1. Tvorba chatovacího robota s využitím hlubokých neuronových sítí
- Author
-
SCHUH, Tom
- Subjects
development ,neuronové sítě ,neural networks ,Chatbot ,umělá inteligence ,vývoj ,LSTM ,artificial intelligence - Abstract
The master thesis focuses on a creating chatbot with using artificial intelligence. At the start of theoretical part, thesis briefly introduce a history of chatbots. Futhuremore the theoretical part introduces some important parts of neural network and explain their principle. Then the master thesis describes some types of neural networks, for exmaple LSTM and GRU, which are often used for creating chatbots. Furthermore, this part of thesis describes methods used for nature language processing, such as Bag-Of-Words or Word Embeggding. The practical part of this master thesis deals with the creation of a retrival-based chatbot and simulation of an aplication for its use in a real environment. Chatbot serves to answer questions about the Covid-19 pandemic. First, practical part introduce the dataset that is used for chatbot training. Furthermore is explained in detail the software implementation of training script and chatbot application as whole. The last part deals with experiments on the model itself, where the effort to achieve the best possible accuracy on the validation data using optimization methods.
- Published
- 2021