~ /github.com/pbizopoulos/pbizopoulos.github.io
Summary
Paschalis Bizopoulos (in Greek: Πασχάλης Μπιζόπουλος) is a computer engineer with experience
in designing, developing, deploying, and maintaining reproducible AI systems using Nix-based
environments, portable tooling, and DevOps/MLOps best practices. He introduced
Sparsely Activated Networks , a method for signal decomposition, compression and pattern discovery.
Education
2014 - 2019, PhD
Electrical and Computer Engineering ,
National Technical University of Athens , Thesis: Sparsely
Activated Networks: A new method for decomposing and compressing data (in Greek)
[abstract]
[pdf]
2011 - 2012, MSc Wireless Communication Systems,
Brunel University , Awarded with distinction, Thesis
(completed at ESIEE , Paris, France): EEGArtimat, An
EEG Artifact Detector and Removal tool
2003 - 2010, BSc
Electrical & Computer Engineering ,
Aristotle University of Thessaloniki , Thesis: Soliton
Solutions of Sigma models in 1+1 and 2+1 dimensions
Selected publication
Paschalis Bizopoulos, and Dimitrios Koutsouris. "Sparsely activated networks." IEEE
transactions on neural networks and learning systems 32.3 (2020): 1304-1313.
[abstract]
[pdf]
[code]
[web application]
Research Experience
Postdoctoral researcher,
Centre for Research & Technology Hellas , Thessaloniki,
Greece
Postgraduate researcher,
National Technical University of Athens , Athens, Greece
2018 Jul - 2020 Jan,
HOLOBALANCE-Horizon 2020
2017 Sep - 2018 Mar, Doctoral dissertation visiting research student,
Harvard-MIT Biomedical Engineering Center ,
Boston, USA
2017 Jul - 2017 Aug,
HarmonicSS-Horizon 2020
2016 Jul - 2017 May,
Center for the Fight against Infarction , Rome, Italy
2015 Apr - 2016 Jun,
HEARTEN-Horizon 2020
2014 Sep - 2015 Jan,
EurHealthAgeing-FP7
Awards
Teaching
2013 - 2015, DSP on Biomedical Signals, Undergraduate laboratory course, National Technical
University of Athens
2013 - 2015, Principles of Biomedical Imaging, Undergraduate laboratory course, National
Technical University of Athens
Contact
printf "cGJpem9wQGdtYWlsLmNvbQ==" | base64 -d