Shanghai Jiao Tong researcher advancing relational table learning with attention-based graph and Transformer models.
2
papers here
2
reviewed by Evolve
1
first-author
0
corresponding
—
works · OpenAlex
0
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—
Publication-output metrics for this profile are not yet available: total works, venue distribution, and yearly output beyond the two 2026 arXiv papers cannot be computed from current data, so this axis is left unscored rather than estimated.
Impact—
Citation-based impact data is insufficient: OpenAlex records an h-index and i10-index of 0 with no usable citation counts, so citation impact cannot yet be meaningfully measured or scored for this researcher.
Focus0/100
The computed focus score is 0, and the raw data shows why: OpenAlex field distribution splits evenly at 0.5/0.5 between 'machine learning (tabular and relational data learning)' and 'machine learning — graph learning on relational databases', while topic counts spread across four areas — advanced database systems and queries (2), advanced graph neural networks (2), topic modeling (1), and recommender systems and techniques (1). Both cataloged papers are nevertheless tightly concentrated on relational table learning.
Collaboration36/100
The collaboration score is 36, based on 5 distinct coauthors in the raw record. Recurring collaborators include Ken Zhong, Zheng Wang, Li Pan, and Jianhua Li, forming a small, apparently Shanghai Jiao Tong University–centered team. Both 2026 papers are multi-author efforts — Li is first author on InRTL and a middle-position author on MetaRTL, neither as corresponding author — consistent with early-career participation in a compact co-authoring group.
Activity32/100
Activity scores 32, reflecting a thin recent record: OpenAlex counts 2 works in the last three years, and countsByYear shows only 2026 with 2 papers, both arXiv preprints in relational table learning. The single-year concentration fits a researcher at the start of a publishing cycle; the score should move as later years accumulate entries.
Reproducibility55/100
Reproducibility scores 55, the mean across 2 repro-scored papers, both rated medium readiness. Evolve's reviews identify concrete gaps: MetaRTL claims state-of-the-art results across 24 tasks but reports no seed variance or runtime-baseline comparison, while InRTL's wins are razor-thin without variance reporting or significance tests. The mid-range score signals workable setups whose evidence quality, rather than availability, limits confidence.
Research summary
Weichen Li is a researcher at Shanghai Jiao Tong University whose cataloged work centers on machine learning for tabular and relational data, specifically relational table learning — applying graph neural networks and attention-based architectures to relational databases linked by primary key–foreign key relationships. The Evolve catalog contains two 2026 arXiv papers. InRTL, on which Li is first author, unifies intra- and inter-table interaction through linearized attention combined with heterogeneous graph neural networks; Evolve's review finds solid benchmarks but razor-thin margins lacking variance reporting or statistical tests. MetaRTL, a co-authored paper, introduces two-stage meta-path attention for relational table learning and reports state-of-the-art results across 24 tasks, though the review notes no seed variance or runtime-baseline comparison. OpenAlex bibliometrics record an h-index and i10-index of 0, two works in the last three years, five distinct coauthors, and an even field split (0.5/0.5) between tabular/relational data learning and graph learning on relational databases. Both papers carry medium reproducibility readiness with a mean reproducibility score of 55, reflecting missing seed variance and baseline comparisons. This is an early-stage profile based on limited bibliographic coverage.
Collaboration pattern
Li's recorded collaboration pattern is a small, stable cluster of 5 distinct coauthors, with Ken Zhong, Zheng Wang, Li Pan, and Jianhua Li recurring — a group consistent with a Shanghai Jiao Tong University lab. Li is first author on InRTL and a middle-position author on MetaRTL, and neither paper lists Li as corresponding author, matching a team-based early-career working mode rather than an independent leading role.
Research keywords
Highlights
InRTL: first-authored unification of intra- and inter-table attention
MetaRTL: two-stage meta-path attention with broad SOTA claims
Medium reproducibility readiness across both papers (mean 55)
Both cataloged papers carry a reproducibility readiness of medium, with a mean reproducibility score of 55 over 2 scored papers; the shared limitation flagged by Evolve is missing seed variance and significance testing rather than unavailability of setups.
Timeline
- 2026First-authored InRTL appears on arXiv, unifying intra- and inter-table interaction learning for relational tables via linearized attention and heterogeneous graph neural networks.
- 2026Co-authored MetaRTL appears on arXiv, introducing two-stage meta-path attention for relational table learning with state-of-the-art claims across 24 tasks.
- 2026Both 2026 papers enter the Evolve catalog with medium reproducibility readiness and a mean reproducibility score of 55; OpenAlex records 2 works for the year.
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 metrics come from OpenAlex only; no ORCID is linked, and the recorded h-index/i10-index of 0 likely reflect limited coverage of arXiv-only output.
- The output and impact axes are unscored because underlying citation and full publication data are not yet available.
- Both cataloged papers are 2026 arXiv preprints; peer-reviewed venue status is not captured in the data.
- 'Weichen Li' is a common name; this profile is name-resolved from public bibliographic data and may over- or under-merge identities.