Anran Fang

Renmin University of China · China

Unclaimed · AI-structured profileOpenAlex

Explainable-AI researcher at Renmin University of China linking algorithmic interpretability with scientometrics

papers here

reviewed by Evolve

first-author

corresponding

—

works · OpenAlex

h-index · OpenAlex

Bibliometric figures come from OpenAlex (2026-10-14) 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 yet scored: this profile has insufficient publication-volume and venue data to compute an output score. The axis would measure cumulative paper count, venue placement and consistency of research production over time.

Impact—

Not yet scored: citation and uptake data are insufficient in the current snapshot. The axis would measure total citations, citation velocity and how frequently this researcher's work is referenced by later studies.

Focus0/100

Scores 0, the lowest concentration band. The field distribution splits evenly (0.5 / 0.5) between machine learning / explainable AI and human-centered explainable AI, and the topic profile spans five areas — scientometrics and bibliometrics research (4 works), XAI (2), computational and text analysis methods (2), topic modeling (1) and complex network analysis (1) — indicating breadth across adjacent areas rather than a single niche.

Collaboration32/100

Scores 32 on collaboration, reflecting a compact network: the OpenAlex snapshot lists only 4 distinct coauthors, with Lemen Chao, Ming Lei, Zixuan Yang and Mingran Sun recurring. Across the two Evolve-reviewed 2026 arXiv papers Anran Fang is last author on ADORE and middle author on the data storytelling paper, and corresponding author on neither — a contributing role within a small, stable team.

Activity47/100

Scores 47 for activity. OpenAlex records 4 works in the last three years, with countsByYear of 4 in 2023, 1 in 2024, 1 in 2025 and 2 in 2026 — output was front-loaded in 2023 and has since settled into a modest, steady cadence that includes the two 2026 arXiv preprints reviewed by Evolve.

Reproducibility55/100

Scores 55 for reproducibility, computed as a mean of 55 across 2 papers. Both Evolve-reviewed preprints — ADORE and the SHAP-to-ABT storytelling framework — carry a medium reproducibility rating; Evolve notes real efficiency gains for ADORE but flags flawed adaptivity logic and missing significance tests, and single-case, uncorrected evidence for the storytelling survey.

Research summary

Anran Fang is an early-career researcher at Renmin University of China working at the intersection of explainable artificial intelligence and scientometrics. The name-resolved OpenAlex profile shows an even 50/50 field split between machine learning / explainable AI and human-centered explainable AI, with topic counts led by scientometrics and bibliometrics research (4), followed by XAI (2) and computational and text analysis methods (2). Two 2026 arXiv preprints are reviewed in the Evolve catalog: ADORE, a unified framework for global and local interpretability built on adaptive derivative-ordered random explanation, and a data-storytelling framework that decodes AI decisions for non-experts using SHAP and ABT narrative structure. On these papers Fang appears as last author and middle author respectively, neither as corresponding author, within a small network of 4 distinct coauthors including Lemen Chao, Ming Lei, Zixuan Yang and Mingran Sun. Bibliometrics show an h-index of 4, an i10-index of 2, and 4 works over the last three years, with yearly counts of 4 (2023), 1 (2024), 1 (2025) and 2 (2026). Reproducibility readiness averages 55 across the 2 reviewed papers, both rated medium. Output and impact dimensions lack sufficient data to score.

Collaboration pattern

Collaboration is compact: only 4 distinct coauthors appear in the OpenAlex snapshot, with Lemen Chao, Ming Lei, Zixuan Yang and Mingran Sun as the recurring cluster, yielding a collaboration score of 32. On the two 2026 Evolve-reviewed arXiv papers Anran Fang is last author on ADORE and middle author on the data storytelling paper, and is corresponding author on neither — a contributing position within a small, stable team rather than an independent group lead so far.

Research keywords

explainable AI (XAI)post-hoc interpretabilityfeature attributioninterpretable machine learningdata storytellingSHAPrandomized SVDhuman-centered AI explanationscientometricsbibliometricstopic modelingcomputational text analysis

Highlights

ADORE: unifying global and local interpretability with adaptive derivative-ordered explanation
This 2026 arXiv preprint (last author) combines finite-difference derivatives, Hessian/Jacobian approximation and randomized SVD into one post-hoc attribution framework covering both global and local interpretability. Evolve's review finds the efficiency gains credible but flags flawed adaptivity logic and the absence of significance tests; reproducibility is rated medium.
Turning SHAP into data stories for non-experts without exposing sensitive details
The second 2026 arXiv preprint (middle author) proposes a SHAP-to-ABT storytelling pipeline, drawing on large language models, to explain AI decisions to non-experts while withholding sensitive data and model details. Evolve's survey reports significant gains, but it rests on single-case evidence with small, uncorrected effects; reproducibility is medium.
A dual-track profile: XAI methods on a scientometrics backbone
The OpenAlex snapshot splits the field distribution evenly (0.5 / 0.5) between machine learning / explainable AI and human-centered explainable AI, while topic counts put scientometrics and bibliometrics research first (4), ahead of XAI (2) and computational and text analysis methods (2). Both reviewed preprints sit on the XAI side of this split.

Timeline

2023The most productive year in the visible record: OpenAlex countsByYear attributes 4 works to this profile in 2023.
2024Activity slows to 1 recorded work, per the OpenAlex countsByYear data.
2025A single work is again recorded in countsByYear, continuing the lower cadence.
20262 works are recorded, both arXiv preprints reviewed by Evolve: the ADORE interpretability framework (last author) and the SHAP-to-ABT data storytelling framework (middle author).

Papers in the Evolve catalog

Frequent co-authors

Lemen Chao×2 Ming Lei×2 Zixuan Yang×1 Mingran Sun×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.

All bibliometric figures come from an OpenAlex identity resolution (identity: openalex); no ORCID is linked, so residual name ambiguity with similar-named authors cannot be fully excluded.
Citation evidence is limited to h-index 4 and i10-index 2; no raw citation counts, citation velocity or venue-quality data are available, which is why the output and impact axes are unscored.
Paper-level judgments (reproducibility mean 55, field split, verdicts) rest on only 2 Evolve-reviewed arXiv preprints, both from 2026; earlier works in countsByYear were not individually reviewed.
Field and topic distributions are assigned by OpenAlex automated classification and may mislabel work that sits between XAI and scientometrics.