The thesis focuses on the use of deep recurrent neural network, architecture Long Short-Term Memory for robust denoising of audio signal. LSTM is currently very attractive due to its characteristics to remember previous weights, or edit them not only according to the used algorithms, but also by examining changes in neighboring cells. The work describes the selection of the initial dataset and used noise along with the creation of optimal test data. For network training, the KERAS framework for Python is selected. Candidate networks for possible solutions are explored and described, followed by several experiments to determine the true behavior of the neural network.

This diploma thesis deals with the realization of a business order, consisting in creating a system of rapid analysis and optimization based on real business data of currency pairs using the MATLAB software. Within this developed topic a customer automated trading system is implemented, including the analysis and the optimization of the system parameters. There was also used a predictive system of the development of the prices, using artificial intelligence methods, based on historical data. The obtained data from the optimization and the simulations of the automated trading system are clearly summarized in graphic outputs.

The master´s thesis is focused on advanced scoring of sleep data, which was performed using deep neural network. Heart rate data and the movement information were used for scoring measured using an Apple Watch smartwatch. After appropriate pre-processing, this data serves as input parameters to the designed networks. The goal of the LSTM network was to classify data into either two groups for sleep and wake or into three groups for wake, Non-REM and REM. The best results were achieved by network doing classification of sleep vs. wake using the accelerometer. The statistical evaluation of this best-designed network reached the values of sensitivity 71,06 %, specificity 57,05 %, accuracy 70,01 % and F1 score 81,42 %.

This thesis is focus on monitoring state of machine parts that are under the most stress. Type of artificial intelligence used in this work is recurrent neural network and its modifications. Chosen type of neural network was used because of the sequential character of used data. This thesis is solving three problems. In first problem algorithm is trying to determine state of mill tool wear using recurrent neural network. Used method for monitoring state is indirect. Second Problem was focused on detecting fault of a bearing and classifying it to specific category. In third problem RNN is used to predict RUL of monitored bearing.

Tato práce se zabývá automatickým generovaním popisů obrázků s využitím několika druhů neuronových sítí. Práce je založena na článcích z MS COCO Captioning Challenge 2015 a znakových jazykových modelech, popularizovaných A. Karpathym. Navržený model je kombinací konvoluční a rekurentní neuronové sítě s architekturou kodér--dekodér. Vektor reprezentující zakódovaný obrázek je předáván jazykovému modelu jako hodnoty paměti LSTM vrstev v síti. Práce zkoumá, na jaké úrovni je model s takto jednoduchou architekturou schopen popisovat obrázky a jak si stojí v porovnání s ostatními současnými modely. Jedním ze závěrů práce je, že navržená architektura není dostatečná pro jakýkoli popis obrázků.

The thesis focuses on the use of deep recurrent neural network, architecture Long Short-Term Memory for robust denoising of audio signal. LSTM is currently very attractive due to its characteristics to remember previous weights, or edit them not only according to the used algorithms, but also by examining changes in neighboring cells. The work describes the selection of the initial dataset and used noise along with the creation of optimal test data. For creation of the training network is selected KERAS framework for Python and are explored and discussed possible candidates for viable solutions.

In this thesis, methods of implementing an online gesture recognition system for the novel Stylaero Board device are investigated. Two methods are evaluated - one based on LSTMs and one based on HMMs - on three kinds of gestures: Tap, circle, and flick motions. A method’s performance was measured in its accuracy in determining both whether any of the above listed gestures were performed and, if so, which gesture, in an online single-pass scenario. Insight was acquired regarding the technical challenges and possible solutions to the online aspect of the problem. Poor performance was, however, observed in both methods, with a likely culprit identified as low quality of training data, due to an arduous and complex gesture performance capturing process. Further research improving on the process of gathering data is suggested.

Chemical sensor based systems that are used for detection, identification, or quantification of various gases are very complex in nature. Sensor response data collected as a multivariate time series signals encounters gradual change of the sensor characteristics(known as sensor drift) due to several reasons. In this thesis, drift component of a silicon carbide Field-Effect Transistor (SiC-FET) sensor data was analyzed using time series. The data was collected from an experiment measuring output response of the sensor with respect to gases emitted by certain experimental object at different temperatures. Augmented Dickey Fuller Test (ADF) was carried out to analyze the sensor drift which revealed that stochastic trend along with deterministic trend characterized the drift components of the sensor. The drift started to rise in daily measurements which contributed to the total drift. / Traditional Autoregressive Integrated Moving Average (ARIMA) and deep learning based Long Short-Term Memory (LSTM) algorithm were carried out to forecast the sensor drift in reduced set of data. However, reduction of the data size degraded the forecasting accuracy and imposed loss of information. Therefore, careful selection of data using only one temperature from the temperature cycle was chosen instead of all time points. This chosen data from sensor array outperformed forecasting of sensor drift than reduced dataset using both traditional and deep learning methods.

The thesis investigates a publicly available dataset which covers morethan three million events in football matches. The aim of the study isto train machine learning models capable of modeling the relationshipbetween a shot event and its outcome. That is, to predict if a footballshot will result in a goal or not. By representing the shot indifferent ways, the aim is to draw conclusion regarding what elementsof a shot allows for a good prediction of its outcome. The shotrepresentation was varied both by including different numbers of eventspreceding the shot and by varying the set of features describing eachevent.The study shows that the performance of the machine learning modelsbenefit from including events preceding the shot. The highestpredictive performance was achieved by a long short-term memory neuralnetwork trained on the shot event and six events preceding the shot.The features which were found to have the largest positive impact onthe shot events were the precision of the event, the position on thefield and how the player was in contact with the ball. The size of thedataset was also evaluated and the results suggest that it issufficiently large for the size of the networks evaluated.

Natural language processing is a field in computer science that is becoming increasingly important. One important part of NLP is the ability to sort text to the past, present or future, depending on when the event came or will come about. The objective of this thesis was to use text classification to classify Swedish sentences based on their tense, either past, present or future. Furthermore, the objective was also to compare how lemmatisation would affect the performance of the models. The problem was tackled by implementing three machine learning models on both lemmatised and not lemmatised data. The machine learning models were Naive Bayes, LSTM and BERT. The result showed that the overall performance was affected negatively when the data was lemmatised. The best performing model was BERT with an accuracy of 96.3\%. The result was useful as the best performing model had very high accuracy and performed well on newly constructed sentences. / Språkteknologi är område inom datavetenskap som som har blivit allt viktigare. En viktig del av språkteknologi är förmågan att sortera texter till det förflutna, nuet eller framtiden, beroende på när en händelse skedde eller kommer att ske. Syftet med denna avhandling var att använda textklassificering för att klassificera svenska meningar baserat på deras tempus, antingen dåtid, nutid eller framtid. Vidare var syftet även att jämföra hur lemmatisering skulle påverka modellernas prestanda. Problemet hanterades genom att implementera tre maskininlärningsmodeller på både lemmatiserade och icke lemmatiserade data. Maskininlärningsmodellerna var Naive Bayes, LSTM och BERT. Resultatet var att den övergripande prestandan påverkades negativt när datan lemmatiserade. Den bäst presterande modellen var BERT med en träffsäkerhet på 96,3 \%. Resultatet var användbart eftersom den bäst presterande modellen hade mycket hög träffsäkerhet och fungerade bra på nybyggda meningar.