Webinar: Teaching Machines to See Any Object
We invite you to online lecture – Training Machines to Detect Any Object, led by RNDr. Andrej Lúčny, PhD. from the Faculty of Mathematics, Physics and Informatics at Comenius University in Bratislava.
Modern artificial intelligence models today enable robots to recognize objects from predefined categories. However, the challenge remains the ability to perceive any object as an independent entity. A major breakthrough in this area came with visual transformers, which, since 2020, have made it possible to infer the position of an object as a byproduct of the image classification process.
Speaker: RNDr. Andrej Lúčny, PhD., Faculty of Mathematics, Physics and Informatics, Comenius University in Bratislava
Date: 13 May, 2025, 10:00 CEST
Venue: online
Registration
Join us for the talk: Teaching Machines to See Any Object
Registered participants will receive an invitation to the MS Teams meeting a few days before the event.
The lecture will present:
- Využitie modelu DINO (2021), trénovaného bez anotácií, na navádzanie pozornosti humanoidného robota.
- Zlepšenie segmentácie obrazu prostredníctvom modelu CUTLER (2023) a metódy MaskCut, založenej na optimalizačnej úlohe NCut.
- New approaches to more efficient generation of fine masks using neural network learning.
The lecture is intended for researchers, professionals, students, and all those interested in artificial intelligence, computer vision, and robotics.
Accountability by Design: Turning AI Standards into Practice and Certification 12 Mar - Exploring how the EU AI Act translates harmonized standards into real-world AI testing and certification.
Discover how the EU AI Act moves from regulatory text to technical implementation — and what this means for organisations developing and deploying AI systems.
Urban buildings awaken: Slovak AI gives a second chance to underused spaces 4 Mar - Mestá sú živé organizmy, ktoré sa neustále menia. Mnohí z nás však v susedstve denne míňajú tiché svedectvá minulosti – prázdne školy, nevyužívané úrady či chátrajúce verejné budovy. Často si kladieme otázky: „Prečo je to zatvorené?“ „Nemohol by tu byť radšej denný stacionár, škôlka alebo kultúrne centrum?“
BeeGFS in Practice — Parallel File Systems for HPC, AI and Data-Intensive Workloads 6 Feb - This webinar introduces BeeGFS, a leading parallel file system designed to support demanding HPC, AI, and data-intensive workloads. Experts from ThinkParQ will explain how parallel file systems work, how BeeGFS is architected, and how it is used in practice across academic, research, and industrial environments.