ML evaluation researcher at Abdelmalek Essaâdi University: embedding geometry, invariance audits, LLM safety.
3
papers here
3
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
8
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—
The output axis — cumulative and per-year publication volume normalized against field and career-stage baselines — is not yet computed for this record, so no score is offered despite the visible yearly work counts in the snapshot.
Impact—
The impact axis — citation accumulation, normalized influence against field peers, and citation-quality signals — lacks sufficient evidence in this snapshot, so it is left unscored rather than estimated from the h-index alone.
Focus0/100
The focus score of 0 reflects extreme dispersion: the five leading topic clusters — educational games and gamification (19), multimedia learning systems (12), online learning and analytics (9), topic modeling (8), and advanced malware detection (8) — share no single core, the field distribution splits evenly (~0.33 each) across three machine-learning subfields, and the 2026 arXiv strand diverges again from the earlier topics.
Collaboration22/100
The collaboration score of 22 mirrors a very small network: only 2 distinct coauthors appear in the snapshot, with Mohammed Ahnouch the recurring partner. All three Evolve-reviewed 2026 arXiv papers place Elaachack in the last-listed position without corresponding authorship, consistent with a compact, stable two-person collaboration pattern rather than large consortium work.
Activity97/100
Activity is the profile's strongest axis at 97: the snapshot records 27 works in the most recent three years and a near-monotonic rise in annual counts, from 2 works in 2014 through 7 in 2020 and 10 in 2021 to a peak of 14 in 2025, with 6 works already logged for partial-year 2026.
Reproducibility85/100
Reproducibility is high, scoring 85 as the mean readiness across 3 assessed papers. Within the reviewed catalog, the whitening-norms and safety-drift papers are rated high readiness, while the invariance-audit paper is rated medium — its verdict notes exact theory and honest small experiments, but a linear-head, digits-only setting and no released code.
Research summary
Lotfi Elaachack (Abdelmalek Essaâdi University, Morocco; ORCID 0000-0001-8437-5800) is a machine learning researcher whose OpenAlex-recorded output spans 2014 to 2026. The snapshot records an h-index of 8 and an i10-index of 8, with 27 works in the most recent three years and clear acceleration: annual works rise from 2 in 2014 to 14 in 2025, with 6 already logged for partial-year 2026. The recorded topic mix is unusually broad — educational games and gamification (19 works), multimedia learning systems (12), online learning and analytics (9), topic modeling (8), and advanced malware detection (8) — which explains the focus score of 0. Evolve's reviewed catalog captures a more recent, theory-heavy strand: three 2026 arXiv preprints on whitened-embedding norms, transformation-invariance auditing, and cross-lingual safety drift in LLM representations, in each of which Elaachack is the last-listed author but not the corresponding author. Reproducibility signals are comparatively strong, with a mean readiness of 85 across three assessed papers. The coauthor footprint is small — 2 distinct coauthors, with Mohammed Ahnouch the recurring partner — while output and impact dimensions remain unscored pending fuller evidence.
Collaboration pattern
The collaboration footprint is small: only 2 distinct coauthors appear in the snapshot, with Mohammed Ahnouch the recurring name. Every Evolve-reviewed 2026 paper positions Elaachack as last-listed author but not corresponding author, suggesting a stable senior role within a compact two-person team rather than large-group or consortium science; no cross-institution clusters are visible in the data provided.
Research keywords
Highlights
Whitened-embedding norms exposed as Mahalanobis atypicality
Pair counts do not equal coverage in transformation audits
A linear-algebra calculus of cross-lingual safety drift
Timeline
- 2014Earliest year in the OpenAlex snapshot: 2 works recorded, the start of a low-output stretch through 2017 (1-3 works per year).
- 2020Annual works reach 7, beginning a sustained climb from the 4 works recorded in each of 2018 and 2019.
- 2021A then-peak of 10 works recorded in a single year.
- 2025Highest annual output in the snapshot: 14 works.
- 20266 works logged so far, including all three Evolve-reviewed arXiv preprints — on whitened embeddings, invariance auditing, and cross-lingual safety drift — each with Elaachack 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 figures come from an OpenAlex snapshot keyed to ORCID 0000-0001-8437-5800; identity is name-resolved from public bibliographic data.
- Output and impact dimensions are unscored because the underlying evidence in the snapshot is insufficient.
- Topic counts reflect the full OpenAlex record (ed-tech, topic modeling, malware detection), which diverges from the recent ML-theory arXiv strand in the Evolve catalog.
- 2026 counts cover only part of the year, and the reviewed papers are arXiv preprints with no peer-review status noted.