|
|
Automatic diagnosis of the 12-lead ECG using deep learning
Blaude, Ondřej ; Chmelík, Jiří (referee) ; Provazník, Valentine (advisor)
The aim of this diploma thesis is to investigate the problematics of automatic ECG diagnostics, namely on twelve-lead recordings. This problem is solved by standard methods such as random forest, artificial neural networks or K-nearest neighbors. However, thanks to its ability to independently extract symptoms, deep learning methods are also popular. All these methods are described in the theoretical part. In the practical part, deep learning models were designed, functionality support was verified using data from the PhysioNet database. Two pilot models were created and subsequently optimized. From the entire parameter optimization procedure, three models are available, of which the best accuracy achieves an F1 score of 87.35% and 83.7%, and the second best achieves an F1 score of 77.74% and an accuracy of 84.53%. The results achieved are discussed and compared with those of similar publications.
|
|
|
Document Information Extraction
Janík, Roman ; Špaňhel, Jakub (referee) ; Hradiš, Michal (advisor)
S rozvojem digitalizace přichází potřeba analýzy historických dokumentů. Důležitou úlohou pro extrakci informací a dolování dat je rozpoznávání pojmenovaných entit. Cílem této práce je vyvinout systém pro extrakci informací z českých historických dokumentů, jako jsou noviny, kroniky a matriční knihy. Byl navržen systém pro extrakci informací, jehož vstupem jsou naskenované historické dokumenty zpracované OCR algoritmem. Systém je založen na modifikovaném modelu RoBERTa. Extrakce informací z českých historických dokumentů přináší výzvy v podobě nutnosti vhodného korpusu pro historickou Češtinu. Pro trénování systému byly použity korpusy Czech Named Entity Corpus (CNEC) a Czech Historical Named Entity Corpus (CHNEC), spolu s mým vlastním vytvořeným korpusem. Systém dosahuje úspěšnosti 88,85 F1 skóre na CNEC a 87,19 F1 skóre na CHNEC. Toto je zlepšení o 1,36 F1 u CNEC a 5,19 F1 u CHNEC a tedy nejlepší známé výsledky.
|
|
|
Classification of board defects in semiconductor manufacturing
Jašek, Filip ; Vágner, Martin (referee) ; Dřínovský, Jiří (advisor)
This diploma thesis focuses on detecting defects in semiconductor wafer manufacturing. It explores methods for identifying faulty chips and controlling yield during production. To classify defects machine learning techniques are used. Initially, ResNet18 architecture was used for inference, but low accuracy was attributed to limited input data. Transfer learning with ResNet50v2 was then attempted, resulting in improved metric with different dataset. Hyperparameter tuning and data augmentations were also explored. The study found that autoencoders for data compression during inference increased speed but led to degraded evaluation metrics.
|
|
|
Playing the Board Game Stratego by Computer
Irovský, Dominik ; Šátek, Václav (referee) ; Zbořil, František (advisor)
The topic of this thesis is the board game of Stratego. This game features incomplete information. The goal of this thesis is research of existing game playing algorithms and, design and implementation of new solution. For the new solution modified version of Monte Carlo Tree Search as well as alfa-beta algorithm and expectimax were used. The solution was implemented as a console application with possibility of future expansion. Functionality of the solution was validated and tested using experiments. Effectivity of the final algorithm was satisfying
|
|
|
Safe and Secure High-Risk AI: Evaluation of Robustness
Binterová, Eliška ; Špelda, Petr (advisor) ; Střítecký, Vít (referee)
The aim of the thesis is to examine Invariant Risk Minimization (IRM) as an existing method for achieving model robustness and assess whether it could potentially serve as means for conformity assessment in the emerging legislative framework of the European Artificial Intelligence Act. Research shows that many cases of erroneous performance in AI systems are caused by machine learning models lacking robustness to changes in data distributions and thus being unable to properly generalize to new environments. In order to achieve reliable performance, the models must exhibit a certain level of robustness to these changes. IRM is a relatively new method designed to achieve such outcomes. This is very much in alignment to the objectives of the EU AI Act that aims for trustworthy AI. The thesis thus examines the congruence of the IRM method and the requirements in the EU AI Act and asks whether IRM can serve as a universal method for ensuring safe and secure AI compliant with European legal requirements through the analysis of existing empirical and theoretical results.
|
|
|
The influence of artificial intelligence on the perception of quality and trustworthiness of content in digital communication
Richter, Martin ; Slavíček, Daniel (advisor) ; Koblovský, Petr (referee)
This master's thesis focuses on the impact of artificial intelligence (AI) on digital communication, content perception, and potential negative consequences associated with generated content. The thesis also presents opportunities that generative artificial intelligence brings in the context of multimedia content creation. At the same time, it addresses factors and biases that influence the perception of such generated content. The thesis also analyzes the awareness and abilities of primary and secondary school students in the field of generative AI. The results show that students have an awareness of the capabilities of generative AI but struggle to recognize generated content, which points to potential risks associated with manipulation and disinformation.
|
|
|
Ethical aspects of personalized content in selected online media
Dorňáková, Tereza ; Moravec, Václav (advisor) ; Lokšík, Martin (referee)
The thesis is focused on personalized content recommendation in Czech online news media and the ethical issues related to its implementation. The subject of the qualitative analysis are semi- structured interviews with representatives of two Czech media houses. The research sample includes respondents from Seznam Zprávy and Seznam, and from E15 and Czech News Center. The interviewees included representatives of newsrooms as well as media and product management. The aim of the study was to find out to what extent they use personalised content recommendation tools, to outline the reasons why they have decided to take this step or are considering it. It turned out that within the studied online news media they are still using personalised content recommendation to a minimum, or they are testing or planning to introduce these tools to a greater extent. Personalised recommendation tools are more widely used on the homepage of Seznam. The subject of this thesis was also to identify ethical issues related to the introduction or use of personalised recommendation. The main areas, according to the respondents' statements, include issues of setting selection parameters, the ability to retain agenda setting by editors, automation and transparency.
|
|
|
Enhancing Security Monitoring with AI-Enabled Log Collection and NLP Modules on a Unified Open Source Platform
Safonov, Yehor ; Zernovic, Michal
The number of computer attacks continues to increasedaily, posing significant challenges to modern securityadministrators to provide security in their organizations. Withthe rise of sophisticated cyber threats, it is becoming increasinglydifficult to detect and prevent attacks using traditional securitymeasures. As a result, security monitoring solutions such asSecurity Information and Event Management (SIEM) have becomea critical component of modern security infrastructures. However,these solutions still face limitations, and administrators areconstantly seeking ways to enhance their capabilities to effectivelyprotect their cyber units. This paper explores how advanced deeplearning techniques can help boost security monitoring capabilitiesby utilizing them throughout all stages of log processing. Thepresented platform has the potential to fundamentally transformand bring about a significant change in the field of securitymonitoring with advanced AI capabilities. The study includes adetailed comparison of modern log collection platforms, with thegoal of determining the most effective approach. The key benefitsof the proposed solution are its scalability and multipurposenature. The platform integrates an open source solution andallows the organization to connect any event log sources or theentire SIEM solution, normalize and filter data, and use thisdata to train and deploy different AI models to perform differentsecurity monitoring tasks more efficiently.
|
|
|
Automatic evaluation of fermentation degree of cocoa beans
Sedlmajer, Jakub ; Kůdela, Jakub (referee) ; Škrabánek, Pavel (advisor)
Standardní metodou zjištění výsledné kvality fermentace kakaových semen je tzv. cut-test. Při tomto testu je kakaové semeno rozříznuto na 2 poloviny, a následně je posouzeno vysoce kvalifikovaným odborníkem. Vzhledem k obrovským objemům zpracovávaného kakaa je automatizace tohoto procesu pomocí strojového vidění nevyhnutelná. Tato práce se zabývá návrhem a vytvořením softwarových nástrojů, které to efektivně umožní.
|
|
| |
guest ::
