Researcher at Dalian University of Technology working across AI-driven drug discovery, LLM agents, and multimodal NLP
3
papers here
3
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—
Output is not yet scored: no publication-count, venue-tier, or productivity metrics are available, since the three catalog entries are recent arXiv preprints. This axis would measure total publication volume, output regularity, and venue quality across the full record.
Impact—
Impact cannot be scored yet: citation counts, h-index, and downstream-usage signals are absent from this snapshot. Once available, this axis would measure citations to the reviewed papers, influence on follow-up work, and recognition within each of the three fields.
Focus0/100
The focus score is 0, the lowest possible reading, because the field distribution is perfectly even across three directions: machine learning for LLM coding agents, computational biology for AI-driven drug discovery, and multimodal NLP for content moderation each take exactly one third (0.33) of the reviewed papers. This breadth spans different benchmark cultures and toolchains, with no single dominant specialization yet.
Collaboration71/100
Collaboration scores 71 out of 100, supported by 34 distinct coauthors across the three reviewed papers. The recurring collaborator pool includes Bowen Ye, Lei Li, Shicheng Li, Zihao Yue, and Linghao Zhang, among twelve frequently listed names, suggesting stable small-team work rather than isolated authorship, although none of the three papers lists Zhao as corresponding author.
Activity—
Activity is not scored: the snapshot lacks yearly publication rhythms, submission cadence, and recency coverage beyond the three 2026 arXiv entries. This axis would measure research pace, year-over-year consistency, and current engagement across venues and preprint servers.
Reproducibility43/100
Reproducibility averages 43 across three scored papers. CodeMidas is rated low — its claimed 5,545 synthesized RL environments and MiMo-V2.5 gains come from single runs without significance tests — while CurvFlow-DTA and ProKDA are rated medium, each reporting solid protocol-matched gains on Davis/KIBA or over explain-then-detect baselines, yet all three lack statistical testing.
Research summary
Liang Zhao is a researcher at Dalian University of Technology whose reviewed output in the Evolve catalog spans three distinct areas in roughly equal measure: AI-driven drug discovery, LLM coding agents, and multimodal content moderation. In computational biology, Zhao is last-listed author on CurvFlow-DTA (2026), which applies dual-graph discrete Forman–Ricci curvature flow with pair-conditioned horizons to drug–target affinity prediction, reporting protocol-matched gains on the Davis and KIBA benchmarks; Evolve judges the results solid but the novelty incremental and the statistics untested. In machine learning, Zhao contributes to CodeMidas (2026), which claims 5,545 reinforcement-learning environments synthesized from source code alone, lifting MiMo-V2.5 on five benchmarks in a single run without significance tests. In multimodal NLP, Zhao works on ProKDA (2026), a progressive knowledge-to-decision alignment pipeline for explainable hateful meme detection that decouples knowledge, detection, and DPO stages, outperforming coupled explain-then-detect SFT. Bibliometrically, the profile shows 34 distinct coauthors (collaboration score 71), a perfectly even three-way field split (focus score 0), and a reproducibility score of 43 averaged over three papers. All three catalog entries are 2026 arXiv preprints, so output, impact, and activity metrics are not yet available; this profile is name-resolved from public bibliographic data.
Collaboration pattern
Across the three reviewed 2026 papers, Liang Zhao works with 34 distinct coauthors, with a recurring core including Bowen Ye, Lei Li, Shicheng Li, Zihao Yue, Linghao Zhang, Hanglong Lv, Yuanxin Liu, Wenhan Ma, Hao Tian, Rang Li, Jinhao Dong, and Yikai Zhao. Authorship positions vary — last author on CurvFlow-DTA and middle positions on CodeMidas and ProKDA — none as corresponding author, consistent with a senior-collaborator or contributing role in team-based projects.
Research keywords
Highlights
CurvFlow-DTA: curvature flow for drug–target affinity
CodeMidas: 5,545 RL environments synthesized from code
ProKDA: learn before you judge hateful memes
Timeline
- 2026Released the ProKDA preprint on progressive knowledge-to-decision alignment for explainable hateful meme detection, contributing as a middle-position author.
- 2026Released CodeMidas, scaling agentic coding RL environments synthesized from source code itself, with reported gains for MiMo-V2.5 on five benchmarks.
- 2026Published the CurvFlow-DTA preprint applying dual-graph discrete Forman–Ricci curvature flow to drug–target affinity prediction as last-listed 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 measures are derived from OpenAlex-based public bibliographic data only.
- No ORCID is linked; 'Liang Zhao' is a common name, so this profile is name-resolved from public bibliographic data and may merge or split identities.
- Output, impact, and activity dimensions are unscored because the only papers in scope are three 2026 arXiv preprints without citation or venue data.
- Reproducibility scores and verdicts reflect Evolve's review of only these three papers and do not represent a full career record.