Flood forecasting maps are essential for rapid disaster response and risk management, yet the computational complexity of physically-based simulations hinders their application for efficient high-resolution spatial flood forecasting. To address the problems of high computational cost and long prediction time, this thesis proposes to develop deep learning neural networks based on a flood simulation dataset, and explore their potential use for flood prediction without learning hydrological modelling knowledge from scratch.  A Fully Convolutional Network (FCN), FCN with multiple outputs (Multioutput FCN), UNet, Graph-based model and their Recurrent Neural Network (RNN) variants are trained on a catchment area with twelve rainfall events, and evaluated on two cases of a specific rainfall event both quantitatively and qualitatively. Among them, Convolution-based models (FCN, Multioutput FCN and UNet) are commonly used to solve problems related to spatial data but do not encode the position and orientation of objects, and Graph-based models can capture the structure of the problem but require higher time and space complexity. RNN-based models are effective for modelling time-series data, however, the computation is slow due to its recurrent nature. The results show that Multioutput FCN and the Graph-based model have significant advantages in predicting deep water depths (>50 cm), and the application of recurrent training greatly improves the long-term flood prediction accuracy of the base deep learning models. In addition, the proposed recurrent training FCN model performs the best and can provide flood predictions with high accuracy.

Title: Hybrid Deep Question Answering Author: Ahmad Aghaebrahimian Institute: Institute of Formal and Applied Linguistics Supervisor: RNDr. Martin Holub, Ph.D., Institute of Formal and Applied Lin- guistics Abstract: As one of the oldest tasks of Natural Language Processing, Question Answering is one of the most exciting and challenging research areas with lots of scientific and commercial applications. Question Answering as a discipline in the conjunction of computer science, statistics, linguistics, and cognitive science is concerned with building systems that automatically retrieve answers to ques- tions posed by humans in a natural language. This doctoral dissertation presents the author's research carried out in this discipline. It highlights his studies and research toward a hybrid Question Answering system consisting of two engines for Question Answering over structured and unstructured data. The structured engine comprises a state-of-the-art Question Answering system based on knowl- edge graphs. The unstructured engine consists of a state-of-the-art sentence-level Question Answering system and a word-level Question Answering system with results near to human performance. This work introduces a new Question An- swering dataset for answering word- and sentence-level questions as well. Start- ing from a...

The neural net computer simulations which will be presented here are based on the acceptance of a set of assumptions that for the last twenty years have been expressed in the fields of information processing, neurophysiology and cognitive sciences. First of all, neural networks and their dynamical behaviors in terms of attractors is the natural way adopted by the brain to encode information. Any information item to be stored in the neural net should be coded in some way or another in one of the dynamical attractors of the brain and retrieved by stimulating the net so as to trap its dynamics in the desired item's basin of attraction. The second view shared by neural net researchers is to base the learning of the synaptic matrix on a local Hebbian mechanism. The last assumption is the presence of chaos and the benefit gained by its presence. Chaos, although very simply produced, inherently possesses an infinite amount of cyclic regimes that can be exploited for coding information. Moreover, the network randomly wanders around these unstable regimes in a spontaneous way, thus rapidly proposing alternative responses to external stimuli and being able to easily switch from one of these potential attractors to another in response to any coming stimulus.<p><p>In this thesis, it is shown experimentally that the more information is to be stored in robust cyclic attractors, the more chaos appears as a regime in the back, erratically itinerating among brief appearances of these attractors. Chaos does not appear to be the cause but the consequence of the learning. However, it appears as an helpful consequence that widens the net's encoding capacity. To learn the information to be stored, an unsupervised Hebbian learning algorithm is introduced. By leaving the semantics of the attractors to be associated with the feeding data unprescribed, promising results have been obtained in term of storing capacity. / Doctorat en sciences appliquées / info:eu-repo/semantics/nonPublished

An automated manufacturing industry makes use of many interacting moving parts and sensors. Data from these sensors generate complex multidimensional data in the production environment. This data is difficult to interpret and also difficult to find patterns in. This project provides tools to get a deeper understanding of Swedsafe’s production data, a company involved in an automated manufacturing business. The project is based on and will show the potential of the multidimensional production data. The project mainly consists of predicting deviations from predefined threshold values in Swedsafe’s production data. Machine learning is a good method of finding relationships in complex datasets. Supervised machine learning classification is used to predict deviation from threshold values in the data. An investigation is conducted to identify the classifier that performs best on Swedsafe's production data. The technique sliding window is used for managing time series data, which is used in this project. Apart from predicting deviations, this project also includes an implementation of live graphs to easily get an overview of the production data. A steady production with stable process values is important. So being able to monitor and predict events in the production environment can provide the same benefit for other manufacturing companies and is therefore suitable not only for Swedsafe. The best performing machine learning classifier tested in this project was the Random Forest classifier. The Multilayer Perceptron did not perform well on Swedsafe’s data, but further investigation in recurrent neural networks using LSTM neurons would be recommended. During the projekt a web based application displaying the sensor data in live graphs is also developed.

Traffic congestion is a severe problem in urban areas and it leads to the emission of greenhouse gases and air pollution. In general, drivers lack knowledge of the location and availability of free parking spaces in urban cities. This leads to people driving around searching for parking places, and about one-third of traffic congestion in cities is due to drivers searching for an available parking lot. In recent years, various solutions to provide parking information ahead have been proposed. The vast majority of these solutions have been applied in large cities, such as Beijing and San Francisco. This thesis has been conducted in collaboration with Knowit and Dukaten to predict parking occupancy in car parks one hour ahead in the relatively small city of Linköping. To make the predictions, this study has investigated the possibility to use long short-term memory and gradient boosting regression trees, trained on historical parking data. To enhance decision making, the predictive uncertainty was estimated using the novel approach Monte Carlo dropout for the former, and quantile regression for the latter. This study reveals that both of the models can predict parking occupancy ahead of time and they are found to excel in different contexts. The inclusion of exogenous features can improve prediction quality. More specifically, we found that incorporating hour of the day improved the models’ performances, while weather features did not contribute much. As for uncertainty, the employed method Monte Carlo dropout was shown to be sensitive to parameter tuning to obtain good uncertainty estimates.

No description available.

This thesis examines two different approaches to parsing event-based vehicular signal data to produce input to a neural network prediction model: event parsing, where the data is kept unevenly spaced over the temporal domain, and slice parsing, where the data is made to be evenly spaced over the temporal domain instead. The dataset used as a basis for these experiments consists of a number of vehicular signal logs taken at Scania AB. Comparisons between the parsing methods have been made by first training long short-term memory (LSTM) recurrent neural networks (RNN) on each of the parsed datasets and then measuring the output error and resource costs of each such model after having validated them on a number of shared validation sets. The results from these tests clearly show that slice parsing compares favourably to event parsing. / Denna avhandling jämför två olika tillvägagångssätt vad gäller parsningen av händelsebaserad signaldata från fordon för att producera indata till en förutsägelsemodell i form av ett neuronnät, nämligen händelseparsning, där datan förblir ojämnt fördelad över tidsdomänen, och skivparsning, där datan är omgjord till att istället vara jämnt fördelad över tidsdomänen. Det dataset som används för dessa experiment är ett antal signalloggar från fordon som kommer från Scania. Jämförelser mellan parsningsmetoderna gjordes genom att först träna ett lång korttidsminne (LSTM) återkommande neuronnät (RNN) på vardera av de skapade dataseten för att sedan mäta utmatningsfelet och resurskostnader för varje modell efter att de validerats på en delad uppsättning av valideringsdata. Resultaten från dessa tester visar tydligt på att skivparsning står sig väl mot händelseparsning.

As an efficient energy storage technology, lithium-ion batteries play a key role in the ongoing electrification of the mobility sector. However, the required modelbased design process, including hardware in the loop solutions, demands precise battery models. In this work, an encoder-decoder model framework based on recurrent neural networks is developed and trained directly on unstructured battery data to replace time consuming characterisation tests and thus simplify the modelling process. A manifold pseudo-random bit stream dataset is used for model training and validation. A mean percentage error (MAPE) of 0.30% for the test dataset attests the proposed encoder-decoder model excellent generalisation capabilities. Instead of the recursive one-step prediction prevalent in the literature, the stage-wise trained encoder-decoder framework can instantaneously predict the battery voltage response for 2000 time steps and proves to be 120 times more time-efficient on the test dataset. Accuracy, generalisation capability and time efficiency of the developed battery model enable a potential online anomaly detection, power or range prediction. The fact that, apart from the initial voltage level, the battery model only relies on the current load as input and thus requires no estimated variables such as the state-of-charge (SOC) to predict the voltage response holds the potential of a battery ageing independent LIB modelling based on raw BMS signals. The intrinsically ageingindependent battery model is thus suitable to be used as a digital battery twin in virtual experiments to estimate the unknown battery SOH on purely BMS data basis.

本研究考慮重要且對台股大盤指數走勢有連動影響的因素，主要納入對台股有領頭作用的美國三大股市，那斯達克(NASDAQ)指數、道瓊工業(Dow Jones)指數、標準普爾500(S&P500)指數;其他對台股緊密連動效果的國際股票市場，香港恆生指數、上海證券綜合指數、深圳證券綜合指數、日經225指數;以及納入左右國際經濟表現的國際原油價格走勢，美國西德州原油、中東杜拜原油和歐洲北海布蘭特原油;在宏觀經濟因素方面則考量失業率、消費者物價指數、匯率、無風險利率、美國製造業重要指標的存貨/銷貨比率、影響貨幣數量甚鉅的M1B;在技術分析方面則納入多種重要的指標，心理線 (PSY) 指標、相對強弱(RSI) 指標、威廉(WMS%R) 指標、未成熟隨機(RSV) 指標、K-D隨機指標、移動平均線(MA)、乖離率(BIAS)、包寧傑%b和包寧傑帶狀寬度(BandWidth%);所有考量因素共計35項，因為納入重要因子比較多，所以完備性較高。 本研究先採用的贏者全拿(Winner-Take-All) 競爭學習策略的自組織映射網路(Self-Organizing Feature Maps, SOM)，藉由將相似資料歸屬到已身的神經元萃取出關聯分類且以計算距離來衡量神經元的離散特徵，對於探索大量且高維度的非線性複雜特徵俱有優良的因素相依性投射效果，將有利於提高預測模式精準度。在線性擬合部分則結合倒傳遞(Back-Propagation, BP)、Elman反饋式和徑向基底函數類網路(Radial-Basis-Function Network, RBF)模式為指數預測輸出，並對台股加權指數隔日收盤指數進行預測和評量。而在傳統的Elman反饋式網路只在隱藏層存在反饋機制，本研究則在輸入層和隱藏層皆建立反饋機制，將儲存在輸入層和隱藏層的過去時間資訊回饋給網路未來參考。在徑向基底函數網路方面，一般選取中心聚類點採用隨機選取方式，若能有效降低中心點個數，可降低網路複雜度，本研究導入垂直最小平方法以求取誤差最小的方式強化非監督式學習選取中心點的能力，以達到網路快速收斂，提昇網路學習品質。 研究資料為台股指數交易收盤價，日期自2001/1/2，至2011/10/31共2676筆資料。訓練資料自2001/1/2至2009/12/31，共2223筆;實證測試資料自2010/1/4至2011/10/31，計453個日數。主要評估指標採用平均相對誤差(AMRE)和平均絕對誤差 (AAE)。在考慮因子較多的狀況下，實證結果顯示，在先透過SOM進行因子聚類分析之後，預測因子被分成四個組別，分別再透過BP、Elman recurrent和RBF方法進行線性擬合，平均表現方面，以RBF模式下的四個群組因子表現最佳，其中RBF模式之下的群組4，其AMRE可達到0.63%，最差的AMRE則是群組1，約為1.05%;而Elman recurrent模式下的四組群組因子之ARME則介於1.01%和1.47%之間;其中預測效果表現最差則是BP模式的預測結果。顯示RBF具有絕佳的股價預測能力。最後，在未來研究建議可以運用本文獻所探討之其他數種類神經網路模式進行股價預測。 / In this study, we considering the impact factors for TWSE index tendency, mainly aimed at the three major American stock markets, NASDAQ index, Dow Jones index, S&P 500, which leading the Taiwan stock market trend; the other international stock markets, such as the Hong Kong Hang-Seng Index, Shanghai Stock Exchange Composite Index, Shenzhen Stock Exchange Composite Index, NIKKEI 225 index, which have close relationship with Taiwan stock market; we also adopt the international oil price trend, such as the West Texas Intermediate Crude Oil in American, the Dubai crude oil in Middle Eastern, North Sea Brent crude oil in European, which affects international economic performance widely; On the side of macroeconomic factors, we considering the Unemployed rate, Consumer Price Index, exchange rate, riskless rate, the Inventory to Sales ratio which it is important index of American manufacturing industry, and the M1b factor which did greatly affect to currency amounts; In the part of Technical Analysis index, we adopt several important indices, such as the Psychology Line Index (PSY), Relative Strength Index (RSI), the Wechsler Memory Scale—Revised Index (WMS%R), Row Stochastic Value Index (RSV), K-D Stochastics Index, Moving Average Line (MA), BIAS, Bollinger %b (%b), Bollinger Band Width (Band Width%);All factors total of 35 which we have considered the important factor is numerous, so the integrity is high. In this study, at first we adopt the Self-Organizing Feature Maps Network which based on the Winner-Take-All competition learning strategy, Similar information by the attribution to the body of the neuron has been extracted related categories and to calculate the distance to measure the discrete characteristics of neurons, it has excellent projection effect by exploring large and complex high-dimensional non-linear characteristics for all the dependency factors , would help to improve the accuracy of prediction models, would be able to help to improve the accuracy of prediction models. The part of the curve fitting combine with the back-propagation (Back-Propagation, BP), Elman recurrent model and radial basis function network (Radial-Basis-Function Network, RBF) model for the index prediction outputs, forecast and assessment the next close price of Taiwan stocks weighted index. In the traditional Elman recurrent network exists only one feedback mechanism in the hidden layer, in this study in the input and hidden layer feedback mechanisms are established, the previous information will be stored in the input and hidden layer and will be back to the network for future reference. In the radial basis function network, the general method is to selecting cluster center points by random selection, if we have the effectively way to reduce the number of the center points, which can reduces network complexity, in this study introduce the Orthogonal Least Squares method in order to obtain the smallest way to strengthen unsupervised learning center points selecting ability, in order to achieve convergence of the network fast, and improve network learning quality. Research data for the Trading close price of Taiwan Stock Index, the date since January 2, 2001 until September 30, 2011, total data number of 2656. since January 2, 2001 to December 31, 2009 a total number of 2223 trading close price as training data; empirical testing data, from January 4, 2010 to September 30, 2011, a total number of 433. The primary evaluation criteria adopt the Average Mean Relative Error (AMRE) and the Average Absolute Error (AAE). In the condition for consider more factors, the empirical results show that, by first through SOM for factor clustering analysis, the prediction factors were divided into four categories and then through BP, Elman recurrent and RBF methods for curve fitting, at the average performance , the four group factors of the RBF models get the best performance, the group 4 of the RBF model, the AMRE can reach 0.63%, the worst AMRE is group 1, about 1.05%; and the four groups of Elman recurrent model of ARME is between 1.01% and 1.47%; the worst prediction model is BP method. RBF has shown excellent predictive ability for stocks index. Finally, the proposal can be used in future studies of the literatures that we have explore several other methods of neural network model for stock trend forecasting.