|
|
Activity of Neural Network in Hidden Layers - Visualisation and Analysis
Fábry, Marko ; Grézl, František (referee) ; Karafiát, Martin (advisor)
Goal of this work was to create system capable of visualisation of activation function values, which were produced by neurons placed in hidden layers of neural networks used for speech recognition. In this work are also described experiments comparing methods for visualisation, visualisations of neural networks with different architectures and neural networks trained with different types of input data. Visualisation system implemented in this work is based on previous work of Mr. Khe Chai Sim and extended with new methods of data normalization. Kaldi toolkit was used for neural network training data preparation. CNTK framework was used for neural network training. Core of this work - the visualisation system was implemented in scripting language Python.
|
|
|
Sorting Networks Design Using Coevolutionary CGP
Fábry, Marko ; Hrbáček, Radek (referee) ; Drahošová, Michaela (advisor)
This paper deals with sorting networks design using Cartesian Genetic Programming and coevolution. Sorting networks are abstract models capable of sorting lists of numbers. Advantage of sorting networks is that they are easily implemented in hardware, but their design is very complex. One of the unconventional and effective ways to design sorting networks is Cartesian Genetic Programming (CGP). CGP is one of evolutionary algorithms that are inspired by Darwinian theory of evolution. Efficiency of the CGP algorithm can be increased by using coevolution. Coevolution is an approach that works with multiple populations, which are influencing one another and constantly evolving, thus prevent the local optima deadlock. In this work it is shown, that with the use of coevolution, it is possible to achieve nearly twice the acceleration compared to evolutionary design.
|
|
|
Activity of Neural Network in Hidden Layers - Visualisation and Analysis
Fábry, Marko ; Grézl, František (referee) ; Karafiát, Martin (advisor)
Goal of this work was to create system capable of visualisation of activation function values, which were produced by neurons placed in hidden layers of neural networks used for speech recognition. In this work are also described experiments comparing methods for visualisation, visualisations of neural networks with different architectures and neural networks trained with different types of input data. Visualisation system implemented in this work is based on previous work of Mr. Khe Chai Sim and extended with new methods of data normalization. Kaldi toolkit was used for neural network training data preparation. CNTK framework was used for neural network training. Core of this work - the visualisation system was implemented in scripting language Python.
|
|
|
Sorting Networks Design Using Coevolutionary CGP
Fábry, Marko ; Hrbáček, Radek (referee) ; Drahošová, Michaela (advisor)
This paper deals with sorting networks design using Cartesian Genetic Programming and coevolution. Sorting networks are abstract models capable of sorting lists of numbers. Advantage of sorting networks is that they are easily implemented in hardware, but their design is very complex. One of the unconventional and effective ways to design sorting networks is Cartesian Genetic Programming (CGP). CGP is one of evolutionary algorithms that are inspired by Darwinian theory of evolution. Efficiency of the CGP algorithm can be increased by using coevolution. Coevolution is an approach that works with multiple populations, which are influencing one another and constantly evolving, thus prevent the local optima deadlock. In this work it is shown, that with the use of coevolution, it is possible to achieve nearly twice the acceleration compared to evolutionary design.
|
guest ::
