Rafaj, Filip - Search Results - Digital Repository

guest :: login Digital Repository
		Search		Submit		Help		About

Home > Search Results: Rafaj, Filip

Search:



Search Tips :: Simple Search

Search collections:

Sort by:	Display results:	Output format:

	Pojmenované entity a ontologie metodami hlubokého učení Rafaj, Filip ; Hajič, Jan (advisor) ; Žabokrtský, Zdeněk (referee) In this master thesis we describe a method for linking named entities in a given text to a knowledge base - Named Entity Linking. Using a deep neural architecture together with BERT contextualized word embeddings we created a semi-supervised model that jointly performs Named Entity Recognition and Named Entity Disambiguation. The model outputs a Wikipedia ID for each entity detected in an input text. To compute contextualized word embeddings we used pre-trained BERT without making any changes to it (no fine-tuning). We experimented with components of our model and various versions of BERT embeddings. Moreover, we tested several different ways of using the contextual embeddings. Our model is evaluated using standard metrics and surpasses scores of models that were establishing the state of the art before the expansion of pre-trained contextualized models. The scores of our model are comparable to current state-of-the-art models. Detailed record
	Pojmenované entity a ontologie metodami hlubokého učení Rafaj, Filip ; Hajič, Jan (advisor) ; Žabokrtský, Zdeněk (referee) In this master thesis we describe a method for linking named entities in a given text to a knowledge base - Named Entity Linking. Using a deep neural architecture together with BERT contextualized word embeddings we created a semi-supervised model that jointly performs Named Entity Recognition and Named Entity Disambiguation. The model outputs a Wikipedia ID for each entity detected in an input text. To compute contextualized word embeddings we used pre-trained BERT without making any changes to it (no fine-tuning). We experimented with components of our model and various versions of BERT embeddings. Moreover, we tested several different ways of using the contextual embeddings. Our model is evaluated using standard metrics and surpasses scores of models that were establishing the state of the art before the expansion of pre-trained contextualized models. The scores of our model are comparable to current state-of-the-art models. Detailed record
	Analýza sondové charakteristiky s využitím neuronových sítí Rafaj, Filip ; Roučka, Štěpán (advisor) ; Kocán, Pavel (referee) Langmuir probes are used on our faculty to measure current-voltage characteristics of low-temperature weakly ionized plasma. From these measurements an electron temperature and an electron number density are obtained. Classical method of doing that is based on the linear least squares fitting. In this thesis we use neural networks as an alternative method of determining the plasma parameters. We train a feedforward neural network with a help of a stochastic gradient descent and a backpropagation algorithms together with a training data based on an analytical model the characteristic. We study network's accuracy, robustness and computational resources demands, all compared with the classical methods. Powered by TCPDF (www.tcpdf.org) Detailed record

Interested in being notified about new results for this query?
Subscribe to the RSS feed.

Digital Repository :: :: :: ::
Powered by v1.1.2
Maintained by

This site is also available in the following languages:
Česky English