Bieito
Beceiro Fernández

Research Contributions

1. B. Beceiro, J. González-Domínguez, and J. Touriño, "CUDA acceleration of feature selection methods based on conditional mutual information", Cluster Computing, vol. 29, Art. no. 280, Jun. 2026. DOI: 10.1007/s10586-026-06076-y
2. B. Beceiro, J. González-Domínguez, L. Morán-Fernández, V. Bolón-Canedo, and J. Touriño, "CUDA acceleration of MI-based feature selection methods", Journal of Parallel and Distributed Computing, vol. 190, Art. no. 104901, Aug. 2024. DOI: 10.1016/j.jpdc.2024.104901
3. B. Beceiro, J. González-Domínguez, and J. Touriño, "Acceleration of information-based feature selection methods for HPC systems", presented at the International HPC Summer School (IHPCSS'23), Atlanta, GA, USA, Jul. 2023.
   Slides: ihpcss23_slides.pdf [en]
4. B. Beceiro, J. González-Domínguez, and J. Touriño, "Biblioteca paralela de selección de características para sistemas clúster", in XXXI Jornadas de Paralelismo (JP'22), Alacant, Spain, Sep. 2022, pp. 23–32. DOI: 10.5281/zenodo.7180809
   Slides: jp22_slides.pdf [es]
5. B. Beceiro, J. González-Domínguez, and J. Touriño, "Parallel-FST: A feature selection library for multicore clusters", Journal of Parallel and Distributed Computing, vol. 169, pp. 106–116, Nov. 2022. DOI: 10.1016/j.jpdc.2022.06.012
6. B. Beceiro, J. González-Domínguez, and J. Touriño, "Acceleration of a feature selection algorithm using high performance computing", in XoveTIC Conference, A Coruña, Spain, Sep. 2020, paper 54. DOI: 10.3390/proceedings2020054054
   Slides: xovetic20_slides.pdf [gl]

Software

• 

  Parallel-FST: Feature Selection Library for Clusters

  Repository: parallel-fst
  Specific versions: jpdc22 [code, dockerhub], jpdc24 [code], cc26 [code]

  Parallel-FST is a collection of feature selection libraries. Currently, it incorporates:

  • CLI-FEAST, a CLI wrapper for the original sequential methods of FEAST.
  • MPI-FEAST, a hybrid MPI/multithreaded implementation for multicore clusters (CLI application).
  • cuFEAST, a CUDA implementation for single-GPU nodes (C++ library and CLI application.

  You can find here some GIFs that show the feature selection process when applying MIM, JMI and mRMR to the MNIST dataset.

• 

  pyCacheSimulator

  Repository: pycachesim
  PyPI package: pyCacheSimulator

  pyCacheSimulator is a Python library designed to simulate cache behavior and explore the effects of different configurations, replacement algorithms, and write policies. It allows to model multi-level cache hierarchies, track memory references, and analyze cache statistics, making it a valuable tool for both teaching and learning about computer architecture.

• bSSG

  Repository: staticsitegenerator
  PyPI package: bssg

  bSSG is a Python static site generator based on staticjinja with support for templates (Jinja), translations (pyBabel), optimized thumbnail generation (Pillow), and build size minimization by excluding unreferenced files. This webpage is created with bSSG.

Short Bio

Bieito Beceiro received the B.S. in Computer Science and the M.S. in High Performance Computing (HPC) from the Universidade da Coruña (UDC), Spain, in 2020 and 2021, respectively.

He is currently a Ph.D. student at the Computer Architecture Group of the UDC. His work is focused on the acceleration of machine learning methods using HPC techniques.