Kvalifikační práce
Procházet dle
Poslední příspěvky
Zobrazují se záznamy 1-10 z 11389
-
Study of physical and chemical properties of tin dioxide-based systems
Výsledek obhajoby: OBHÁJENO(Univerzita Karlova, Matematicko-fyzikální fakulta, 2022)Datum obhajoby: 28. 6. 2022Title: Study of physical and chemical properties of tin dioxide-based systems Author: Oleksandr Leiko Department: Department of Surface and Plasma Science Supervisor: doc. RNDr. Karel Mašek, Dr. Abstract: In the presented ... -
A Spatiotemporal Graph Neural Network for Early Cancer Detection via Cytokine Signaling Dynamics
Výsledek obhajoby: OBHÁJENO(Univerzita Karlova, Matematicko-fyzikální fakulta, 2025)Datum obhajoby: 9. 6. 2025Title: A Spatiotemporal Graph Neural Network for Early Cancer Detection via Cytokine Signaling Dynamics Author: Robert Goldwein Department: Department of Theoretical Computer Science and Mathematical Logic Supervisor of ... -
Machine learning methods use to extract domain control knowledge in planning problems
Výsledek obhajoby: OBHÁJENO(Univerzita Karlova, Matematicko-fyzikální fakulta, 2025)Datum obhajoby: 9. 6. 2025Domain Control Knowledge (DCK) could significantly improve planning efficiency of general planners. Automatic creation of such knowledge with computers could be very helpful. In this thesis, two machine learning techniques ... -
Scalable and Interpretable Conjugate Gradient Techniques
Výsledek obhajoby: OBHÁJENO(Univerzita Karlova, Matematicko-fyzikální fakulta, 2025)Datum obhajoby: 9. 6. 2025The widespread use of neural networks and their increasing complexity necessitate effective training algorithms to optimize their performance. While second-order methods like Scaled Conjugate Gradient (SCG) offer potential ... -
Parameter-efficient unlearning of sensitive data in Large Language Models
Výsledek obhajoby: OBHÁJENO(Univerzita Karlova, Matematicko-fyzikální fakulta, 2025)Datum obhajoby: 9. 6. 2025Large language models show remarkable performance in many natural language pro- cessing tasks. However, their pre-training on web-scraped data raises ethical concerns and potentially legal issues. Due to their ability to ... -
Predicting the Popularity of Social Network Posts
Výsledek obhajoby: OBHÁJENO(Univerzita Karlova, Matematicko-fyzikální fakulta, 2025)Datum obhajoby: 9. 6. 2025 -
Segmentation of fuel rod end plugs
Výsledek obhajoby: OBHÁJENO(Univerzita Karlova, Matematicko-fyzikální fakulta, 2025)Datum obhajoby: 9. 6. 2025We present deep learning models for the automatic analysis of nuclear fuel assembly inspection images. The main task is measurement of relative rod height in milimeters. For this purpose we suggest segmentation of rod ... -
Development of Software for Defect Measurement on Threaded Inserts
Výsledek obhajoby: OBHÁJENO(Univerzita Karlova, Matematicko-fyzikální fakulta, 2025)Datum obhajoby: 9. 6. 2025This thesis addresses the development of a software pipeline for defect inspection of threaded inserts from video scans. The pipeline has two main components: the One Image Overview (OIO) generation and Depth Map (DM) ... -
Construction of phase diagrams of spin models using unsupervised machine learning
Výsledek obhajoby: OBHÁJENO(Univerzita Karlova, Matematicko-fyzikální fakulta, 2025)Datum obhajoby: 9. 6. 2025 -
Improved graph pruning for multi-agent pathfinding using heuristics
Výsledek obhajoby: OBHÁJENO(Univerzita Karlova, Matematicko-fyzikální fakulta, 2025)Datum obhajoby: 9. 6. 2025In this work, we focus on improving the solution finding process of the multi-agent pathfinding (MAPF) problem by extending graph-pruning techniques through heuristic- based selection of ground vertices, which they rely ...
RSS 1.0