03610ntm a22005177i 4500 000701263 CZ-PrVSE 20231021093613.0 m d cr n|||||||||| 231021s2023 xr fsbm 000 0 eng d NEZPRACOVANÝ IMPORT ABA006 cze ABA006 ABA006 rda Melenová, Markéta ISIS:163386 dis Analýza návrhu Aktu EU o umělé inteligenci eng Analysis of the proposal for the EU Artificial Intelligence Act / Markéta Melenová 2023 ?? stran : digital, PDF soubor Vedoucí práce: Tomáš Kliegr Diplomová práce (Ing.)—Vysoká škola ekonomická v Praze. Fakulta informatiky a statistiky, 2023 Obsahuje bibliografii Textový (vysokoškolská kvalifikační práce) Rok obhajoby 2023 This work attempted to explore the new landscape of proposed EU Artificial Intelligence act. For this the feedback submitted by general public was collected and analysed using semi-automated approach. Agglomerative clustering was used to identify the types of feedbacks. Neverheless, when used and tested in a standard way, such model assigned a feedback document, which typically covered whole range of topics, to only one cluster. Due to this, it was impossible to interpret, what topic characterizes given cluster or verify manually if the assignment is correct. In order to overcome the problem of clustering that assigns the whole document to 1 cluster only, to get a more detailed granularity, I decided to split the feedback files to smaller chunks based on an approximate paragraph pattern. This approach enhances the selected model’s suitability for the purposes of deeper analysis of selected topics, if needed, mapping the main topics of feedback to individual AIA articles and reflecting on the manner how the feedback was reflected in the final AIA version. This approach is characterised by high level of interpretability together with relatively low number of documents in the training set, which makes it possible to identify any misattributions in manual review without much time or effort spent. This is one of the by-products of the thesis. The work provides an overview of the main types of feedback and answers the research question if the processing of individual thematic parts of the EU Artificial Intelligence Act proposal aligns with the extent of feedback submitted during the designated feedback period. Způsob přístupu: Internet kognitivní informatika [obor dipl. práce] diplomové práce fd132022 czenas master's theses eczenas regulation AI Act artificial intelligence text analysis agglomerative clustering Kliegr, Tomáš ISIS:8484 ths Vadinský, Ondřej ISIS:43066 opn Vysoká škola ekonomická v Praze. Fakulta informatiky a statistiky kn20010709399 dgg https://insis.vse.cz/zp/82540/podrobnosti VŠKP v InSIS https://insis.vse.cz/zp/82540 Hlavní práce https://insis.vse.cz/zp/82540/posudek/vedouci Hodnocení vedoucího https://insis.vse.cz/zp/82540/posudek/oponent/79916 Oponentura https://insis.vse.cz/zp/82540/podrobnosti dc:identifier NEPOSILAT VSKP vse82540 231005 82540