularity and small labeled data as checks to run. Coverage is thin — one reviewed paper cannot establish a reproducibility pattern, so treat this score as indicative rather than established.
Research summary
Ruihong Qiu is a machine-learning researcher at The University of Queensland whose OpenAlex-resolved profile concentrates on topic modeling (22 works), advanced graph neural networks (17), recommender systems (14), domain adaptation and few-shot learning (14), and multimodal machine learning (12). The snapshot shows a fast-moving record: 45 works in the recent three-year window, annual counts rising from 1 in 2020 to 18 in 2026, an h-index of 13 and an i10-index of 18. The Evolve catalog holds two 2026 arXiv papers. The first introduces WaterBERT, whose SciBERT domain-adaptive pretraining on 2.97B tokens tops water-treatment BERT baselines cheaply; Evolve finds it convincing while flagging LLM-judge circularity and small labeled data, with reproducibility readiness rated low. The second, 'Call Neighbours Yourself: Graph Walks with Destination-Conditioned On-Policy Self-Distillation', extends the graph-learning line and awaits review. Collaboration scores 46 with 9 distinct recent co-authors; frequent partners include Zi Huang, Yanran Tang, Mudi Zhai, T. David Waite and Bing-Jie Ni, spanning the machine-learning and water-treatment NLP projects in the catalog. Focus scores 100 on that concentrated topic core, activity 100 on the rising counts; output and impact axes await venue and citation-distribution data, and reproducibility currently sits at 20 from a single assessed paper. The profile is name-resolved from public bibliographic records via ORCID 0000-0001-8349-6475.
Collaboration pattern
The recent 3-year window lists 9 distinct co-authors (collaboration score 46/100). Nine frequent co-authors are recorded, including Zi Huang, Yanran Tang, Mudi Zhai and Haoran Duan alongside Boyu Luo, Yilun Liu, Qingyun Zeng, T. David Waite and Bing-Jie Ni; neither reviewed paper lists Qiu as corresponding author (middle position in both). The mix is consistent with project-based teams spanning the machine-learning core and the water-treatment NLP work.
Research keywords
Highlights
WaterBERT: cheap domain-adaptive pretraining tops water-treatment BERT baselines
Destination-conditioned self-distillation for graph walks
The second 2026 arXiv catalog entry, 'Call Neighbours Yourself', proposes graph walks with destination-conditioned on-policy self-distillation, extending the Advanced Graph Neural Networks topic line (17 works). Evolve has not yet issued a verdict, keywords or reproducibility assessment for it.
Rising, concentrated research trajectory (2020–2026)
OpenAlex counts-by-year climb from 1 work (2020) to 18 (2026), with 45 works in the last three years and both cataloged papers dated 2026 — pairing the top activity score (100) with the top focus score (100).
Timeline
- 2020First year in the OpenAlex counts-by-year series: 1 recorded work, before output climbs steeply.
- 202110 works recorded — a tenfold jump from 2020's single work.
- 202310 works recorded, matching the 2021 level and resuming growth after 2022's 7.
- 202514 works recorded, then the highest annual count in the snapshot.
- 2026Peak year with 18 works; includes both Evolve-cataloged arXiv papers — the reviewed WaterBERT literature-mining study and 'Call Neighbours Yourself' on destination-conditioned self-distillation graph walks.
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 indicators (h-index 13, i10-index 18, counts by year, topics) come from an OpenAlex snapshot; no Scopus or Web of Science cross-check.
- The 'output' and 'impact' radar axes are not scored because venue-level and citation-distribution data are missing from this snapshot.
- Only two papers from the Evolve catalog were reviewed; the second lacks a verdict, keywords, reproducibility assessment and catalog slug.
- Identity is resolved from public bibliographic records (OpenAlex) linked to ORCID 0000-0001-8349-6475; disambiguation for a common name relies on that linkage.