Aleix Alcacer Sales
Assistant Professor · Universitat Jaume I
I work on archetypal analysis: representing data as mixtures of extreme cases.
Selected papers
- A Survey on Archetypal Analysis. With I. Epifanio, S. Mair, M. Mørup. IEEE TPAMI, 2026.
- Comparative Insights: Fuzzy Clustering Versus Archetypal Analysis in Vector Quantisation for Blosc2. With I. Epifanio. Electronics Letters, 2026.
- Biarchetype Analysis: Simultaneous Learning of Observations and Features Based on Extremes. With I. Epifanio, X. Gual-Arnau. IEEE TPAMI, 2024.
- Outlier detection of clustered functional data with image and signal processing applications by archetype analysis. With I. Epifanio. PLOS One, 2024.
- A data-driven classification of 3D foot types by archetypal shapes based on landmarks. With I. Epifanio, M. V. Ibáñez, A. Simó, A. Ballester. PLOS One, 2020.
- All publications →
Collaborations
- ironArray (2021–2024) — Knowledge-transfer agreement with Universitat Jaume I on new Blosc2 codecs. Paper: Electronics Letters 2026.
- Technical University of Denmark (2022, 2023) — Research stays on archetypal analysis for tensors. Paper: IEEE TPAMI 2026.
- Instituto de Biomecánica de Valencia (2018–2021) — Research collaboration on 3D foot shape and footwear sizing. Paper: PLOS One 2020.
Software
- archetypes (2020–present) — Author. Python library for archetypal analysis, compatible with scikit-learn.
- Blosc2 (2018–2022) — Maintainer. High-performance compressor for binary data. I developed its multidimensional array support.
Background
- Assistant Professor (2026–present)
- Postdoctoral researcher, PROMETEO project (2024–2026)
- PhD in Sciences, FPU fellowship (2020–2024)
- Software analyst at ironArray (2018–2021)
- MSc (2020) and BSc (2018) in Computational Mathematics