Použití velkého jazykového modelu jako asistenta pro konceptuální modelování
| Název práce v češtině: | Použití velkého jazykového modelu jako asistenta pro konceptuální modelování |
|---|---|
| Název v anglickém jazyce: | Using large language model as an assistant for conceptual modeling |
| Klíčová slova: | konceptuální modelování|velké jazykové modely|AI asistant |
| Klíčová slova anglicky: | conceptual modeling|large language models|AI assistant |
| Akademický rok vypsání: | 2023/2024 |
| Typ práce: | diplomová práce |
| Jazyk práce: | čeština |
| Ústav: | Katedra softwarového inženýrství (32-KSI) |
| Vedoucí / školitel: | doc. Mgr. Martin Nečaský, Ph.D. |
| Řešitel: | skrytý - zadáno a potvrzeno stud. odd. |
| Datum přihlášení: | 15.03.2024 |
| Datum zadání: | 19.03.2024 |
| Datum potvrzení stud. oddělením: | 19.03.2024 |
| Datum a čas obhajoby: | 11.09.2024 09:00 |
| Datum odevzdání elektronické podoby: | 18.07.2024 |
| Datum odevzdání tištěné podoby: | 18.07.2024 |
| Datum proběhlé obhajoby: | 11.09.2024 |
| Oponenti: | doc. RNDr. Irena Holubová, Ph.D. |
| Konzultanti: | Mgr. Štěpán Stenchlák |
| Zásady pro vypracování |
| When a team works on any project, unifying the language and defining the terms everyone will use is beneficial. This is addressed by creating conceptual models. In a nutshell, conceptual models capture entities, their attributes, and relationships. However, creating these conceptual models takes a non-trivial amount of time. This problem often means conceptual modeling is skipped in practice despite severe consequences later. One possible solution to deal with this problem is to develop a large language model (LLM) based assistant to, for example, provide suggestions on what the user can add to his conceptual model based on some domain description. This could reduce the workload of the conceptual modeling.
This thesis aims to develop such an LLM-based assistant. Based on a baseline LLM-based assistant developed in a student research project, the thesis will provide guidelines on which methods and configurations to use with LLMs to achieve better results with tasks such as attributes extraction, relationships extraction, and entities extraction from an unstructured plain text by experimenting with: (1) different types of LLM prompt techniques, such as few-shot prompting, the chain of thoughts, and the tree of thoughts, (2) different approaches, such as retrieval augmented generation, and (3) different LLMs. The student will design and perform the experiments and compile the guidelines based on the experiments. Using the guidelines, the student will extend the baseline LLM-based assistant and finally integrate it into the conceptual modeling component of the Dataspecer tool. |
| Seznam odborné literatury |
| [1] CHEN, Kua, et al. Automated Domain Modeling with Large Language Models: A Comparative Study. In: 2023 ACM/IEEE 26th International Conference on Model Driven Engineering Languages and Systems (MODELS). IEEE, 2023. p. 162-172.
[2] WEI, Jason, et al. Chain-of-thought prompting elicits reasoning in large language models. Advances in Neural Information Processing Systems, 2022, 35: 24824-24837. [3] LONG, Jieyi. Large Language Model Guided Tree-of-Thought. arXiv preprint arXiv:2305.08291, 2023. [4] GAO, Yunfan, et al. Retrieval-augmented generation for large language models: A survey. arXiv preprint arXiv:2312.10997, 2023. [5] Dataspecer tool. https://dataspecer.com/ |