Home
About
News
Partners & People
Research
Gallery
Contact
THRISS
Comenius University Bratislava
Enhancement of 3D Camera Synthetic Training Data with Noise Models
The goal of this paper is to assess the impact of noise in 3D camera-captured data by modeling the noise of the imaging process and …
Katarína Osvaldová
,
Lukáš Gajdošech
,
Viktor Kocur
,
Martin Madaras
Cite
DOI
Tuning-less Object Naming with a Foundation Model
We implement a real-time object naming system that enables learning a set of named entities never seen. Our approach employs an …
Andrej Lúčny
,
Pavel Petrovič
PDF
DOI
arXiv
Novel Synthetic Data Tool for Data-Driven Cardboard Box Localization
Application of neural networks in industrial settings, such as automated factories with bin-picking solutions requires costly …
Peter Kravár
,
Lukáš Gajdošech
,
Martin Madaras
DOI
ICANN 2023
QuasiNet: a Neural Network with Trainable Product Layers
Classical neural networks achieve only limited convergence in hard problems such as XOR or parity when the number of hidden neurons is …
Kristína Malinovská
,
Slavomír Holenda
,
Ľudovít Malinovský
DOI
ICANN 2023
Robot at the Mirror: Learning to Imitate via Associating Self-Supervised Models
We introduce an approach to building a custom model from ready-made self-supervised models via their associating instead of training …
Andrej Lúčny
,
Kristína Malinovská
,
Igor Farkaš
DOI
ICANN 2023
Safe Reinforcement Learning in a Simulated Robotic Arm
Reinforcement learning (RL) agents need to explore their environments in order to learn optimal policies. In many environments and …
Luka Kovač
,
Igor Fargaš
Cite
DOI
ICANN 2023
Controlling the Output of a Generative Model by Latent Feature Vector Shifting
State-of-the-art generative models (e.g. StyleGAN3) often generate photorealistic images based on vectors sampled from their latent …
Robert Belanec
,
Peter Lacko
,
Kristína Malinovská
Cite
DOI
DISA 2023
Merge, Explain, Iterate: A Combination of MHS and MXP in an ABox Abduction Solver
Classical neural networks achieve only limited convergence hard problems such as XOR or parity when the number of hidden neurons is …
Martin Homola
,
Júlia Pukancová
,
Janka Boborová
,
Iveta Balintová
PDF
Cite
DOI
Spike Up Prime Interest in Science and Technology Through Constructionist Games
Robotics sets have been successfully used in elementary and secondary schools in conformance with the ’learning through …
Pavel Petrovič
,
Fedor Agarshev
PDF
DOI
EDULEARN
Using Programmable Drone in Educational Projects and Competitions
The main stream of educational robotics platforms orbits the various versions of versatile robotics sets and kits, while interesting …
Pavel Petrovič
,
Peter Verčimák
PDF
DOI
EDULEARN
»
Cite
×