Ahmed Ammar Kubba

University of Sharjah · United Arab Emirates

Unclaimed · AI-structured profileOpenAlex

University of Sharjah researcher using multimodal AI for liver-cancer bioinformatics and UAE architectural heritage

papers here

reviewed by Evolve

first-author

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 data is insufficient for scoring: the snapshot records 12 recent works but lacks a complete, verified publication list with venues. This axis would measure total scholarly output, publication venues, and year-over-year productivity.

Impact—

Impact cannot be scored yet: citation coverage in this snapshot is too thin, with only an early h-index of 4 and i10-index of 1 recorded. This axis would measure citation uptake, field-weighted influence, and the scholarly reach of the work.

Focus0/100

The focus score of 0 reflects an evenly spread portfolio: the OpenAlex field distribution splits exactly one third each between computer vision for architectural heritage, AI for oncology, and bioinformatics. Recorded topics — hepatocellular carcinoma treatment and prognosis, GIS studies, AI cancer detection, and gene-expression classification — confirm this breadth across two quite different domains.

Collaboration46/100

A collaboration score of 46 is based on 9 distinct coauthors over the last three years. Recurring named partners include Manar Wasif Abu Talib, Ali Bou Nassif, Qassim Nasir, Jibran Sualeh Muhammad, Jens U. Marquardt, Darko Castven, and Iman Ibrahim, a list consistent with cross-disciplinary teams spanning the computing and clinically oriented projects. Two of the three catalog papers list this researcher as first author.

Activity75/100

The activity score of 75 is supported by the year-by-year counts: 2 works in 2024, rising to 6 in 2025, and 4 already recorded in 2026, giving 12 works over the last three years. The 2026 output includes three arXiv preprints reviewed in the Evolve catalog, indicating a steep ramp-up in early-career productivity.

Reproducibility20/100

Reproducibility scores 20, the mean across the 3 papers assessed. Evolve's reviews rate all three catalog preprints as low reproducibility readiness, citing circular evaluation where labels derive from clustered features, inflated accuracy figures, data leakage, and missing per-class metrics — the key methodological issues to fix going forward.

Research summary

Ahmed Ammar Kubba is an early-career researcher at the University of Sharjah, United Arab Emirates, whose recorded work spans three fields in equal measure according to the OpenAlex field distribution: computer vision for architectural heritage, AI for oncology, and bioinformatics. The snapshot records 12 works over the last three years — 2 in 2024, 6 in 2025, and 4 in 2026 — with 9 distinct coauthors, an h-index of 4, and an i10-index of 1. Three 2026 arXiv preprints appear in the Evolve catalog, two as first author: a CLIP+SVM classifier for residential architectural styles in the UAE, and a 770-sample, five-stage hepatocellular carcinoma gene-expression dataset built with XGBoost self-training; a third, on AI-identified liver-cancer biomarkers with in-vitro DNAJB14 validation, lists the researcher as a middle author. Evolve's reviews credit the oncology work with real translational value, but rate all three preprints low on reproducibility readiness, citing circular evaluation, accuracy claims of 98% and 96.5% that are inflated, and data leakage. The profile is name-resolved from OpenAlex bibliographic data; the output and impact axes remain unscored pending fuller coverage.

Collaboration pattern

Collaboration data shows 9 distinct coauthors in the last three years, yielding a collaboration score of 46. Recurring names include Manar Wasif Abu Talib, Ali Bou Nassif, Qassim Nasir, Abdalla Sayed Ali, and Abdalla Sayed Mohamed, alongside clinically oriented partners such as Jibran Sualeh Muhammad, Jens U. Marquardt, Darko Castven, and Iman Ibrahim — a pattern consistent with cross-disciplinary teams on the heritage-computing and cancer-bioinformatics projects. Two of the three catalog papers are first-authored.

Research keywords

Multimodal learningCLIP embeddingsArchitectural style classificationCultural heritageHepatocellular carcinomaGene expressionBiomarker discoverySemi-supervised learningXGBoostExplainable AI (SHAP)Deep learning

Highlights

First-author CLIP+SVM classifier for UAE residential architectural styles
A 2026 arXiv preprint reports state-of-the-art 98% accuracy classifying residential building styles in the UAE from CLIP embeddings and an SVM. Evolve's review finds the evaluation circular: style labels are derived from clustering the same features used for classification. Reproducibility readiness is rated low.
Semi-supervised gene-expression dataset for multi-stage hepatocellular carcinoma
A first-author 2026 preprint assembles a 770-sample dataset covering five HCC stages by mapping labels and applying XGBoost self-training on TCGA and GSE89377 data. Evolve finds the reported 96.5% accuracy inflated by circular evaluation and rates readiness low.
AI-screened diagnostic and prognostic biomarkers for early-stage liver cancer
As a middle author, the researcher contributed to a 2026 preprint claiming 90.74% five-stage HCC classification with in-vitro validation of DNAJB14. Evolve sees real translational value but flags data leakage and absent per-class metrics; readiness is low.

Timeline

20242 works recorded in the OpenAlex year-by-year counts — the earliest activity in the available snapshot.
2025Annual output rises to 6 works, the highest yearly count in the recorded window.
20264 works recorded so far; three arXiv preprints enter the Evolve catalog — two as first author (CLIP+SVM architectural style classification; semi-supervised HCC biomarker dataset) and one as a middle author (AI biomarkers for early-stage liver cancer).

Papers in the Evolve catalog

Frequent co-authors

Manar Wasif Abu Talib×3 Ali Bou Nassif×2 Darko Castven×2 Jibran Sualeh Muhammad×2 Jens U. Marquardt×2 Iman Ibrahim×1 Qassim Nasir×1 Abdalla Sayed Ali×1 Abdalla Sayed Mohamed×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.

Bibliometric snapshot is OpenAlex-based (identity: openalex); this profile is name-resolved from public bibliographic data, and the name Ahmed Ammar Kubba may be shared by more than one author.
No ORCID is linked in the data, so author disambiguation relies solely on automated OpenAlex matching; totals may be incomplete or include name variants.
The output and impact dimensions are unscored because citation and full-publication coverage in this snapshot is insufficient.
Reproducibility verdicts cover only the three 2026 arXiv preprints reviewed by Evolve; no peer-reviewed journal work has been assessed.