National competence centre for HPC @ Hack Košice 2022

#April 23 – 24, 2022

#Confidential machine learning challenge

#live in Košice

Confidential computing is a hot emerging topic, gaining significance with the increasing cloud services popularity. These tools and techniques help computational services providers to guarantee the client’s data safety while it’s being processed. Such services are typically offered by cloud computing companies or HPC (High-Performance Computing) providers.

Find out more about the challenge!

The winning team gets reMarkable tablet, Apple airpods, fitness band and more!


In the role of a service “provider” your task is to design and train suitable Deep Neural Network (DNN) and/or Gradient-Boosted Decision Trees (GBDT) model on a data set provided by us. It has to be done in such a way that an arbitrary end-user, “client”, can use this model for inference on his / her data without the provider being able to decode/decipher it. You can adapt the DNN / GBDT model, pre- and post-process the input and/or output data (both in role of provider and/or client) anyway you want, but that inference from the model has to be readable to clients, obviously. We don’t make any restriction on hardware or software components used, as long as they are available on market for a general customer. Open-source solutions are preferred.



EVALUATION criteria:

Winning solution must comply with the requirement that the “provider” has no way to make sense of “client’s” input data (i.e. must be encrypted / scrambled, etc.). Solution will rank higher on the evaluation scale if it provides better throughput (more data inferred per second) and facilitates horizontal scalability (i.e. more computer nodes / CPUs / CPU cores / accelerators other hardware components, if applicable).

WORKSHOP about HPC – why do we need it and how do we use it?
saturday april 23, 19:00 cest

Come learn more about the HPC (High-Performance Computing) and related technologies: where is it most typically used, what applications are the most perspective, how are these computational resources managed and what are the problems we face in this field. We will show you some interesting case studies, as well.

We will also touch upon the motivation and background behind our Confidential machine learning hackathon challenge.

The HPC activities in Slovakia and the role of the National competence centre for HPC will be outlined in the context of major European initiatives such as EuroHPC Joint Undertaking.


Michal Pitoňák – HPC applications specialist:

Michal studied chemical physics at the Faculty of Natural Sciences, Comenius University in Bratislava, where he earned his PhD., as well. As a scientist he focused on weak intermolecular interactions in biomolecules and parallel quantum chemistry programs development. He works at the Centre of Operations, Slovak Academy of Sciences as an HPC applications specialist.

Marián Gall – hybrid applications mentor:

Marian is a lecturer/researcher at Slovak University of Technology in Bratislava, Department of Mathematics.
His research is oriented towards calculations in quantum chemistry, development of novel quantum chemistry methods and parallel computer programs.
His recent interests also tackle the field of machine learning, neural networks and general topics on parallel computing, where he investigates mutual opportunities with HPC technologies.

Our experts and mentors are available for consultations during the hackathon and will be happy to help! Follow our latest news on the NCC webpage or social media.