|
|
Utilising Large Pretrained Language Models for Configuration and Support of a Clinical Information System
Sova, Michal ; Burget, Radek (referee) ; Rychlý, Marek (advisor)
The aim of this work is to get acquainted with the essence and use of large pre-trained language models, to get acquainted with the configuration options of the clinical information system FONS Enterprise and the possibility of its adaptation to the specific environment of customers. The work first presents large pre-trained language models and the FONS Enterprise clinical information system. This work examines possibilities of training models and implementing RAG methods on data from the clinical system. The implementation of the RAG architecture is supported by the tools LangChain and LlamaIndex. The results show that the RAG method with the Gemma model and the bge-m3 embedding model provides the most relevant answers on basic questions, but struggles to understand more complex questions. The method of pre-training the model does not produce the expected results, even after adjusting the training parameters.
|
|
|
Detection of key information in emergency calls
Sarvaš, Marek ; Plchot, Oldřich (referee) ; Schwarz, Petr (advisor)
Tiesňové volania sa zvyčajne uskutočňujú v extrémne stresujúcich podmienkach, kde volajúci často poskytuje dôležité informácie rýchlo, čo sťažuje operátorom tiesňovej linky presne zachytiť všetky podrobnosti. To môže viesť k opakovaným otázkam o už poskytnutých informáciách a oneskoreniu reakcie pohotovostnej služby. Cieľom tejto práce je zmierniť tento problém a potenciálne urýchliť reakciu pohotovostných služieb nasadením neurónovej siete na extrakciu informácií, konkrétne so zameraním na úlohu Rozpoznávania pomenovaných entít (NER). Táto práca skúma rôzne prístupy založené na architektúre typu Transformers, ako sú predtrénované enkodér modely, enkodér-dekodér (sequence-2-sequence) a veľké jazykové modely. Vybrané modely dosiahli zatiaľ najlepšie výsledky na verejne dostupných českých NER datasetoch. Okrem toho boli vytvorené nové NER datasety z poskytnutých nahrávok skutočných tiesňových volaní a odpovedajúcich metadát. Predstavené modely boli natrénované a vyhodnotené na týchto novovytvorených datasetoch a úspešne dosiahli rozumné výsledky pre extrakciu mien a polohy.
|
|
|
Large Language Models for Generating Code Focusing on Embedded Systems
Vadovič, Matej ; Nosko, Svetozár (referee) ; Smrž, Pavel (advisor)
The goal of this work was to adapt a pre-trained language model for the purpose of generating code in the field of embedded systems. The work introduces a new dataset for fine-tuning code generation models, consisting of 50,000 pairs of source code and comments focused on embedded systems programming. This dataset is composed of collected source code from the GitHub platform. Two new language models for code generation, based on transformer architecture pre-trained models, were fine-tuned on the data of the new corpus. Model MicroCoder is based on the CodeLLaMA-Instruct 7B model, and during its fine-tuning, the QLoRA technique was used to minimize computational requirements. The second model, MicroCoderFIM, is based on the StarCoderBase 1B model and supports code infilling. The individual models were compared based on BLEU, CodeBLEU, ChrF++, and ROUGE-L metrics. Model MicroCoderFIM achieves the best adaptation results to the new task, with over 120% improvement in all measured metrics. The weights of the models along with the new dataset are freely accessible on a public repository.
|
|
| |
|
|
Sentiment Analysis of Czech and Slovak Social Networks and Web Discussions
Slúka, Dušan ; Doležal, Jan (referee) ; Smrž, Pavel (advisor)
This bachelor’s thesis deals with the issue of extraction and analysis of data obtained from social networks to understand public opinion on various social topics. The goal is systematic categorization and interpretation of contents. The problem is solved through a platform for opinion extraction and automatic data classification, which allows the creation of thematic subcategories and sorting into them. The result of the work is a system that analyzes social networks and provides deeper insight into public opinion on social topics. The system enables organizations to better understand the dynamics of online discourse. The benefit of this work is the provision of a new tool for the analysis of social issues, which can serve the academic sphere as well as organizations from practice.
|
|
|
Human web browsing simulation
Doležal, Jáchym ; Setinský, Jiří (referee) ; Hranický, Radek (advisor)
Tato práce představuje slibný nástroj pro automatizovanou webovou navigaci a plnění specifických cílů podle rozhodnutí daných velkým jazykovým modelem, který používá aktuální informace z dané stránky. Výsledky simulátoru s modelem GPT 4 Turbo demonstrují efektivitu tohoto nástroje, který dosahuje úspěšnosti přes 80% v dokončování předem definovaných cílů. Výsledky dokazují použiteljnost tohoto nástroje v reálných případech užití.
|
|
|
Generating Documentation to Source Code in Python
Novosád, Juraj ; Nosko, Svetozár (referee) ; Smrž, Pavel (advisor)
The aim of this work is to adapt selected language models on domain data and to develop a system that would allow their use on commonly available hardware. The models have been adapted to generate documentation for undocumented source code in the Python progra- mming language to follow the Google Style convention. A prerequisite of model adaptation was to obtain domain data and process it appropriately for the purpose of model fine-tuning. This work focuses on fine-tuning models with fewer than one billion parameters, for the sake of enabling inference even on commonly available hardware. Part of the work was to objectively evaluate the quality of the adapted models. For this reason, I developed a tool that evaluates the quality of the generated documentation on a selected corpus of models. The evaluation of the adapted models showed that they achieve comparable performance to multiply larger models for general tasks, such as gpt-3.5-turbo-0125. The result of this work is a server capable of horizontal scaling that integrates the capabilities of more than just the adapted models through an easy-to-use API.
|
|
|
Brno Communication Agent
Neprašová, Kateřina ; Fajčík, Martin (referee) ; Smrž, Pavel (advisor)
This thesis focuses on domain-specific communication agents, the aim is to create a functional communication agent for both tourists and locals in Brno, providing relevant and up-to-date information according to individual user needs. It describes large language models, analyses existing technologies for domain-specific communication agents and their construction. It focuses on the creation of a knowledge base and the implementation of an efficient dialogue interface using Retrieval-Augmented Generation (RAG), while comparing selected language models on different tasks.
|
|
|
Failure Modes of Large Language Models
Milová, Soňa ; Špelda, Petr (advisor) ; Střítecký, Vít (referee)
Failure Modes of Large Language Models Soňa Milová Abstract Diploma thesis "The failure modes of Large Language Models" focuses on addressing failure modes of Large Language Models (LLMs) from the ethical, moral and security point of view. The method of the empirical analysis is document analysis that defines the existing study, and the process by which failure modes are selected from it and analysed further. It looks closely at OpenAI's Generative Pre-trained Transformer 3 (GPT-3) and its improved successor Instruct Generative Pre-trained Transformer (IGPT). The thesis initially investigates model bias, privacy violations and fake news as the main failure modes of GPT-3. Consequently, it utilizes the concept of technological determinism as an ideology to evaluate whether IGPT has been effectively designed to address all the aforementioned concerns. The core argument of the thesis is that the utopic and dystopic view of technological determinism need to be combined with the additional aspect of human control. LLMs are in need of human involvement to help machines better understand context, mitigate failure modes, and of course, to ground them in reality. Therefore, contextualist view is portrayed as the most accurate lens through which to look at LLMs as it argues they depend on the responsibilities,...
|
guest ::
