|
|
Recurrent Neural Networks for Speech Recognition
Nováčik, Tomáš ; Karafiát, Martin (referee) ; Veselý, Karel (advisor)
This master thesis deals with the implementation of various types of recurrent neural networks via programming language lua using torch library. It focuses on finding optimal strategy for training recurrent neural networks and also tries to minimize the duration of the training. Furthermore various types of regularization techniques are investigated and implemented into the recurrent neural network architecture. Implemented recurrent neural networks are compared on the speech recognition task using AMI dataset, where they model the acustic information. Their performance is also compared to standard feedforward neural network. Best results are achieved using BLSTM architecture. The recurrent neural network are also trained via CTC objective function on the TIMIT dataset. Best result is again achieved using BLSTM architecture.
|
|
| |
|
|
Proposal for Increasing Business Effectiveness of an E-shop
Kuna, Martin ; Veselý, Karel (referee) ; Dydowicz, Petr (advisor)
This bachelor’s thesis evaluates the most important factors influencing business effectiveness of Anglia-Trading.cz e-commerce web site. The first part contains a theoretical introduction to the topic of on-line selling and business effectiveness. In the second, practical part of this work, the theoretical findings are used to evaluate the effectiveness of the above mentioned on-line store. This evaluation is focused mainly on uncovering possible critical points and undiscovered opportunities. In the following chapter, several improvement suggestions and possible solutions of problems uncovered during the analysis are outlined. In the final part, main outcomes of application of these suggestions and solutions are described.
|
|
|
Speech Recognition for Air Traffic Communication
Žmolíková, Kateřina ; Burget, Lukáš (referee) ; Veselý, Karel (advisor)
This thesis deals with speech recognition. The aim is to build a speech recognition system based on neural networks and test it on recordings of air traffic communication. Final acoustic model will be used in project A-PiMod. The system reached word error rate 29.5%. Next task of this thesis was to experiment with neural networks which are part of acoustic model. First experiments explored its simplification and acceleration and its impact on error rate. Next experiments dealt with activation function rectifier and convolutional neural networks. Experiments with convolutional neural networks achieved 1.5% improvement, so the final result was 0.4% better than fully connected network with the same architecture.
|
|
|
Building deep networks using autoencoders
Lohniský, Michal ; Veselý, Karel (referee) ; Hradiš, Michal (advisor)
This thesis deals with pretraining deep networks by autoencoders. Components of neural networks are described in first chapters. Rest of chapters aims to deep network trainings and to results of experiments where autoencoder pretraining and Backpropagation algorithm are compared. Results showed positive contribution of autoencoder pretraining, mainly in combination with Finetuning.
|
|
|
Parallel Training of Neural Networks for Speech Recognition
Veselý, Karel ; Fousek, Petr (referee) ; Burget, Lukáš (advisor)
This thesis deals with different parallelizations of training procedure for artificial neural networks. The networks are trained as phoneme-state acoustic descriptors for speech recognition. Two effective parallelization strategies were implemented and compared. The first strategy is data parallelization, where the training is split into several POSIX threads. The second strategy is node parallelization, which uses CUDA framework for general purpose computing on modern graphic cards. The first strategy showed a 4x speed-up, while using the second strategy we observed nearly 10x speed-up. The Stochastic Gradient Descent algorithm with error backpropagation was used for the training. After a short introduction, the second chapter of this thesis shows the motivation and introduces the neural networks into the context of speech recognition. The third chapter is theoretical, the anatomy of a neural network and the used training method are discussed. The following chapters are focused on the design and implementation of the project, while the phases of the iterative development are described. The last extensive chapter describes the setup of the testing system and reports the experimental results. Finally, the obtained results are concluded and the possible extensions of the project are proposed.
|
|
|
Multi-Task Neural Networks for Speech Recognition
Egorova, Ekaterina ; Veselý, Karel (referee) ; Karafiát, Martin (advisor)
První část této diplomové práci se zabývá teoretickým rozborem principů neuronových sítí, včetně možnosti jejich použití v oblasti rozpoznávání řeči. Práce pokračuje popisem viceúkolových neuronových sítí a souvisejících experimentů. Praktická část práce obsahovala změny software pro trénování neuronových sítí, které umožnily viceúkolové trénování. Je rovněž popsáno připravené prostředí, včetně několika dedikovaných skriptů. Experimenty představené v této diplomové práci ověřují použití artikulačních characteristik řeči pro viceúkolové trénování. Experimenty byly provedeny na dvou řečových databázích lišících se kvalitou a velikostí a representujících různé jazyky - angličtinu a vietnamštinu. Artikulační charakteristiky byly také kombinovány s jinými sekundárními úkoly, například kontextem, s záměrem ověřit jejich komplementaritu. Porovnaní je provedeno s neuronovými sítěmi různých velikostí tak, aby byl popsán vztah mezi velikostí neuronových sítí a efektivitou viceúkolového trénování. Závěrem provedených experimentů je, že viceúkolové trénování s použitím artikulačnich charakteristik jako sekundárních úkolů vede k lepšímu trénování neuronových sítí a výsledkem tohoto trénování může být přesnější rozpoznávání fonémů. V závěru práce jsou viceúkolové neuronové sítě testovány v systému rozpoznávání řeči jako extraktor příznaků.
|
|
|
Hybrid Recognizer of Isoladed Words
Veselý, Karel ; Černocký, Jan (referee) ; Grézl, František (advisor)
The speaker independent isolated words recignizer has various practical applications. For example it can be used to control home gadgets by PC. Even more interesting is possibility that it can be built in the user interface of any application or even into operating system to perform command based control such as invocation of applications, or execution of any other specific action. The most remarkable application of isolated recognition is in electronical dictionaries. A voice controlled word lookup could be new feature of the next generation dictionaries. Very useful is the ability to ouptut ordered list of the most likely words, which gives the user ability to learn and distinguish similar words.
|
|
|
Accelerating Face Anti-Spoofing Algorithms
Beňuš, Ondřej ; Havel, Jiří (referee) ; Veselý, Karel (advisor)
Tato práce se specializuje na akceleraci algoritmu z oblasti obličejově zaměřených anti-spoofing algoritmů s využitím grafického hardware jakožto platformy pro paralelní zpracování dat. Jako framework je použita technologie OpenCL která umožňuje použití od výkoných stolních počítačů po přenosná zařízení, od různých akcelerátorů jako grafické čipy, či ASIC až po procesory typu x86 bez vazby na konkrétního výrobce či operační systém. Autor předkládá čtenáři rozbor a akcelerovanou implementaci široce používaného algoritmu a dopadu urychlení výpočtu.
|
|
| |
guest ::
RSS feed.