Mohd. Arshad

—

Unclaimed · AI-structured profileOpenAlex

Researcher on robust ML evaluation: imbalanced learning with noisy labels and medical vision-language model audits

papers here

reviewed by Evolve

first-author

corresponding

—

works · OpenAlex

—

h-index · OpenAlex

Bibliometric figures come from OpenAlex (2026-10-17) 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 scored: the input contains no publication-count or productivity metrics for this name beyond the three arXiv preprints listed in the Evolve catalog. This axis would measure total and per-year scholarly output across venues.

Impact—

Impact is not scored: no citation counts, h-index or venue-level influence data are available in the input. This axis would measure the citation-based influence of the work, and no estimate is made here.

Focus0/100

The focus score is 0/100, reflecting a field distribution split exactly evenly between machine learning (0.5) and medical imaging machine learning with an evaluation-methodology emphasis (0.5). The three catalog papers span general imbalanced-learning methods, a broad learning system under data uncertainty, and an audit of chest X-ray TB screening models — methodologically coherent around robust evaluation, but spread across two distinct application fields.

Collaboration41/100

Collaboration scores 41/100 on a modest footprint of 7 distinct co-authors. On all three catalog preprints Arshad appears in the last author position, none as corresponding author, consistent with a senior slot on small teams. Recurring partners include Mushir Akhtar, M. Tanveer, Akarsh J., A. Varshney, A. Quadir and A. Rahaman; M. Tanveer appears twice in the frequent-co-author list, indicating repeat collaboration.

Activity—

Activity is not scored: the input lacks multi-year publication timing, so recent cadence cannot be computed. This axis would measure publishing frequency and momentum over recent years; only 2026 preprint years are present in the data.

Reproducibility55/100

Reproducibility averages 55/100 across the two papers Evolve has reviewed (reproPapers = 2), each rated medium repro-readiness. The reviewed TB-screening audit and the BARGE method paper both present quantitative protocols, but the reviews note that rankings and thresholds fail to port across cohorts and prompts in the first, and that noisy long-tailed baselines are missing in the second, keeping readiness at medium rather than high.

Research summary

Mohd. Arshad is a machine learning researcher whose recent work centers on robust, honest model evaluation — in imbalanced learning with noisy labels and in medical imaging screening. Evolve's catalog lists three 2026 arXiv preprints on each of which he appears as last author: an audit of four medical vision-language models for chest X-ray tuberculosis screening; BARGE, a bounded prior-adjusted scoring method with reliability-guided embedding for imbalanced learning with noisy labels; and a robust broad learning system with a wave loss for classification under data uncertainty. His field profile splits evenly between machine learning (0.5) and medical imaging machine learning with an evaluation-methodology emphasis (0.5), yielding a focus score of 0/100. Reproducibility averages 55/100 across the two reviewed papers, both rated medium readiness. Collaboration scores 41/100 over 7 distinct co-authors, including recurring partners Mushir Akhtar, M. Tanveer, Akarsh J., A. Varshney, A. Quadir and A. Rahaman. Output, impact and activity lack sufficient data and are not scored, and no institution, country or ORCID is available.

Collaboration pattern

Arshad's collaboration footprint in this snapshot is small: 7 distinct co-authors and a collaboration score of 41/100. On all three catalog preprints he sits in the last author position, none as corresponding author. The frequent-co-author list is dominated by one cluster — Mushir Akhtar, M. Tanveer (listed twice), Akarsh J., A. Varshney, A. Quadir and A. Rahaman — suggesting a stable in-house team rather than a wide external network.

Research keywords

chest X-raytuberculosis screeningvision-language modelszero-shot classificationprompt sensitivitynegative spectrumlong-tailed learninglabel noiseproper scoring rulesdensity-power divergencelogit adjustmentclass imbalance

Highlights

Auditing medical VLMs for chest X-ray TB screening
Last author on a 2026 arXiv audit of 4 medical vision-language models for chest X-ray tuberculosis screening. Evolve's verdict: rankings, reliability and thresholds often fail to port across cohorts and prompts; called a rigorous, honest audit (repro-readiness medium).
BARGE for imbalanced learning with noisy labels
Last author on BARGE (2026, arXiv): a bounded prior-adjusted score with reliability-guided embedding geometry for imbalanced learning with noisy labels. Evolve notes clean results trail LA, the edge under noise is modest (70.0 vs 72.3 MBE avg), and no noisy long-tailed baselines are included.
Robust broad learning system with wave loss
A 2026 arXiv preprint on a robust broad learning system with a wave loss for classification under data uncertainty, with Arshad as last author. Evolve has no review verdict, keywords or slug for this paper yet, so no quality assessment is available.

Timeline

2026Last author on the arXiv preprint 'Robust Broad Learning System with Wave Loss for Classification under Data Uncertainty' (no Evolve review yet).
2026Last author on the arXiv preprint 'Bounded Adjustment with Reliability-Guided Embedding for Imbalanced Learning with Noisy Labels' (BARGE); reviewed by Evolve with medium repro-readiness.
2026Last author on the arXiv audit 'Beyond Benchmark Scores: Auditing Medical Vision-Language Models for Chest X-Ray Tuberculosis Screening'; Evolve calls it a rigorous, honest audit.

Papers in the Evolve catalog

Frequent co-authors

Mushir Akhtar×3 M. Tanveer×2 Akarsh J.×1 A. Varshney×1 A. Quadir×1 A. Rahaman×1 M. Tanveer×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.

Radar scores are computed from OpenAlex-based bibliometric data only; no other scholarly databases are reflected in this profile.
No institution, country or ORCID is available, and 'Mohd. Arshad' is a fairly common name: this profile is name-resolved from public bibliographic data and may merge or split identities.
Output, impact and activity dimensions are unscored because the underlying bibliometric data is insufficient.
One of the three catalog papers ('Robust Broad Learning System with Wave Loss') has no Evolve review verdict, keywords, repro-readiness rating or slug; it is described from metadata only.