|
Word2vec Models with Added Context Information
Šůstek, Martin ; Rozman, Jaroslav (referee) ; Zbořil, František (advisor)
This thesis is concerned with the explanation of the word2vec models. Even though word2vec was introduced recently (2013), many researchers have already tried to extend, understand or at least use the model because it provides surprisingly rich semantic information. This information is encoded in N-dim vector representation and can be recall by performing some operations over the algebra. As an addition, I suggest a model modifications in order to obtain different word representation. To achieve that, I use public picture datasets. This thesis also includes parts dedicated to word2vec extension based on convolution neural network.
|
|
Visual detection of small objects using available tools in MATLAB
Sladký, Jiří ; Dobossy, Barnabás (referee) ; Appel, Martin (advisor)
This thesis investigates possibilities of small object detection in pictures using YOLO method, a deep learning algorithm available in MATLAB. In the thesis, a detector was designed and trained to detect cows from top-down view. A tool was created, that performs detection using the proposed model even on high resolution images and counts the present objects. A generator of synthetic images was programmed, which helped with training the model. Various experiments were performed that found the limits of YOLO and validated contribution of the proposed improvements.
|
|
Bayesian and Neural Networks
Hložek, Bohuslav ; Rozman, Jaroslav (referee) ; Zbořil, František (advisor)
This paper introduces Bayesian neural network based on Occams razor. Basic knowledge about neural networks and Bayes rule is summarized in the first part of this paper. Principles of Occams razor and Bayesian neural network are explained. A real case of use is introduced (about predicting landslide). The second part of this paper introduces how to construct Bayesian neural network in Python. Such an application is shown. Typical behaviour of Bayesian neural networks is demonstrated using example data.
|
|
Tool for deep neural network design
Hladík, Jakub ; Harár, Pavol (referee) ; Uher, Václav (advisor)
The aim of this thesis was to create a program for visualization of artificial neural networks. The first part contains definition of artificial neural networks, comparison with biological neural networks and comparison with processors used in personal computers. Also contains brief description of their function and advantages/disadvantages of artificial neural networks. The second part contains brief explanation of deep learning. There are described methods and models used for learning. The third part contains introduction to Caffe framework and summary of the most frequently occuring layers in neural networks, that are created by using Caffe. Next part contains brief analysis of created program as well as detailed description of classes, that program contains.
|
|
Utilization of inverse reliability analysis tools for probability based design of selected structural parameters
Lipowczan, Martin ; Novák, Drahomír (referee) ; Lehký, David (advisor)
This bachelor thesis deals with the application of methodology and tools of inverse analysis in regards to probabilistic design of selected design parameters of structure. The first step was to get familiar with the probabilistic design and analysis, then understanding of the inverse analysis methodology itself which is based on artificial neural networks. After researching the topic we could get to the actual issue. To put the theory in practice easier examples were used at first. These were mathematical functions and one practical-based example, whereas the results were known in advance. This simplified a process of checking achieved values. Using software tools and especially DLNNET software allowed us to take on practical exercises. Used exercises are chosen from earlier undergraduate studies at the Faculty of Civil Engineering, Brno. The first of these was a design of reinforced concrete slab, where desired parameters were slab’s height and area of reinforcement. The second one was a design of a diagonal truss screw connection, aimed to size the screw diameter and its quantity.
|
|
Interconnection of Restricted Boltzmann machine method with statistical physics and its implementation in the processing of spectroscopic data
Vrábel, Jakub ; Hrdlička, Aleš (referee) ; Pořízka, Pavel (advisor)
Práca sa zaoberá spojeniami medzi štatistickou fyzikou a strojovým učením s dôrazom na základné princípy a ich dôsledky. Ďalej sa venuje obecným vlastnostiam spektroskopických dát a ich zohľadnení pri pokročilom spracovaní dát. Začiatok práce je venovaný odvodeniu partičnej sumy štatistického systému a štúdiu Isingovho modelu pomocou "mean field" prístupu. Následne, popri základnom úvode do strojového učenia, je ukázaná ekvivalencia medzi Isingovým modelom a Hopfieldovou sieťou - modelom strojového učenia. Na konci teoretickej časti je z Hopfieldovej siete odvodený model Restricted Boltzmann Machine (RBM). Vhodnosť použitia RBM na spracovanie spektroskopických dát je diskutovaná a preukázaná na znížení dimenzie týchto dát. Výsledky sú porovnané s bežne používanou Metódou Hlavných Komponent (PCA), spolu so zhodnotením prístupu a možnosťami ďalšieho zlepšovania.
|
|
Recognition of electrochemical signals using artificial neuronal network
Šílený, Jan ; Kuchta, Radek (referee) ; Hubálek, Jaromír (advisor)
Automatical electrochemical measurements are sources of large data sets intended for further analysis. This work deals with classification, evaluation and processing of electrochemical signals using artificial neural networks. Due to high dimensionality of input data, an autoassociative neural network (AANN) is used in this work. This type of network performs dimensionality reduction via filtering the input data into relatively small number of principal parameters at the bottleneck output. These extracted parameters can be used for classification, evaluation and additional modelling of analyzed data trough the reconstructive part of this network. Furthermore, this work deals with implementation of a feedforward neural network in OpenCL language.
|
|
Classification of heart beats from multilead ECG using principal component analysis
Vlček, Milan ; Vítek, Martin (referee) ; Ronzhina, Marina (advisor)
The resume of this master´s thesis is to introduce reader into principal component analysis (PCA), namely, the use of PCA for analysis of ECG. This method allows to reduce quantity of the data without loss of useful information. That is why PCA is widespread for preprocessing of the data for further classification, which this thesis also deals. Data available at the Department of Biomedical Engineering at the University of Technology in Brno were used in this work. All the methods were realized using Matlab.
|
|
The Use of Artificial Intelligence on Stock Market
Lajczyk, Pavel ; Budík, Jan (referee) ; Dostál, Petr (advisor)
This master's thesis deals with artificial neural networks and possibilities of their use on stock market. In next chapters of this thesis there are provided design and implementation of stock prices prediction tool. The implementation is done with use of the MATLAB software. The created prediction tool is then tested in a simple trading simulation and achieved results are discussed in the end
|
| |