Haoxuan Li

Tsinghua University · China

Unclaimed · AI-structured profileOpenAlex

Tsinghua researcher pairing LLM reasoning and knowledge-graph data management for Traditional Chinese Medicine

papers here

reviewed by Evolve

first-author

corresponding

—

works · OpenAlex

h-index · OpenAlex

Bibliometric figures come from OpenAlex (2026-10-08) 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 scoring is unavailable. OpenAlex currently reports an h-index of 0 and an i10-index of 0 with only 2 recent works — far too thin to score. This axis would measure publication volume and cumulative productivity over time.

Impact—

Impact scoring is unavailable: citation-based signals are absent or zero (h-index 0, i10-index 0) across the current record. This axis would measure citation uptake, scholarly influence, and the reach of the published work.

Focus0/100

The focus score of 0 mirrors the raw field distribution: output splits evenly at 0.5 / 0.5 between natural language processing (medical AI / TCM) and database systems / knowledge graph data management, and all five OpenAlex topics — from Semantic Web and Ontologies to TCM studies — carry a count of just 1. With only two 2026 papers, no topic yet dominates, so concentration cannot be established.

Collaboration44/100

Collaboration scores 44, based on 8 distinct coauthors across two 2026 arXiv papers, both held in middle-author positions without corresponding-author credit. Named frequent collaborators include Zheng Chen, Peiwu Qin, Zhicheng Du, Yuzhu Li, Zhongde Zhang, and Lianshun Jin — a pattern of steady team-based TCM projects rather than an independently led line.

Activity32/100

Activity scores 32: OpenAlex records only 2 works in the recent three-year window, both dated 2026 (countsByYear shows 2026 = 2 and no earlier years). This very thin, brand-new record fits either an early-stage researcher or incomplete indexing; pre-2026 output is absent from the data, so longer-run productivity cannot be assessed.

Reproducibility78/100

Reproducibility is the strongest measured axis at 78 (reproMean across 2 papers). TCMaster is rated high readiness — confidence-aware querying and workload-guided physical design on a 221K-node TCM knowledge graph with honest, modest gains — while the prescription generation paper is medium, its '7B beats GPT-5' claim resting on inconsistent PQS reporting and tiny margins.

Research summary

Haoxuan Li is a researcher at Tsinghua University whose 2026 output sits at the intersection of two areas: natural language processing for medical AI — specifically Traditional Chinese Medicine (TCM) — and database systems for knowledge-graph data management. OpenAlex records the field distribution as an even 0.5 / 0.5 split, with five topics each counted once. The current profile rests on two arXiv preprints, both middle-author contributions without corresponding credit. The first, TCMaster, adds confidence-aware querying — Cypher rewriting and path ranking — plus workload-guided physical design to a 221K-node multi-source TCM knowledge graph, reporting honest, modest gains; Evolve rates its reproducibility readiness high, while noting the 20pp QA win rests on a self-built benchmark. The second applies structured reasoning, chain-of-thought distillation, and DPO alignment to TCM prescription generation under medication-safety constraints such as Shi Ba Fan; Evolve cautions that its '7B outperforms GPT-5' comparison depends on inconsistent PQS reporting and very small margins. Bibliometric signals are early-stage: h-index 0, i10-index 0, 2 works in the recent three-year window, and 8 distinct coauthors. Reproducibility, at 78, is the strongest measured dimension (collaboration 44, activity 32, focus 0); output and impact remain unscored pending more evidence.

Collaboration pattern

Both 2026 papers list Haoxuan Li as a middle author without corresponding credit, and OpenAlex counts 8 distinct coauthors — a team-based pattern around Tsinghua TCM projects. Recurring collaborators include Zheng Chen, Peiwu Qin, Zhicheng Du, Yuzhu Li, Zhongde Zhang, and Lianshun Jin. The collaboration dimension scores 44, moderate: consistent joint work, but no first or corresponding authorship yet in the reviewed record.

Research keywords

TCM prescription generationLarge language modelsChain-of-thought distillationDPO alignmentMedication safetyKnowledge graphConfidence-aware queryingCypher rewritingPath rankingWorkload-guided physical designTraditional Chinese Medicine studiesKnowledge graph data management

Highlights

Confidence-aware querying on a 221K-node TCM knowledge graph
In TCMaster, confidence-aware querying (Cypher rewriting, path ranking) combines with skew-aware, workload-guided physical design over a multi-source TCM knowledge graph of 221K nodes; Evolve notes honest, modest gains, with the 20pp QA win resting on a self-built benchmark. Reproducibility readiness: high.
Structured reasoning and DPO alignment for safer TCM prescriptions
The prescription generation pipeline pairs structured reasoning and knowledge-driven alignment (chain-of-thought distillation, DPO) with medication-safety constraints such as Shi Ba Fan; Evolve cautions that the '7B outperforms GPT-5' headline rests on inconsistent PQS reporting and very small margins. Reproducibility readiness: medium.
Reproducibility is the strongest measured dimension (78/100)
Across the 2 reviewed papers, reproMean is 78: TCMaster rated high readiness and the prescription pipeline medium — the highest of the four scored radar dimensions (collaboration 44, activity 32, focus 0).
A deliberately two-sided TCM research agenda
OpenAlex fieldDist splits evenly (0.5 / 0.5) between natural language processing (medical AI / TCM) and database systems / knowledge graph data management, with five topics counted once each — modelling, reasoning, and data infrastructure for the same domain.

Timeline

2026Co-authored (middle position) the TCM prescription generation paper, combining structured reasoning, chain-of-thought distillation, and DPO alignment under medication-safety constraints such as Shi Ba Fan.
2026Co-authored (middle position) TCMaster, adding confidence-aware querying and workload-guided physical design to a 221K-node multi-source TCM knowledge graph; Evolve rates its reproducibility readiness high.
2026OpenAlex logs 2 works in the recent three-year window, all in 2026, split evenly between medical NLP and knowledge-graph data management (fieldDist 0.5 / 0.5).
2026Worked with 8 distinct coauthors across both papers, with Zheng Chen and Peiwu Qin among the most frequent collaborators; reproducibility across the 2 reviewed papers averages 78.

Papers in the Evolve catalog

Frequent co-authors

Zheng Chen×2 Haoxuan Li×2 Peiwu Qin×2 Zhicheng Du×1 Yuzhu Li×1 Zhongde Zhang×1 Lianshun Jin×1 Peiwu Qin×1

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.

Bibliometrics come from OpenAlex only; h-index and i10-index are currently 0, so the output and impact dimensions are unscored for lack of data.
No ORCID is provided and 'Haoxuan Li' is a common name, so this profile is name-resolved from public bibliographic data and identity may be ambiguous.
The profile rests on only two 2026 arXiv preprints with middle-author positions; no career history, funding, or award data is available.
The coauthor list contains duplicates ('Peiwu Qin' appears twice) and includes the profile name itself, so the distinct-coauthor count of 8 may be inflated.