Dissertations / Theses on the topic 'RNN NETWORK'
To see the other types of publications on this topic, follow the link: RNN NETWORK.
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the top 50 dissertations / theses for your research on the topic 'RNN NETWORK.'
Next to every source in the list of references, there is an 'Add to bibliography' button. Press on it, and we will generate automatically the bibliographic reference to the chosen work in the citation style you need: APA, MLA, Harvard, Chicago, Vancouver, etc.
You can also download the full text of the academic publication as pdf and read online its abstract whenever available in the metadata.
Browse dissertations / theses on a wide variety of disciplines and organise your bibliography correctly.
1
Bäärnhielm, Arvid. "Multiple time-series forecasting on mobile network data using an RNN-RBM model." Thesis, Uppsala universitet, Datalogi, 2017. http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-315782.
Full textAbstract:
The purpose of this project is to evaluate the performance of a forecasting model based on a multivariate dataset consisting of time series of traffic characteristic performance data from a mobile network. The forecasting is made using machine learning with a deep neural network. The first part of the project involves the adaption of the model design to fit the dataset and is followed by a number of simulations where the aim is to tune the parameters of the model to give the best performance. The simulations show that with well tuned parameters, the neural network performes better than the baseline model, even when using only a univariate dataset. If a multivariate dataset is used, the neural network outperforms the baseline model even when the dataset is small.
2
Vikström, Filip. "A recurrent neural network approach to quantification of risks surrounding the Swedish property market." Thesis, Umeå universitet, Institutionen för matematik och matematisk statistik, 2016. http://urn.kb.se/resolve?urn=urn:nbn:se:umu:diva-126192.
Full textAbstract:
As the real estate market plays a central role in a countries financial situation, as a life insurer, a bank and a property developer, Skandia wants a method for better assessing the risks connected to the real estate market. The goal of this paper is to increase the understanding of property market risk and its covariate risks and to conduct an analysis of how a fall in real estate prices could affect Skandia’s exposed assets.This paper explores a recurrent neural network model with the aim of quantifying identified risk factors using exogenous data. The recurrent neural network model is compared to a vector autoregressive model with exogenous inputs that represent economic conditions.The results of this paper are inconclusive as to which method that produces the most accurate model under the specified settings. The recurrent neural network approach produces what seem to be better results in out-of-sample validation but both the recurrent neural network model and the vector autoregressive model fail to capture the hypothesized relationship between the exogenous and modeled variables. However producing results that does not fit previous assumptions, further research into artificial neural networks and tests with additional variables and longer sample series for calibration is suggested as the model preconditions are promising.
3
Billingsley, Richard John. "Deep Learning for Semantic and Syntactic Structures." Thesis, The University of Sydney, 2014. http://hdl.handle.net/2123/12825.
Full textAbstract:
Deep machine learning has enjoyed recent success in vision and speech-to-text tasks, using deep multi-layered neural networks. They have obtained remarkable results particularly where the internal representation of the task is unclear. In parsing, where the structure of syntax is well studied and understood from linguistics, neural networks have so far not performed so well. State-of-the-art parsers use a tree-based graphical model that requires a large number of equivalent classes to represent each parse node and its phrase label. A recursive neural network (RNN) parser has been developed that works well on short sentences, but falls short of the state-of-the-art results on longer sentences. This thesis aims to investigate deep learning and improve parsing by examining how neural networks could perform state-of-the-art parsing by comparison with PCFG parsers. We hypothesize that a neural network could be configured to implement an algorithm parallel to PCFG parsers, and examine their suitability to this task from an analytic perspective. This highlights a missing term that the RNN parser is unable to model, and we identify the role of this missing term in parsing. We finally present two methods to improve the RNN parser by building upon the analysis in earlier chapters, one using an iterative process similar to belief propagation that yields a 0.38% improvement and another replacing the scoring method with a deeper neural model yielding a 0.83% improvement. By examining an RNN parser as an exemplar of a deep neural network, we gain insights to deep machine learning and some of the approximations it must make by comparing it with well studied non-neural parsers that achieve state-of-the-art results. In this way, our research provides a better understanding of deep machine learning and a step towards improvements in parsing that will lead to smarter algorithms that can learn more accurate representations of information and the syntax and semantics of text.
4
Liu, Chang. "Data Analysis of Minimally-Structured Heterogeneous Logs : An experimental study of log template extraction and anomaly detection based on Recurrent Neural Network and Naive Bayes." Thesis, KTH, Skolan för datavetenskap och kommunikation (CSC), 2016. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-191334.
Full textAbstract:
Nowadays, the ideas of continuous integration and continuous delivery are under heavy usage in order to achieve rapid software development speed and quick product delivery to the customers with good quality. During the process ofmodern software development, the testing stage has always been with great significance so that the delivered software is meeting all the requirements and with high quality, maintainability, sustainability, scalability, etc. The key assignment of software testing is to find bugs from every test and solve them. The developers and test engineers at Ericsson, who are working on a large scale software architecture, are mainly relying on the logs generated during the testing, which contains important information regarding the system behavior and software status, to debug the software. However, the volume of the data is too big and the variety is too complex and unpredictable, therefore, it is very time consuming and with great efforts for them to manually locate and resolve the bugs from such vast amount of log data. The objective of this thesis project is to explore a way to conduct log analysis efficiently and effectively by applying relevant machine learning algorithms in order to help people quickly detect the test failure and its possible causalities. In this project, a method of preprocessing and clusering original logs is designed and implemented in order to obtain useful data which can be fed to machine learning algorithms. The comparable log analysis, based on two machine learning algorithms - Recurrent Neural Network and Naive Bayes, is conducted for detecting the place of system failures and anomalies. Finally, relevant experimental results are provided and analyzed.
5
Li, Edwin. "LSTM Neural Network Models for Market Movement Prediction." Thesis, KTH, Skolan för elektroteknik och datavetenskap (EECS), 2018. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-231627.
Full textAbstract:
Interpreting time varying phenomena is a key challenge in the capital markets. Time series analysis using autoregressive methods has been carried out over the last couple of decades, often with reassuring results. However, such methods sometimes fail to explain trends and cyclical fluctuations, which may be characterized by long-range dependencies or even dependencies between the input features. The purpose of this thesis is to investigate whether recurrent neural networks with LSTM-cells can be used to capture these dependencies, and ultimately be used as a complement for index trading decisions. Experiments are made on different setups of the S&P-500 stock index, and two distinct models are built, each one being an improvement of the previous model. The first model is a multivariate regression model, and the second model is a multivariate binary classifier. The output of each model is used to reason about the future behavior of the index. The experiment shows for the configuration provided that LSTM RNNs are unsuitable for predicting exact values of daily returns, but gives satisfactory results when used to predict the direction of the movement.Att förstå och kunna förutsäga hur index varierar med tiden och andra parametrar är ett viktigt problem inom kapitalmarknader. Tidsserieanalys med autoregressiva metoder har funnits sedan årtionden tillbaka, och har oftast gett goda resultat. Dessa metoder saknar dock möjligheten att förklara trender och cykliska variationer i tidsserien, något som kan karaktäriseras av tidsvarierande samband, men även samband mellan parametrar som indexet beror utav. Syftet med denna studie är att undersöka om recurrent neural networks (RNN) med long short-term memory-celler (LSTM) kan användas för att fånga dessa samband, för att slutligen användas som en modell för att komplettera indexhandel. Experimenten är gjorda mot en modifierad S&P-500 datamängd, och två distinkta modeller har tagits fram. Den ena är en multivariat regressionsmodell för att förutspå exakta värden, och den andra modellen är en multivariat klassifierare som förutspår riktningen på nästa dags indexrörelse. Experimenten visar för den konfiguration som presenteras i rapporten att LSTM RNN inte passar för att förutspå exakta värden för indexet, men ger tillfredsställande resultat när modellen ska förutsäga indexets framtida riktning.
6
Ďuriš, Denis. "Detekce ohně a kouře z obrazového signálu." Master's thesis, Vysoké učení technické v Brně. Fakulta elektrotechniky a komunikačních technologií, 2020. http://www.nusl.cz/ntk/nusl-412968.
Full textAbstract:
This diploma thesis deals with the detection of fire and smoke from the image signal. The approach of this work uses a combination of convolutional and recurrent neural network. Machine learning models created in this work contain inception modules and blocks of long short-term memory. The research part describes selected models of machine learning used in solving the problem of fire detection in static and dynamic image data. As part of the solution, a data set containing videos and still images used to train the designed neural networks was created. The results of this approach are evaluated in conclusion.
7
Racette, Olsén Michael. "Electrocardiographic deviation detection : Using long short-term memory recurrent neural networks to detect deviations within electrocardiographic records." Thesis, Linnéuniversitetet, Institutionen för datavetenskap (DV), 2018. http://urn.kb.se/resolve?urn=urn:nbn:se:lnu:diva-76411.
Full textAbstract:
Artificial neural networks have been gaining attention in recent years due to theirimpressive ability to map out complex nonlinear relations within data. In this report,an attempt is made to use a Long short-term memory neural network for detectinganomalies within electrocardiographic records. The hypothesis is that if a neuralnetwork is trained on records of normal ECGs to predict future ECG sequences, it isexpected to have trouble predicting abnormalities not previously seen in the trainingdata. Three different LSTM model configurations were trained using records fromthe MIT-BIH Arrhythmia database. Afterwards the models were evaluated for theirability to predict previously unseen normal and anomalous sections. This was doneby measuring the mean squared error of each prediction and the uncertainty of over-lapping predictions. The preliminary results of this study demonstrate that recurrentneural networks with the use of LSTM units are capable of detecting anomalies.
8
Carman, Benjamin Andrew. "Translating LaTeX to Coq: A Recurrent Neural Network Approach to Formalizing Natural Language Proofs." Ohio University Honors Tutorial College / OhioLINK, 2021. http://rave.ohiolink.edu/etdc/view?acc_num=ouhonors161919616626269.
Full text
9
Ljungehed, Jesper. "Predicting Customer Churn Using Recurrent Neural Networks." Thesis, KTH, Skolan för datavetenskap och kommunikation (CSC), 2017. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-210670.
Full textAbstract:
Churn prediction is used to identify customers that are becoming less loyal and is an important tool for companies that want to stay competitive in a rapidly growing market. In retail, a dynamic definition of churn is needed to identify churners correctly. Customer Lifetime Value (CLV) is the monetary value of a customer relationship. No change in CLV for a given customer indicates a decrease in loyalty. This thesis proposes a novel approach to churn prediction. The proposed model uses a Recurrent Neural Network to identify churners based on Customer Lifetime Value time series regression. The results show that the model performs better than random. This thesis also investigated the use of the K-means algorithm as a replacement to a rule-extraction algorithm. The K-means algorithm contributed to a more comprehensive analytical context regarding the churn prediction of the proposed model.Illojalitet prediktering används för att identifiera kunder som är påväg att bli mindre lojala och är ett hjälpsamt verktyg för att ett företag ska kunna driva en konkurrenskraftig verksamhet. I detaljhandel behöves en dynamisk definition av illojalitet för att korrekt kunna identifera illojala kunder. Kundens livstidsvärde är ett mått på monetärt värde av en kundrelation. En avstannad förändring av detta värde indikerar en minskning av kundens lojalitet. Denna rapport föreslår en ny metod för att utföra illojalitet prediktering. Den föreslagna metoden består av ett återkommande neuralt nätverk som används för att identifiera illojalitet hos kunder genom att prediktera kunders livstidsvärde. Resultaten visar att den föreslagna modellen presterar bättre jämfört med slumpmässig metod. Rapporten undersöker också användningen av en k-medelvärdesalgoritm som ett substitut för en regelextraktionsalgoritm. K-medelsalgoritm bidrog till en mer omfattande analys av illojalitet predikteringen.
10
Смішний, Денис Миколайович. "Система прогнозування економічних показників." Master's thesis, КПІ ім. Ігоря Сікорського, 2019. https://ela.kpi.ua/handle/123456789/30950.
Full textAbstract:
Магістерська дисертація: 88 с., 20 рис., 27 табл., 1 додаток, 33 джерел.
Актуальність проблеми. Глобалізація та збільшення числа населення сприяють розвитку глобальної економіки, а отже — появі нових видів госпо-дарської діяльності та нових гравців на ринку праці. При реалізації власного підприємства важливо правильно оцінити ризики ринку, проаналізувавши та спробувавши спрогнозувати рух котирувань на найближчий час задля мінімальних фінансових втрат.
Зв’язок роботи з науковими програмами, планами, темами. Наразі, не має конкретних зв’язків з науковими програмами чи планами.
Мета і задачі дослідження. Завданням цієї роботи є дослідження мож-ливості прогнозування економічних параметрів підприємств на прикладі цін на акції компаній на фондовій біржі. Метою є розроблення системи, побудо-ваної на базі нейронної мережі, здатної проаналізувати задані економічні по-казники та, на основі отриманих даних спрогнозувати їхню динаміку.
Об’єкт дослідження. Процес прогнозування економічних показників з використанням елементів нейронної мережі.
Предмет дослідження. Методи аналізу та обробки економічних даних за певний період.
Новизна. Отримання програмного продукту, що здатний прогнозувати коливання економічних показників. Дослідження можливості реалізації мо-делі на основі нейронної мережі для виконання поставленої мети та завдань.Master's Thesis: 88 pp., 20 figs., 27 tables, 1 appendix, 33 sources. The urgency of the problem. Globalization and population growth are con-tributing to the development of the global economy and, consequently, to the emergence of new types of economic activity and new players in the labor market. When implementing your own business it is important to properly evaluate the risks of the market, analyzing and trying to predict the movement of quotations in the near future for minimal financial losses. Relationship with working with scientific programs, plans, topics. Cur-rently, it has no specific links to scientific programs or plans. The purpose and objectives of the study. The purpose of this work is re-search possibility of forecasting the economic parameters of enterprises on the ex-ample of stock prices of companies on the stock exchange. The purpose is to de-velop a system based on a neural network, capable of analyzing specified economic indicators and, based on the data obtained, to predict their dynamics. Object of study. The process of forecasting economic performance using neural network elements. Subject of study. Methods of analysis and processing of economic data for a certain period. Novelty. Obtaining a software product capable of predicting economic fluc-tuations. Investigation of the possibility of creating a universal model based on a neural network, which would not require specialization and would be able to work effectively with any set of input data without further training.
11
Wen, Tsung-Hsien. "Recurrent neural network language generation for dialogue systems." Thesis, University of Cambridge, 2018. https://www.repository.cam.ac.uk/handle/1810/275648.
Full textAbstract:
Language is the principal medium for ideas, while dialogue is the most natural and effective way for humans to interact with and access information from machines. Natural language generation (NLG) is a critical component of spoken dialogue and it has a significant impact on usability and perceived quality. Many commonly used NLG systems employ rules and heuristics, which tend to generate inflexible and stylised responses without the natural variation of human language. However, the frequent repetition of identical output forms can quickly make dialogue become tedious for most real-world users. Additionally, these rules and heuristics are not scalable and hence not trivially extensible to other domains or languages. A statistical approach to language generation can learn language decisions directly from data without relying on hand-coded rules or heuristics, which brings scalability and flexibility to NLG. Statistical models also provide an opportunity to learn in-domain human colloquialisms and cross-domain model adaptations. A robust and quasi-supervised NLG model is proposed in this thesis. The model leverages a Recurrent Neural Network (RNN)-based surface realiser and a gating mechanism applied to input semantics. The model is motivated by the Long-Short Term Memory (LSTM) network. The RNN-based surface realiser and gating mechanism use a neural network to learn end-to-end language generation decisions from input dialogue act and sentence pairs; it also integrates sentence planning and surface realisation into a single optimisation problem. The single optimisation not only bypasses the costly intermediate linguistic annotations but also generates more natural and human-like responses. Furthermore, a domain adaptation study shows that the proposed model can be readily adapted and extended to new dialogue domains via a proposed recipe. Continuing the success of end-to-end learning, the second part of the thesis speculates on building an end-to-end dialogue system by framing it as a conditional generation problem. The proposed model encapsulates a belief tracker with a minimal state representation and a generator that takes the dialogue context to produce responses. These features suggest comprehension and fast learning. The proposed model is capable of understanding requests and accomplishing tasks after training on only a few hundred human-human dialogues. A complementary Wizard-of-Oz data collection method is also introduced to facilitate the collection of human-human conversations from online workers. The results demonstrate that the proposed model can talk to human judges naturally, without any difficulty, for a sample application domain. In addition, the results also suggest that the introduction of a stochastic latent variable can help the system model intrinsic variation in communicative intention much better.
12
Abdulaziz, Ali Haseeb Mohamed. "Passive gesture recognition on unmodified smartphones using Wi-Fi RSSI." Thesis, KTH, Robotik, perception och lärande, RPL, 2017. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-216390.
Full textAbstract:
The smartphone is becoming a common device carried by hundreds of millions of individual humans worldwide, and is used to accomplish a multitude of different tasks like basic communication, internet browsing, online shopping and fitness tracking. Limited by its small size and tight energy storage, the human-smartphone interface is largely bound to the smartphones small screens and simple keypads. This prohibits introducing new rich ways of interaction with smartphones. The industry and research community are working extensively to find ways to enrich the human-smartphone interface by either seizing the existing smartphones resources like microphones, cameras and inertia sensors, or by introducing new specialized sensing capabilities into the smartphones like compact gesture sensing radar devices. The prevalence of Radio Frequency (RF) signals and their limited power needs, led us towards investigating using RF signals received by smartphones to recognize gestures and activities around smartphones. This thesis introduces a solution for recognizing touch-less dynamic hand gestures from the Wi-Fi Received Signal Strength (RSS) received by the smartphone using a recurrent neural network (RNN) based probabilistic model. Unlike other Wi-Fi based gesture recognition solutions, the one introduced in this thesis does not require a change to the smartphone hardware or operating system, and performs the hand gesture recognition without interfering with the normal operation of other smartphone applications. The developed hand gesture recognition solution achieved a mean accuracy of 78% detecting and classifying three hand gestures in an online setting involving different spatial and traffic scenarios between the smartphone and Wi-Fi access points (AP). Furthermore the characteristics of the developed solution were studied, and a set of improvements have been suggested for further future work.Smarta telefoner bärs idag av hundratals miljoner människor runt om i världen, och används för att utföra en mängd olika uppgifter, så som grundläggande kommunikation, internetsökning och online-inköp. På grund av begränsningar i storlek och energilagring är människa-telefon-gränssnitten dock i hög grad begränsade till de förhållandevis små skärmarna och enkla knappsatser. Industrin och forskarsamhället arbetar för att hitta vägar för att förbättra och bredda gränssnitten genom att antingen använda befintliga resurser såsom mikrofoner, kameror och tröghetssensorer, eller genom att införa nya specialiserade sensorer i telefonerna, som t.ex. kompakta radarenheter för gestigenkänning. Det begränsade strömbehovet hos radiofrekvenssignaler (RF) inspirerade oss till att undersöka om dessa kunde användas för att känna igen gester och aktiviteter i närheten av telefoner. Denna rapport presenterar en lösning för att känna igen gester med hjälp av ett s.k. recurrent neural network (RNN). Till skillnad från andra Wi-Fi-baserade lösningar kräver denna lösning inte en förändring av vare sig hårvara eller operativsystem, och ingenkänningen genomförs utan att inverka på den normala driften av andra applikationer på telefonen. Den utvecklade lösningen når en genomsnittlig noggranhet på 78% för detektering och klassificering av tre olika handgester, i ett antal olika konfigurationer vad gäller telefon och Wi-Fi-sändare. Rapporten innehåller även en analys av flera olika egenskaper hos den föreslagna lösningen, samt förslag till vidare arbete.
13
Fagerholm, Christian. "Time series analysis and forecasting : Application to the Swedish Power Grid." Thesis, Linnéuniversitetet, Institutionen för datavetenskap och medieteknik (DM), 2019. http://urn.kb.se/resolve?urn=urn:nbn:se:lnu:diva-88615.
Full textAbstract:
n the electrical power grid, the power load is not constant but continuouslychanging. This depends on many different factors, among which the habits of theconsumers, the yearly seasons and the hour of the day. The continuous change inenergy consumption requires the power grid to be flexible. If the energy provided bygenerators is lower than the demand, this is usually compensated by using renewablepower sources or stored energy until the power generators have adapted to the newdemand. However, if buffers are depleted the output may not meet the demandedpower and could cause power outages. The currently adopted practice in the indus-try is based on configuring the grid depending on some expected power draw. Thisanalysis is usually performed at a high level and provide only some basic load aggre-gate as an output. In this thesis, we aim at investigating techniques that are able topredict the behaviour of loads with fine-grained precision. These techniques couldbe used as predictors to dynamically adapt the grid at run time. We have investigatedthe field of time series forecasting and evaluated and compared different techniquesusing a real data set of the load of the Swedish power grid recorded hourly throughyears. In particular, we have compared the traditional ARIMA models to a neuralnetwork and a long short-term memory (LSTM) model to see which of these tech-niques had the lowest forecasting error in our scenario. Our results show that theLSTM model outperformed the other tested models with an average error of 6,1%.
14
Hamerník, Pavel. "Využití hlubokého učení pro rozpoznání textu v obrazu grafického uživatelského rozhraní." Master's thesis, Vysoké učení technické v Brně. Fakulta informačních technologií, 2019. http://www.nusl.cz/ntk/nusl-403823.
Full textAbstract:
Optical character recognition (OCR) has been a topic of interest for many years. It is defined as the process of digitizing a document image into a sequence of characters. Despite decades of intense research, OCR systems with capabilities to that of human still remains an open challenge. In this work there is presented a design and implementation of such system, which is capable of detecting texts in graphical user interfaces.
15
Ridhagen, Markus, and Petter Lind. "A comparative study of Neural Network Forecasting models on the M4 competition data." Thesis, Uppsala universitet, Statistiska institutionen, 2021. http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-445568.
Full textAbstract:
The development of machine learning research has provided statistical innovations and further developments within the field of time series analysis. This study seeks to investigate two different approaches on artificial neural network models based on different learning techniques, and answering how well the neural network approach compares with a basic autoregressive approach, as well as how the artificial neural network models compare to each other. The models were compared and analyzed in regards to the univariate forecast accuracy on 20 randomly drawn time series from two different time frequencies from the M4 competition dataset. Forecasting was made dependent on one time lag (t-1) and forecasted three and six steps ahead respectively. The artificial neural network models outperformed the baseline Autoregressive model, showing notably lower mean average percentage error overall. The Multilayered perceptron models performed better than the Long short-term memory model overall, whereas the Long short-term memory model showed improvement on longer prediction time dimensions. As the training were done univariately on a limited set of time steps, it is believed that the one layered-approach gave a good enough approximation on the data, whereas the added layer couldn’t fully utilize its strengths of processing power. Likewise, the Long short-term memory model couldn’t fully demonstrate the advantagements of recurrent learning. Using the same dataset, further studies could be made with another approach to data processing. Implementing an unsupervised approach of clustering the data before analysis, the same models could be tested with multivariate analysis on models trained on multiple time series simultaneously.
16
Almqvist, Olof. "A comparative study between algorithms for time series forecasting on customer prediction : An investigation into the performance of ARIMA, RNN, LSTM, TCN and HMM." Thesis, Högskolan i Skövde, Institutionen för informationsteknologi, 2019. http://urn.kb.se/resolve?urn=urn:nbn:se:his:diva-16974.
Full textAbstract:
Time series prediction is one of the main areas of statistics and machine learning. In 2018 the two new algorithms higher order hidden Markov model and temporal convolutional network were proposed and emerged as challengers to the more traditional recurrent neural network and long-short term memory network as well as the autoregressive integrated moving average (ARIMA). In this study most major algorithms together with recent innovations for time series forecasting is trained and evaluated on two datasets from the theme park industry with the aim of predicting future number of visitors. To develop models, Python libraries Keras and Statsmodels were used. Results from this thesis show that the neural network models are slightly better than ARIMA and the hidden Markov model, and that the temporal convolutional network do not perform significantly better than the recurrent or long-short term memory networks although having the lowest prediction error on one of the datasets. Interestingly, the Markov model performed worse than all neural network models even when using no independent variables.
17
Kišš, Martin. "Rozpoznávání historických textů pomocí hlubokých neuronových sítí." Master's thesis, Vysoké učení technické v Brně. Fakulta informačních technologií, 2018. http://www.nusl.cz/ntk/nusl-385912.
Full textAbstract:
The aim of this work is to create a tool for automatic transcription of historical documents. The work is mainly focused on the recognition of texts from the period of modern times written using font Fraktur. The problem is solved with a newly designed recurrent convolutional neural networks and a Spatial Transformer Network. Part of the solution is also an implemented generator of artificial historical texts. Using this generator, an artificial data set is created on which the convolutional neural network for line recognition is trained. This network is then tested on real historical lines of text on which the network achieves up to 89.0 % of character accuracy. The contribution of this work is primarily the newly designed neural network for text line recognition and the implemented artificial text generator, with which it is possible to train the neural network to recognize real historical lines of text.
18
Yazdi, Mohammad Hamed. "Software Test Strategies for the RNC RNH Subsystem." Thesis, KTH, Kommunikationsnät, 2013. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-116599.
Full textAbstract:
This work concerns software testing strategies for the Radio Network Controller (RNC) RadioNetwork Handler (RNH) subsystem at the WCDMA development department at Ericsson AB. Due to the rapid development in the area of radio communication It is crucial to constantly develop and deliver new software components without errors in the code, which has to be tested and proved to work on a regular basis. Since development teams are working in parallel, one cannot uphold another team for long periods for testing purposes. It should be easy and straightforward to implement and maintain RNH tests. The main goal is to propose the best way of software testing for the RNH subsystem with respect to the agile way of working. In the first part of this work an investigation of the RNH software was done. This was to define a template for code classification. The aim of the classification is to identify a smallest testable unit for different testing levels. The data classes were considered as smallest testable unit for testing on low level. In the second part, unit test was deployed to two different blocks to evaluate unit testing and prove testability of data classes on a low level. In addition, the automated regression test framework was evaluated with respect to node level testing performance. In the third part, unit test was evaluated in comparison to the current testing level at RNH. The major result of this investigation shows all testing levels are required for the RNH subsystem, because each level focuses on a specific area of software testing. Furthermore, unit testing is recommended to be a permanent testing level at RNH subsystem, since unit testing is promoted by agile testing strategies (test as early as possible). Besides, when more detailed test on low level (unit testing) is applied, it may lead to the less testing effort on higher level.
19
Andersson, Aron, and Shabnam Mirkhani. "Portfolio Performance Optimization Using Multivariate Time Series Volatilities Processed With Deep Layering LSTM Neurons and Markowitz." Thesis, KTH, Matematisk statistik, 2020. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-273617.
Full textAbstract:
The stock market is a non-linear field, but many of the best-known portfolio optimization algorithms are based on linear models. In recent years, the rapid development of machine learning has produced flexible models capable of complex pattern recognition. In this paper, we propose two different methods of portfolio optimization; one based on the development of a multivariate time-dependent neural network,thelongshort-termmemory(LSTM),capable of finding lon gshort-term price trends. The other is the linear Markowitz model, where we add an exponential moving average to the input price data to capture underlying trends. The input data to our neural network are daily prices, volumes and market indicators such as the volatility index (VIX).The output variables are the prices predicted for each asset the following day, which are then further processed to produce metrics such as expected returns, volatilities and prediction error to design a portfolio allocation that optimizes a custom utility function like the Sharpe Ratio. The LSTM model produced a portfolio with a return and risk that was close to the actual market conditions for the date in question, but with a high error value, indicating that our LSTM model is insufficient as a sole forecasting tool. However,the ability to predict upward and downward trends was somewhat better than expected and therefore we conclude that multiple neural network can be used as indicators, each responsible for some specific aspect of what is to be analysed, to draw a conclusion from the result. The findings also suggest that the input data should be more thoroughly considered, as the prediction accuracy is enhanced by the choice of variables and the external information used for training.Aktiemarknaden är en icke-linjär marknad, men många av de mest kända portföljoptimerings algoritmerna är baserad på linjära modeller. Under de senaste åren har den snabba utvecklingen inom maskininlärning skapat flexibla modeller som kan extrahera information ur komplexa mönster. I det här examensarbetet föreslår vi två sätt att optimera en portfölj, ett där ett neuralt nätverk utvecklas med avseende på multivariata tidsserier och ett annat där vi använder den linjära Markowitz modellen, där vi även lägger ett exponentiellt rörligt medelvärde på prisdatan. Ingångsdatan till vårt neurala nätverk är de dagliga slutpriserna, volymerna och marknadsindikatorer som t.ex. volatilitetsindexet VIX. Utgångsvariablerna kommer vara de predikterade priserna för nästa dag, som sedan bearbetas ytterligare för att producera mätvärden såsom förväntad avkastning, volatilitet och Sharpe ratio. LSTM-modellen producerar en portfölj med avkastning och risk som ligger närmre de verkliga marknadsförhållandena, men däremot gav resultatet ett högt felvärde och det visar att vår LSTM-modell är otillräckligt för att använda som ensamt predikteringssverktyg. Med det sagt så gav det ändå en bättre prediktion när det gäller trender än vad vi antog den skulle göra. Vår slutsats är därför att man bör använda flera neurala nätverk som indikatorer, där var och en är ansvarig för någon specifikt aspekt man vill analysera, och baserat på dessa dra en slutsats. Vårt resultat tyder också på att inmatningsdatan bör övervägas mera noggrant, eftersom predikteringsnoggrannheten.
20
Arvidsson, Philip, and Tobias Ånhed. "Sequence-to-sequence learning of financial time series in algorithmic trading." Thesis, Högskolan i Borås, Akademin för bibliotek, information, pedagogik och IT, 2017. http://urn.kb.se/resolve?urn=urn:nbn:se:hb:diva-12602.
Full textAbstract:
Predicting the behavior of financial markets is largely an unsolved problem. The problem hasbeen approached with many different methods ranging from binary logic, statisticalcalculations and genetic algorithms. In this thesis, the problem is approached with a machinelearning method, namely the Long Short-Term Memory (LSTM) variant of Recurrent NeuralNetworks (RNNs). Recurrent neural networks are artificial neural networks (ANNs)—amachine learning algorithm mimicking the neural processing of the mammalian nervoussystem—specifically designed for time series sequences. The thesis investigates the capabilityof the LSTM in modeling financial market behavior as well as compare it to the traditionalRNN, evaluating their performances using various measures.Prediktion av den finansiella marknadens beteende är i stort ett olöst problem. Problemet hartagits an på flera sätt med olika metoder så som binär logik, statistiska uträkningar ochgenetiska algoritmer. I den här uppsatsen kommer problemet undersökas medmaskininlärning, mer specifikt Long Short-Term Memory (LSTM), en variant av rekurrentaneurala nätverk (RNN). Rekurrenta neurala nätverk är en typ av artificiellt neuralt nätverk(ANN), en maskininlärningsalgoritm som ska efterlikna de neurala processerna hos däggdjursnervsystem, specifikt utformat för tidsserier. I uppsatsen undersöks kapaciteten hos ett LSTMatt modellera finansmarknadens beteenden och jämförs den mot ett traditionellt RNN, merspecifikt mäts deras effektivitet på olika vis.
21
Talevi, Luca, and Luca Talevi. "“Decodifica di intenzioni di movimento dalla corteccia parietale posteriore di macaco attraverso il paradigma Deep Learning”." Master's thesis, Alma Mater Studiorum - Università di Bologna, 2019. http://amslaurea.unibo.it/17846/.
Full textAbstract:
Le Brain Computer Interfaces (BCI) invasive permettono di restituire la mobilità a pazienti che hanno perso il controllo degli arti: ciò avviene attraverso la decodifica di segnali bioelettrici prelevati da aree corticali di interesse al fine di guidare un arto prostetico. La decodifica dei segnali neurali è quindi un punto critico nelle BCI, richiedendo lo sviluppo di algoritmi performanti, affidabili e robusti.
Tali requisiti sono soddisfatti in numerosi campi dalle Deep Neural Networks, algoritmi adattivi le cui performance scalano con la quantità di dati forniti, allineandosi con il crescente numero di elettrodi degli impianti.
Impiegando segnali pre-registrati dalla corteccia di due macachi durante movimenti di reach-to-grasp verso 5 oggetti differenti, ho testato tre basilari esempi notevoli di DNN – una rete densa multistrato, una Convolutional Neural Network (CNN) ed una Recurrent NN (RNN) – nel compito di discriminare in maniera continua e real-time l’intenzione di movimento verso ciascun oggetto. In particolare, è stata testata la capacità di ciascun modello di decodificare una generica intenzione (single-class), la performance della migliore rete risultante nel discriminarle (multi-class) con o senza metodi di ensemble learning e la sua risposta ad un degrado del segnale in ingresso. Per agevolarne il confronto, ciascuna rete è stata costruita e sottoposta a ricerca iperparametrica seguendo criteri comuni.
L’architettura CNN ha ottenuto risultati particolarmente interessanti, ottenendo F-Score superiori a 0.6 ed AUC superiori a 0.9 nel caso single-class con metà dei parametri delle altre reti e tuttavia maggior robustezza. Ha inoltre mostrato una relazione quasi-lineare con il degrado del segnale, priva di crolli prestazionali imprevedibili.
Le DNN impiegate si sono rivelate performanti e robuste malgrado la semplicità, rendendo eventuali architetture progettate ad-hoc promettenti nello stabilire un nuovo stato dell’arte nel controllo neuroprotesico.
22
Max, Lindblad. "The impact of parsing methods on recurrent neural networks applied to event-based vehicular signal data." Thesis, KTH, Skolan för elektroteknik och datavetenskap (EECS), 2018. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-223966.
Full textAbstract:
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.
23
Gu, Chenghong. "Long-run network pricing for security of supply in distribution networks." Thesis, University of Bath, 2010. https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.527128.
Full text
24
Thainesh, Joseph S. "Radio access network (RAN) signalling architecture for dense mobile network." Thesis, University of Surrey, 2016. http://epubs.surrey.ac.uk/811126/.
Full textAbstract:
Small cells are becoming a promising solution for providing enhanced coverage and increasing system capacity in a large-scale small cell network. In such a network, the large number of small cells may cause mobility signalling overload on the core network (CN) due to frequent handovers, which impact the users Quality of Experience (QoE). This is one of the major challenges in dense small cell networks. Such a challenge has been considered, this thesis addresses this challenging task to design an effective signalling architecture in dense small cell networks. First, in order to reduce the signalling overhead incurred by path switching operations in the small cell network, a new mobility control function, termed the Small Cell Controller (SCC) is introduced to the existing base station (BS) on the Radio-Access-Network(RAN)-side. Based on the signalling architecture, a clustering optimisation algorithm is proposed in order to select the optimal SCC in a highly user density environment. Specifically, this algorithm is designed to select multiple optimal SCCs due to the growth in number of small cells in the large-scale environment. Finally, a scalable architecture for handling the control plane failures in heterogeneous networks is proposed. In that architecture, the proposed SCC scheme controls and manages the affected small cells in a clustered fashion during the macro cell fail-over period. Particularly, the proposed SCC scheme can be flexibly configured into a hybrid scenario. For operational reduction (reducing a number of direct S1 connections to the CN), better scalability (reducing a number of S1 bearers on the CN) and reduction of signalling load on the CN, the proposed radio access network (RAN) signalling architecture is a viable and preferable option for dense small cell networks. Besides, the proposed signalling architecture is evaluated through realistic simulation studies.
25
Nilsson, Mathias, and Corswant Sophie von. "How Certain Are You of Getting a Parking Space? : A deep learning approach to parking availability prediction." Thesis, Linköpings universitet, Institutionen för datavetenskap, 2020. http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-166989.
Full textAbstract:
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.
26
Hartvigsen, Thomas. "Adaptively-Halting RNN for Tunable Early Classification of Time Series." Digital WPI, 2018. https://digitalcommons.wpi.edu/etd-theses/1257.
Full textAbstract:
Early time series classification is the task of predicting the class label of a time series before it is observed in its entirety. In time-sensitive domains where information is collected over time it is worth sacrificing some classification accuracy in favor of earlier predictions, ideally early enough for actions to be taken. However, since accuracy and earliness are contradictory objectives, a solution to this problem must find a task-dependent trade-off. There are two common state-of-the-art methods. The first involves an analyst selecting a timestep at which all predictions must be made. This does not capture earliness on a case-by-case basis, so if the selecting timestep is too early, all later signals are missed, and if a signal happens early, the classifier still waits to generate a prediction. The second method is the exhaustive search for signals, which encodes no timing information and is not scalable to high dimensions or long time series. We design the first early classification model called EARLIEST to tackle this multi-objective optimization problem, jointly learning (1) to decide at which time step to halt and generate predictions and (2) how to classify the time series. Each of these is learned based on the task and data features. We achieve an analyst-controlled balance between the goals of earliness and accuracy by pairing a recurrent neural network that learns to classify time series as a supervised learning task with a stochastic controller network that learns a halting-policy as a reinforcement learning task. The halting-policy dictates sequential decisions, one per timestep, of whether or not to halt the recurrent neural network and classify the time series early. This pairing of networks optimizes a global objective function that incorporates both earliness and accuracy. We validate our method via critical clinical prediction tasks in the MIMIC III database from the Beth Israel Deaconess Medical Center along with another publicly available time series classification dataset. We show that EARLIEST out-performs two state-of-the-art LSTM-based early classification methods. Additionally, we dig deeper into our model's performance using a synthetic dataset which shows that EARLIEST learns to halt when it observes signals without having explicit access to signal locations. The contributions of this work are three-fold. First, our method is the first neural network-based solution to early classification of time series, bringing the recent successes of deep learning to this problem. Second, we present the first reinforcement-learning based solution to the unsupervised nature of early classification, learning the underlying distributions of signals without access to this information through trial and error. Third, we propose the first joint-optimization of earliness and accuracy, allowing learning of complex relationships between these contradictory goals.
27
Sulieman, Nabeel Ibrahim. "Diversity and Network Coded 5G Wireless Network Infrastructure for Ultra-Reliable Communications." Scholar Commons, 2019. https://scholarcommons.usf.edu/etd/7961.
Full textAbstract:
This dissertation is directed towards improving the performance of 5G Wireless Fronthaul Networks and Wireless Sensor Networks, as measured by reliability, fault recovery time, energy consumption, efficiency, and security of transmissions, beyond what is achievable with conventional error control technology. To achieve these ambitious goals, the research is focused on novel applications of networking techniques, such as Diversity Coding, where a feedforward network design uses forward error control across spatially diverse paths to enable reliable wireless networking with minimal delay, in a wide variety of application scenarios. These applications include Cloud-Radio Access Networks (C-RANs), which is an emerging 5G wireless network architecture, where Remote Radio Heads (RRHs) are connected to the centralized Baseband Unit (BBU) via fronthaul networks, to enable near-instantaneous recovery from link/node failures. In addition, the ability of Diversity Coding to recover from multiple simultaneous link failures is demonstrated in many network scenarios. Furthermore, the ability of Diversity Coding to enable significantly simpler and thus lower-cost routing than other types of restoration techniques is demonstrated.
Achieving high throughput for broadcasting/multicasting applications, with the required level of reliability is critical for the efficient operation of 5G wireless infrastructure networks. To improve the performance of C-RAN networks, a novel technology, Diversity and Network Coding (DC-NC), which synergistically combines Diversity Coding and Network Coding, is introduced. Application of DC-NC to several 5G fronthaul networks, enables these networks to provide high throughput and near-instant recovery in the presence of link and node failures. Also, the application of DC-NC coding to enhance the performance of downlink Joint Transmission-Coordinated Multi Point (JT-CoMP) in 5G wireless fronthaul C-RANs is demonstrated. In all these scenarios, it is shown that DC-NC coding can provide efficient transmission and reduce the resource consumption in the network by about one-third for broadcasting/multicasting applications, while simultaneously enabling near-instantaneous latency in recovery from multiple link/node failures in fronthaul networks. In addition, it is shown by applying the DC-NC coding, the number of redundant links that uses to provide the required level of reliability, which is an important metric to evaluate any protection system, is reduced by about 30%-40% when compared to that of Diversity Coding.
With the additional goal of further reducing of the recovery time from multiple link/node failures and maximizing the network reliability, DC-NC coding is further improved to be able to tolerate multiple, simultaneous link failures with less computational complexity and lower energy consumption. This is accomplished by modifying Triangular Network Coding (TNC) and synergistically combining TNC with Diversity Coding to create enhanced DC-NC (eDC-NC), that is applied to Fog computing-based Radio Access Networks (F-RAN) and Wireless Sensor Networks (WSN). Furthermore, it is demonstrated that the redundancy percentage for protecting against n link failures is inversely related to the number of source data streams, which illustrates the scalability of eDC-NC coding. Solutions to enable synchronized broadcasting are proposed for different situations.
The ability of eDC-NC coding scheme to provide efficient and secure broadcasting for 5G wireless F-RAN fronthaul networks is also demonstrated. The security of the broadcasting data streams can be obtained more efficiently than standardized methods such as Secure Multicasting using Secret (Shared) Key Cryptography.
28
Lyazidi, Mohammed Yazid. "Dynamic resource allocation and network optimization in the Cloud Radio Access Network." Thesis, Paris 6, 2017. http://www.theses.fr/2017PA066549/document.
Full textAbstract:
Le Cloud Radio Access Network (C-RAN) est une future direction dans les réseaux de communications sans fils pour déployer des systèmes cellulaires 4G et renforcer la migration des opérateurs vers la nouvelle génération 5G. En comparaison avec l'architecture traditionnelle des stations de base distribuées, l'architecture C-RAN apporte un lot d'avantages à l'opérateur: meilleure utilisation des ressources radio, flexibilité du réseau, minimisation de la puissance consommée et amenuisement des coûts de déploiement. Dans cette thèse, nous adressons le problème d'allocation dynamique des ressources et minimisation de la puissance des communications à liaison descendante dans le C-RAN. Notre recherche vise à allouer les ressources radio à des flux dynamiques d'utilisateurs, tout en trouvant les meilleures combinaisons entre points d'accès et unités de calculs, pour satisfaire la demande de trafic. Il s'agit en outre, d'un problème d'optimisation non linéaire et NP-difficile, comprenant plusieurs contraintes relatives aux demandes de ressources des utilisateurs, gestion d'interférences, capacités fixes des unités de calcul dans le Cloud et des liaisons de transport ainsi que la limitation de la puissance transmise maximale. Afin de surmonter la complexité inhérente à cette problématique du C-RAN, nous présentons différentes approches pour l'allocation dynamique des ressources en trois principales contributions. Les résultats de nos simulations prouvent l'efficacité de nos méthodes, comparé à celles existantes dans la littérature, en termes de taux de débit de satisfaction, nombre d'antennes actives, puissance consommée dans le Cloud, résilience et coût opérationnel du C-RANCloud Radio Access Network (C-RAN) is a future direction in wireless communications for deploying cellular radio access subsystems in current 4G and next-generation 5G networks. In the C-RAN architecture, BaseBand Units (BBUs) are located in a pool of virtual base stations, which are connected via a high-bandwidth low latency fronthaul network to Radio Remote Heads (RRHs). In comparison to standalone clusters of distributed radio base stations, C-RAN architecture provides significant benefits in terms of centralized resource pooling, network flexibility and cost savings. In this thesis, we address the problem of dynamic resource allocation and power minimization in downlink communications for C-RAN. Our research aims to allocate baseband resources to dynamic flows of mobile users, while properly assigning RRHs to BBUs to accommodate the traffic and network demands. This is a non-linear NP-hard optimization problem, which encompasses many constraints such as mobile users' resources demands, interference management, BBU pool and fronthaul links capacities, as well as maximum transmission power limitation. To overcome the high complexity involved in this problem, we present several approaches for resource allocation strategies and tackle this issue in three stages. Obtained results prove the efficiency of our proposed strategies in terms of throughput satisfaction rate, number of active RRHs, BBU pool processing power, resiliency, and operational budget cost
29
Cid, Samper Fernando 1991. "Computational approaches to characterize RNP granules." Doctoral thesis, Universitat Pompeu Fabra, 2020. http://hdl.handle.net/10803/668449.
Full textAbstract:
Ribonucleoprotein granules (RNP granules) are liquid-liquid phase separated
complexes composed mainly by proteins and RNA. They are responsible of many
processes involved in RNA regulation. Alterations in the dynamics of these proteinRNA complexes are associated with the appearance of several neurodegenerative
disorders such as Amyotrophic Lateral Sclerosis ALS or Fragile X Tremor Ataxia
Syndrome FXTAS. Yet, many aspects of their organization as well as the specific
roles of the RNA on the formation and function of these complexes are still unknown.
In order to study RNP granules structure and formation, we integrated several state of
the art high-throughput datasets. This includes protein and RNA composition obtained
from RNP pull-downs, protein-RNA interaction data from eCLIP experiments and
transcriptome-wide secondary structure information (produced by PARS). We used
network analysis and clustering algorithms to understand the fundamental properties
of granule RNAs. By integrating these properties, we produced a model to identify
scaffolding RNA. Scaffolding RNAs are able to recruit many protein components into
RNP granules. We found that the main protein components of stress granules (a kind
of RNP granules) are connected through protein-RNA interactions. We also analyzed
the contribution of RNA-RNA interactions and RNA post-transcriptional
modifications on the granule internal organization.
We applied these findings to understand the biochemical pathophysiology of FXTAS
disease, employing as well some novel experimental data. In FXTAS, a mutation on
the FMR1 gene produces a 5´microsatellite repetition that enhances its scaffolding
ability. This mutated mRNA is able to sequester some important proteins into nuclear
RNP granules, such as TRA2A (i.e. a splicing factor), impeding their normal function
and therefore producing some symptoms associated with the progress of the disease.
The better understanding of the principles governing granules formation and structure
will enable to develop novel therapies (e.g. aptamers) to mitigate the development of
several neurodegenerative diseases.Los gránulos ribonucleoproteicos (gránulos RNP, por sus siglas en inglés) son complejos producidos mediante separación líquido-líquido y están constituidos principalmente por proteínas y ARN. Son responsables de numerosos procesos involucrados con la regulación del ARN. Alteraciones en la dinámica de estos complejos de proteínas y ARN están asociadas con la aparición de diversas enfermedades neurodegenerativas como el ELA o FXTAS. Sin embargo, todavía se desconocen muchos aspectos relativos a su organización interna así como las contribuciones específicas del RNA en la formación y funcionamiento de estos complejos. A fin de estudiar la estructura y formación de los gránulos RNP, hemos integrado varias bases de datos de alto rendimiento de reciente aparición. Esto incluye datos sobre la composición proteica y en ARN de los RNP, sobre la interacción de proteínas y ARN extraída de experimentos de eCLIP y sobre la estructura secundaria del transcriptoma (producida mediante PARS). Todos estos datos han sido procesados para comprender las propiedades fundamentales de los ARNs que integran los gránulos, mediante el empleo de métodos computacionales como el análisis de redes o algoritmos de agrupamiento. De esta manera, hemos producido un modelo que integra varias de estas propiedades e identifica candidatos denominados ARNs de andamiaje. Definimos ARNs de andamiaje como moléculas de ARN con una alta propensión a formar gránulos y reclutar un gran número de componentes proteicos a los gránulos RNP. También hemos encontrado que las interacciones proteína-ARN conectan los principales componentes proteicos de consenso de los gránulos de estrés (un tipo específico de gránulos RNP). También hemos estudiado la contribución de las interacciones ARN-ARN y las modificaciones post-transcriptionales del RNA en la organización interna del gránulo. Hemos aplicado estos resultados para la comprensión de la fisiopatología molecular de FXTAS, empleando también algunos datos experimentales originales. En FXTAS, una mutación en el gen FMR1 produce una repetición de microsatélite en 5´ que incrementa su capacidad como ARN de andamiaje. Este mARN mutado es capaz de secuestrar algunas proteínas importantes como TRA2A (un factor de ayuste alternativo) en gránulos RNP nucleares, impidiendo su normal funcionamiento y por consiguiente produciendo algunos síntomas asociados con el progreso de la enfermedad. Una mejor comprensión de los principios que gobiernan la formación y estructura de los gránulos puede permitir desarrollar nuevas terapias (ej: aptámeros) para mitigar el desarrollo de diversas enfermedades neurodegenerativas.
30
Arsene, Simon. "Pre-evolutionary dynamics in autocatalytic RNA networks." Thesis, Sorbonne Paris Cité, 2018. http://www.theses.fr/2018USPCC156/document.
Full textAbstract:
Les réseaux de molécules interdépendantes sont depuis quelque temps considérés comme de potentiels candidats pour avoir amorcé la transition de la biologie à la chimie. Bien qu'ils aient été intensivement examinés en théorie, il n'existe toujours aucune preuve expérimentale pour confirmer ou infirmer leur supposé rôle crucial dans les origines de la vie. En particulier, il nous manque encore une démonstration empirique des trois ingrédients habituellement présentés comme requis pour l'évolution darwinienne: l'hérédité, la variation et la sélection. Un système qui posséderait les trois tout en étant couplé à un processus de réplication en compartiments serait théoriquement capable d’évoluer au sens darwinien du terme. Par exemple, cela a été montré théoriquement pour les Ensembles Collectivement Autocatalytiques (CAS pour Collectively Aucatalytic Sets en anglais) où chaque molécule de l'ensemble est formée catalytiquement par un autre membre de l'ensemble. Ici, nous utilisons le système de ribozyme Azoarcus, qui catalysent des réactions de recombinaisons, pour former expérimentalement des CASs structurellement divers afin d’explorer leurs propriétés évolutives. Dans ce système, les ribozymes peuvent catalyser la formation d'autres ribozymes à partir de fragments plus petits, présents dans l'environnement. Nous utilisons un dispositif de microfluidique en gouttes associé au séquençage haut-débit pour mener une étude à grande échelle sur des milliers de CASs Azoarcus. Nous développons une approche perturbative pour identifier les paramètres topologiques importants contrôlant les variations observées dans les CAS à la suite de perturbations de l’environnement, ici l'ajout d'une nouvelle espèce. Nous déterminons ensuite l’ensemble restreint de caractéristiques du réseau régissant la mémoire des conditions initiales dans les CASs Azoarcus, un prérequis pour l'hérédité, en utilisant un modèle théorique validé par des données expérimentales. Enfin, nous démontrons qu’il existe dans les CASs Azoarcus des processus cataboliques qui les rendent robustes aux perturbations des fragments qui composent leur substrat et donc plus pertinent d’un point de vue prébiotique. Ces résultats démontrent le rôle crucial des CASs à base d’ARN dans les origines de la vie et illustrent comment la structure de leur réseau peut être adaptée pour obtenir des CASs avec des propriétés intéressantes d’un point de vue évolutif, ouvrant la voie à une démonstration expérimentale de l'évolution darwinienne avec système purement moléculaireNetworks of interdependent molecules are considered plausible candidates for initiating the transition from biology to chemistry. Though they have been intensively scrutinized theoretically, there is still no experimental evidence for confirming or denying their supposed crucial role in the origins of life. In particular, we are still lacking experimental proofs of any of the three ingredients usually presented as required for Darwinian evolution: heredity, variation and selection. A system that would possess the three while being coupled to some sort of encapsulated replication process would theoretically be able to undergo Darwinian evolution. As a matter of fact, this has been shown theoretically for Collectively Autocatalytic Sets (CAS) where each molecule of the set is catalytically formed by another member of the ensemble. Here we use the Azoarcus recombination ribozyme system to experimentally form structurally diverse CASs to explore their evolutionary properties. In this system, the ribozymes can catalyze the assembly of other ribozymes from smaller fragments, present in the food set. We first use a droplet microfluidics set-up coupled with next-generation sequencing to conduct a large scale study on thousands of Azoarcus CASs. We develop a perturbative approach to identify the important topological parameters that control variations in CASs as a result of environmental perturbations, here the addition of a new species. We then determine the small set of network features governing memory of the initial conditions in Azoarcus CAS, a pre-requisite for heredity, by using a computational model validated by experimental data. Finally, we demonstrate that Azoarcus CAS possess catabolic processes which make them robust to perturbations in the food set and thus more prebiotic relevant. These results provide evidence for the crucial role of RNA CASs in the origins of life and illustrate how the network structure can be tailored to obtain CASs with properties interesting from an evolutionary point of view, paving the way to an experimental demonstration of Darwinian evolution with a purely molecular system
31
Cavallie, Mester Jon William. "Using LSTM Neural Networks To Predict Daily Stock Returns." Thesis, Linnéuniversitetet, Institutionen för datavetenskap och medieteknik (DM), 2021. http://urn.kb.se/resolve?urn=urn:nbn:se:lnu:diva-106124.
Full textAbstract:
Long short-term memory (LSTM) neural networks have been proven to be effective for time series prediction, even in some instances where the data is non-stationary. This lead us to examine their predictive ability of stock market returns, as the development of stock prices and returns tend to be a non-stationary time series. We used daily stock trading data to let an LSTM train models at predicting daily returns for 60 stocks from the OMX30 and Nasdaq-100 indices. Subsequently, we measured their accuracy, precision, and recall. The mean accuracy was 49.75 percent, meaning that the observed accuracy was close to the accuracy one would observe by randomly selecting a prediction for each day and lower than the accuracy achieved by blindly predicting all days to be positive. Finally, we concluded that further improvements need to be made for models trained by LSTMs to have any notable predictive ability in the area of stock returns.
32
Rendel, Mark. "Neural network structure in an exhaustive RNA genotype-phenotype map." Thesis, University of Oxford, 2008. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.494400.
Full text
33
Yeates, Jessica Anne Mellor. "The Foundations of Network Dynamics in an RNA Recombinase System." PDXScholar, 2016. http://pdxscholar.library.pdx.edu/open_access_etds/2919.
Full textAbstract:
How life originated from physical and chemical processes is one of the great questions still unanswered today. Studies towards this effort have transitioned from the notion of a single self-replicating entity to the idea that a network of interacting molecules made this initial biological leap. In order to understand the chemical kinetic and thermodynamic mechanisms that could engender pre-life type networks we present an empirical characterization of a network of RNA recombinase molecules. We begin with 1-, 2-, and 3-molecular ensembles and provide a game theoretic analysis to describe the frequency dependent dynamics of competing and cooperating RNA genotypes. This is then extended to 4- and 5-membered networks where varying topologies are compared and mechanisms that could lead to preferential growth and selection of genotypes are described. At the core of these network connections is ribozyme catalysis initiated through a 3-nucleotide base-pairing interface. With the development of a fluorescence anisotropy method, we are able to illustrate a correlation between these binding thermodynamics and network outcomes. Finally, we consider how the heterogeneity of the environment could impact network dynamics and develop a spectrum of spatial inducing methods in which our chemical populations can be probed. These experiments illustrate simple chemical dynamics of RNA interactions, yet these very processes are the foundation for building complexity and ultimately from where selection and evolvability derive.
34
Di, Cicco Nicola. "Scalable Algorithms for Cloud Radio Access Network (C-RAN) Optimization." Master's thesis, Alma Mater Studiorum - Università di Bologna, 2021. http://amslaurea.unibo.it/23755/.
Full textAbstract:
In the evolving scenario of 5G networks, resource allocation algorithms for the Cloud Radio Access Network (C-RAN) model have proven to be the key for managing ever increasing Capital Expenditure (CAPEX) and Operating Expenditure (OPEX) for mobile networks while ensuring high Quality of Service (QoS).
In Chapter 1 a brief overview of the main elements of the C-RAN and of the methodologies that are employed in this work is provided.
In Chapter 2, an exact scalable methodology for a static traffic scenario, based on lexicographic optimization, is proposed for the solution of a multi-objective optimization problem to achieve, among other goals, the minimization of the number of active nodes in the C-RAN while supporting reliability and meeting latency constraints. The optimal solution of the most relevant objectives for networks of several tens of nodes is obtained in few tens of seconds of computational time in the worst case. For the least relevant objective a heuristic is developed, providing near optimal solutions in few seconds of computing time.
In Chapter 3, an optimization framework for dynamic C-RAN reconfiguration is developed. The objective is to maintain C-RAN cost optimization, while minimizing the cost of virtual network function migration. Significant savings in terms of migrations (above 82% for primary virtual BBU functions and above 75% for backup virtual BBU functions) can be obtained with respect to a static traffic scenario, with execution time of the optimization algorithm below 20 seconds in the worst cases, making its application feasible for dynamic scenarios.
In Chapter 4, an alternative Column Generation model formulation is developed, and the quality of the computed lower bounds is evaluated. Further extensions from this baseline (e.g. Column Generation based heuristics, exact Branch&Price algorithms) are left as future work.
In Chapter 5, the main results achieved in this work are summarized, and several possible extensions are proposed.
35
Rabia, Tarek. "Virtualisation des fonctions d'un Cloud Radio Access Network(C-RAN)." Thesis, Sorbonne université, 2018. http://www.theses.fr/2018SORUS009/document.
Full textAbstract:
La nouvelle génération de réseaux mobiles (5G) devrait faire face, durant les cinq prochaines années, à une importante croissance du volume de données, échangé entre plusieurs milliards d'objets et d'applications connectés. En outre, l'émergence de nouvelles technologies, telles que Internet of Things (IoT), conduite autonome et réalité augmentée, impose de plus fortes contraintes de performance et de qualité de service (QoS). Répondre aux besoins cités, tout en réduisant les dépenses d'investissement et d'exploitation (CAPEX/OPEX), sont les objectifs poursuivis par les opérateurs télécom, qui ont défini une nouvelle architecture d'accès radio, appelée Cloud Radio Access Network (C-RAN). Le principe du C-RAN est de centraliser, au sein d'un pool, les parties de traitement, BaseBand Unit (BBU), d'un RAN traditionnel. Les BBU sont alors dissociées de la station de base et de la partie radio, Remote Radio Unit (RRU). Ces deux parties restent néanmoins connectées à travers un réseau intermédiaire appelé Fronthaul (FH). Dans cette thèse, nous allons concevoir une nouvelle architecture C-RAN partiellement centralisée qui intègrera une plateforme de virtualisation basée sur un environnement Xen, nommée " Metamorphic Network " (MNet). A travers cette architecture, nous viserons à : i) mettre en place un pool, dans lequel des ressources physiques (processeurs, mémoire, ports réseaux, etc.) seront partagées entre des BBU virtualisées et d'autres applications, ii) établir un réseau FH ouvert aux fournisseurs de services et aux tierces parties, facilitant ainsi le déploiement des services au plus près des utilisateurs, pour une meilleure qualité d'expérience, iii) exploiter, à travers le FH, les infrastructures Ethernet existantes pour réduire les CAPEX/OPEX et enfin, iv) atteindre les performances réseau préconisées pour la 5G. Dans la première contribution, nous allons définir une nouvelle architecture Xen pour la plateforme MNet, intégrant le framework de packet processing, OpenDataPlane (ODP), au sein d’un domaine Xen privilégié, nommé « Driver Domain ». Notre objectif, à travers cette architecture, est d’accélérer le traitement des paquets de données transitant par MNet, en évitant la surutilisation, par ODP, des cœurs du processeur physique (CPU) de la plateforme. Pour cela, des cœurs CPU virtuels (vCPU) seront alloués dans le Driver Domain pour être exploités durant le traitement des paquets par ODP. Cette nouvelle plateforme MNet servira de base pour notre architecture C-RAN. Dans la seconde contribution, nous allons implémenter, au sein du FH, deux solutions réseau. La première solution, consistera à déployer le réseau de couche 2, Transparent Interconnection of Lots of Links (TRILL), pour connecter les différents éléments de notre architecture C-RAN. La seconde solution, consistera à déployer un réseau Software Defined Network (SDN), géré par le contrôleur distribué ONOS, qui sera virtualisé dans le pool BBU. Une comparaison des performances réseau sera réalisée entre ces deux solutionsOver the next five years, the new generation of mobile networks (5G) would face a significant growth of the data volume, exchanged between billions of connected objects and applications. Furthermore, the emergence of new technologies, such as Internet of Things (IoT), autonomous driving and augmented reality, imposes higher performance and quality of service (QoS) requirements. Meeting these requirements, while reducing the Capital and Operation Expenditures (CAPEX/OPEX), are the pursued goals of the mobile operators. Consequently, Telcos define a new radio access architecture, called Cloud Radio Access Network (C-RAN). The C-RAN principle is to centralize, within a pool, the processing unit of a radio interface, named BaseBand Unit (BBU). These two units are interconnected through a Fronthaul (FH) network. In this thesis, we design a new partially centralized C-RAN architecture that integrates a virtualization platform, based on a Xen environment, called Metamorphic Network (MNet). Through this architecture, we aim to: i) implement a pool in which physical resources (processors, memory, network ports, etc.) are shared between virtualized BBUs and other applications; ii) establish an open FH network that can be used by multiple operators, service providers and third parties to deploy their services and Apps closer to the users for a better Quality of Experience (QoE); iii) exploit, through the FH, the existing Ethernet infrastructures to reduce CAPEX/OPEX; and finally iv) provide the recommended network performance for the 5G. In the first contribution, we define a new Xen architecture for the MNet platform integrating the packet-processing framework, OpenDataPlane (ODP), within a privileged Xen domain, called Driver Domain (DD). This new architecture accelerates the data packet processing within MNet, while avoiding the physical CPUs overuse by ODP. Thus, virtual CPU cores (vCPU) are allocated within DD and are used by ODP to accelerate the packet processing. This new Xen architecture improves the MNet platform by 15%. In the second contribution, we implement two network solutions within the FH. The first solution consist of deploying a layer 2 network protocol, Transparent Interconnection of Lots of Links (TRILL), to connect multiple elements of our C-RAN architecture. The second solution consists of implementing a Software Defined Network (SDN) model managed by Open Network Operating System (ONOS), a distributed SDN controller that is which is virtualized within BBU pool. Moreover, a network performance comparison is performed between these two solutions
36
Nedunchelliyan, Chitra. "Peer-to-Peer Directory Service in Resource Area Network." Ohio University / OhioLINK, 2007. http://rave.ohiolink.edu/etdc/view?acc_num=ohiou1194935888.
Full text
37
Martins, Helder. "Predicting user churn on streaming services using recurrent neural networks." Thesis, KTH, Skolan för datavetenskap och kommunikation (CSC), 2017. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-217109.
Full textAbstract:
Providers of online services have witnessed a rapid growth of their user base in the last few years. The phenomenon has attracted an increasing number of competitors determined on obtaining their own share of the market. In this context, the cost of attracting new customers has increased significantly, raising the importance of retaining existing clients. Therefore, it has become progressively more important for the companies to improve user experience and ensure they keep a larger share of their users active in consuming their product. Companies are thus compelled to build tools that can identify what prompts customers to stay and also identify the users intent on abandoning the service. The focus of this thesis is to address the problem of predicting user abandonment, also known as "churn", and also detecting motives for user retention on data provided by an online streaming service. Classical models like logistic regression and random forests have been used to predict the churn probability of a customer with a fair amount of precision in the past, commonly by aggregating all known information about a user over a time period into a unique data point. On the other hand, recurrent neural networks, especially the long short-term memory (LSTM) variant, have shown impressive results for other domains like speech recognition and video classification, where the data is treated as a sequence instead. This thesis investigates how LSTM models perform for the task of predicting churn compared to standard nonsequential baseline methods when applied to user behavior data of a music streaming service. It was also explored how different aspects of the data, like the distribution between the churning and retaining classes, the size of user event history and feature representation influences the performance of predictive models. The obtained results show that LSTMs has a comparable performance to random forest for churn detection, while being significantly better than logistic regression. Additionally, a framework for creating a dataset suitable for training predictive models is provided, which can be further explored as to analyze user behavior and to create retention actions that minimize customer abandonment.Leverantörer av onlinetjänster har bevittnat en snabb användartillväxt under de senaste åren. Denna trend har lockat ett ökande antal konkurrenter som vill ta del av denna växande marknad. Detta har resulterat i att kostnaden för att locka nya kunder ökat avsevärt, vilket även ökat vikten av att behålla befintliga kunder. Det har därför gradvis blivit viktigare för företag att förbättra användarupplevelsen och se till att de behåller en större andel avanvändarna aktiva. Företag har därför ett starkt intresse avatt bygga verktyg som kan identifiera vad som driver kunder att stanna eller vad som får dem lämna. Detta arbete fokuserar därför på hur man kan prediktera att en användare är på väg att överge en tjänst, så kallad “churn”, samt identifiera vad som driver detta baserat på data från en onlinetjänst. Klassiska modeller som logistisk regression och random forests har tidigare använts på aggregerad användarinformation över en given tidsperiod för att med relativt god precision prediktera sannolikheten för att en användare kommer överge produkten. Under de senaste åren har dock sekventiella neurala nätverk (särskilt LSTM-varianten Long Short Term Memory), där data istället behandlas som sekvenser, visat imponerande resultat för andra domäner såsom taligenkänning och videoklassificering. Detta arbete undersöker hur väl LSTM-modeller kan användas för att prediktera churn jämfört med traditionella icke-sekventiella metoder när de tillämpas på data över användarbeteende från en musikstreamingtjänst. Arbetet undersöker även hur olika aspekter av data påverkar prestandan av modellerna inklusive distributionen mellan gruppen av användare som överger produkten mot de som stannar, längden av användarhändelseshistorik och olika val av användarfunktioner för modeller och användardatan. De erhållna resultaten visar att LSTM har en jämförbar prestanda med random forest för prediktering av användarchurn samt är signifikant bättre än logistisk regression. LSTMs visar sig således vara ett lämpligt val för att förutsäga churn på användarnivå. Utöver dessa resultat utvecklades även ett ramverk för att skapa dataset som är lämpliga för träning av prediktiva modeller, vilket kan utforskas ytterligare för att analysera användarbeteende och för att skapa förbättrade åtgärder för att behålla användare och minimera antalet kunder som överger tjänsten.
38
Kost, Alex. "Applying Neural Networks for Tire Pressure Monitoring Systems." DigitalCommons@CalPoly, 2018. https://digitalcommons.calpoly.edu/theses/1827.
Full textAbstract:
A proof-of-concept indirect tire-pressure monitoring system is developed using neural net- works to identify the tire pressure of a vehicle tire. A quarter-car model was developed with Matlab and Simulink to generate simulated accelerometer output data. Simulation data are used to train and evaluate a recurrent neural network with long short-term memory blocks (RNN-LSTM) and a convolutional neural network (CNN) developed in Python with Tensorflow. Bayesian Optimization via SigOpt was used to optimize training and model parameters. The predictive accuracy and training speed of the two models with various parameters are compared. Finally, future work and improvements are discussed.
39
Berlati, Alessandro. "Ambiguity in Recurrent Models: Predicting Multiple Hypotheses with Recurrent Neural Networks." Master's thesis, Alma Mater Studiorum - Università di Bologna, 2018. http://amslaurea.unibo.it/16611/.
Full textAbstract:
Multiple Hypothesis Prediction (MHP) models have been introduced to deal with uncertainty in feedforward neural networks, in particular it has been shown how to easily convert a standard single-prediction neural network into one able to show many feasible outcomes.
Ambiguity, however, is present also in problems where feedback model are needed, such as sequence generation and time series classification. In our work, we propose an extension of MHP to Recurrent Neural Networks (RNNs), especially those consisting of Long Short-Term Memory units. We test the resulting models on both regression and classification problems using public datasets, showing promising results. Our way to build MHP models can be used to retrofit other works, leading the way towards further research.
We can find many possible application scenarios in the autonomous driv- ing environment. For example, trajectory prediction, for humans and cars, or intention classification (e.g. lane change detection) are both tasks where ambiguous situations are frequent.
40
Motyčka, Jan. "Implementace mechanismů zajišťujících “RAN Slicing” v simulačním nástroji Network Simulator 3." Master's thesis, Vysoké učení technické v Brně. Fakulta elektrotechniky a komunikačních technologií, 2021. http://www.nusl.cz/ntk/nusl-442360.
Full textAbstract:
This thesis deals with the topic of network slicing technology in 5G networks, mainly on the RAN part. In the theoretical part, basic principles of 5G network slicing in core network part and RAN part are presented. Practical part contains a simulation scenario created in NS3 simulator with LENA 5G module. Results of this simulation are presented and discussed with the emphasis on RAN slicing.
41
Shojaee, Ali B. S. "Bacteria Growth Modeling using Long-Short-Term-Memory Networks." University of Cincinnati / OhioLINK, 2021. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1617105038908441.
Full text
42
Rendel, Mark D. "The evolutionary dynamics of neutral networks : lessons from RNA." Thesis, University of Oxford, 2008. http://ora.ox.ac.uk/objects/uuid:85107ca7-fada-4582-95e7-17b5bbb038cd.
Full textAbstract:
The evolutionary options of a population are strongly influenced by the avail- ability of adaptive mutants. In this thesis, I use the concept of neutral networks to show that neutral drift can actually increase the accessibility of adaptive mu- tants, and therefore facilitate adaptive evolutionary change. Neutral networks are groups of unique genotypes which all code for the same phenotype, and are connected by simple point mutations. I calculate the size and shape of the networks in a small but exhaustively enumerated space of RNA genotypes by mapping the sequences to RNA secondary structure phenotypes. The qual- itative results are similar to those seen in many other genotype–phenotype map models, despite some significant methodological differences. I show that the boundary of each network has single point–mutation connections to many more phenotypes than the average individual genotype within that network. This means that paths involving a series of neutral point–mutation steps across a network can allow evolution to adaptive phenotypes which would otherwise be extremely unlikely to arise spontaneously. This can be likened to walking along a flat ridge in an adaptive landscape, rather than traversing or jumping across a lower fitness valley. Within this model, when a genotype is made up of just 10 bases, the mean neutral path length is 1.88 point mutations. Furthermore, the map includes some networks that are so convoluted that the path through the network is longer than the direct route between two sequences. A minimum length adaptive walk across the genotype space usually takes as many neutral steps as adaptive ones on its way to the optimum phenotype. Finally I show that the shape of a network can have a very important affect on the number of generations it takes a population to drift across it, and that the more routes between two sequences, the fewer generations required for a population to find an advantageous sequence. My conclusion is that, within the RNA map at least, the size, shape and connectivity of neutral networks all have a profound effect on the way that sequences change and populations evolve, and by not considering them, we risk missing an important evolutionary mechanism.
43
Näslund, Per. "Artificial Neural Networks in Swedish Speech Synthesis." Thesis, KTH, Tal-kommunikation, 2018. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-239350.
Full textAbstract:
Text-to-speech (TTS) systems have entered our daily lives in the form of smart assistants and many other applications. Contemporary re- search applies machine learning and artificial neural networks (ANNs) to synthesize speech. It has been shown that these systems outperform the older concatenative and parametric methods. In this paper, ANN-based methods for speech synthesis are ex- plored and one of the methods is implemented for the Swedish lan- guage. The implemented method is dubbed “Tacotron” and is a first step towards end-to-end ANN-based TTS which puts many differ- ent ANN-techniques to work. The resulting system is compared to a parametric TTS through a strength-of-preference test that is carried out with 20 Swedish speaking subjects. A statistically significant pref- erence for the ANN-based TTS is found. Test subjects indicate that the ANN-based TTS performs better than the parametric TTS when it comes to audio quality and naturalness but sometimes lacks in intelli- gibility.Talsynteser, också kallat TTS (text-to-speech) används i stor utsträckning inom smarta assistenter och många andra applikationer. Samtida forskning applicerar maskininlärning och artificiella neurala nätverk (ANN) för att utföra talsyntes. Det har visats i studier att dessa system presterar bättre än de äldre konkatenativa och parametriska metoderna. I den här rapporten utforskas ANN-baserade TTS-metoder och en av metoderna implementeras för det svenska språket. Den använda metoden kallas “Tacotron” och är ett första steg mot end-to-end TTS baserat på neurala nätverk. Metoden binder samman flertalet olika ANN-tekniker. Det resulterande systemet jämförs med en parametriskt TTS genom ett graderat preferens-test som innefattar 20 svensktalande försökspersoner. En statistiskt säkerställd preferens för det ANN- baserade TTS-systemet fastställs. Försökspersonerna indikerar att det ANN-baserade TTS-systemet presterar bättre än det parametriska när det kommer till ljudkvalitet och naturlighet men visar brister inom tydlighet.
44
Woodcock, M. Ryan. "Network Analysis and Comparative Phylogenomics of MicroRNAs and their Respective Messenger RNA Targets Using Twelve Drosophila species." VCU Scholars Compass, 2010. http://scholarscompass.vcu.edu/etd/155.
Full textAbstract:
MicroRNAs represent a special class of small (~21–25 nucleotides) non-coding RNA molecules which exert powerful post-transcriptional control over gene expression in eukaryotes. Indeed microRNAs likely represent the most abundant class of regulators in animal gene regulatory networks. This study describes the recovery and network analyses of a suite of homologous microRNA targets recovered through two different prediction methods for whole gene regions across twelve Drosophila species. Phylogenetic criteria under an accepted tree topology were used as a reference frame to 1) make inference into microRNA-target predictions, 2) study mathematical properties of microRNA-gene regulatory networks, 3) and conduct novel phylogenetic analyses using character data derived from weighted edges of the microRNA-target networks. This study investigates the evidences of natural selection and phylogenetic signatures inherent within the microRNA regulatory networks and quantifies time and mutation necessary to rewire a microRNA regulatory network. Selective factors that appear to operate upon seed aptamers include cooperativity (redundancy) of interactions and transcript length. Topological analyses of microRNA regulatory networks recovered significant enrichment for a motif possessing a redundant link in all twelve species sampled. This would suggest that optimization of the whole interactome topology itself has been historically subject to natural selection where resilience to attack have offered selective advantage. It seems that only a modest number of microRNA–mRNA interactions exhibit conservation over Drosophila cladogenesis. The decrease in conserved microRNA-target interactions with increasing phylogenetic distance exhibited a cure typical of a saturation phenomena. Scale free properties of a network intersection of microRNA target predictions methods were found to transect taxonomic hierarchy.
45
Santos, Bruno Acácio de Castro Moreira dos. "Small RNAs in gene regulatory networks." Thesis, University of Cambridge, 2015. https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.708543.
Full text
46
Schmidt, Robert. "Slicing in heterogeneous software-defined radio access networks." Electronic Thesis or Diss., Sorbonne université, 2021. http://www.theses.fr/2021SORUS525.
Full textAbstract:
Les réseaux 5G sont envisagés comme un changement de paradigme vers des réseaux orientés services. Dans cette thèse, nous étudions comment combiner efficacement le découpage en tranches et le SD-RAN afin de fournir le niveau requis de flexibilité et de programmabilité dans l'infrastructure RAN pour réaliser des réseaux multi-locataires orientés services. Premièrement, nous concevons une abstraction d'une station de base pour représenter les stations de base logiques et décrire un service de réseau virtualisé. Deuxièmement, nous proposons une nouvelle plateforme SD-RAN conforme aux normes, appelée FlexRIC, sous la forme d'un kit de développement logiciel (SDK). Troisièmement, nous fournissons une conception modulaire pour un cadre d'ordonnancement MAC tenant compte des tranches afin de gérer et de contrôler efficacement les ressources radio dans un environnement multiservice avec un support de qualité de service (QoS). Enfin, nous présentons une couche de virtualisation SD-RAN dynamique basée sur le SDK FlexRIC et le cadre d'ordonnancement MAC pour composer de manière flexible une infrastructure SD-RAN multiservice et fournir une programmabilité pour de multiples contrôleurs SD-RAN5G networks are envisioned to be a paradigm shift towards service-oriented networks. In this thesis, we investigate how to efficiently combine slicing and SD-RAN to provide the required level of flexibility and programmability in the RAN infrastructure to realize service-oriented multi-tenant networks. First, we devise an abstraction of a base station to represent logical base stations and describe a virtualized network service. Second, we propose a novel standard-compliant SD-RAN platform, named FlexRIC, in the form of a software development kit (SDK). Third, we provide a modular design for a slice-aware MAC scheduling framework to efficiently manage and control the radio resources in a multi-service environment with quality-of-service (QoS) support. Finally, we present a dynamic SD-RAN virtualization layer based on the FlexRIC SDK and MAC scheduling framework to flexibly compose a multi-service SD-RAN infrastructure and provide programmability for multiple SD-RAN controllers
47
Wang, Huajun. "Interplay between capacity and energy consumption in C-RAN transport network design." Thesis, KTH, Skolan för informations- och kommunikationsteknik (ICT), 2016. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-204939.
Full textAbstract:
Current mobile network architecture is facing a big challenge as the traffic demands have been increasing dramatically these years. Explosive mobile data demands are driving a significant growth in energy consumption in mobile networks, as well as the cost and carbon footprints [1]. In 2010, China Mobile Research Institute proposed Cloud Radio Access Network (C-RAN) [2], which has been regarded as one of the most promising architecture to solve the challenge of operators. In C-RAN, the baseband units (BBU) are decoupled from the remote radio units (RRH) and centralized in one or more locations. The feasibility of combination of implementing the very tight radio coordination schemes and sharing baseband processing and cooling system resources proves to be the two main advantages of C-RAN compared to traditional RAN. More importantly, mobile operators can quickly deploy RRHs to expand and make upgrades to their networks. Therefore, the C-RAN has been advocated by both operators and equipment vendors as a means to achieve the significant performance gains required for 5G [3]. However, one of the biggest barriers has shown up in the deployment of C-RAN as the novel architecture imposes very high capacity requirement on the transport network between the RRHs and BBUs, which is been called fronthaul network. With the implementation of 5G wireless system using advanced multi-antenna transmission (MIMO), the capacity requirement would go further up, as well as the power consumption. One solution has been proposed to solve the problem is to have the baseband functions divided, partially staying with RRHs and other functions would be centralized in BBU pool. Different splitting solutions has been proposed in [4] [5] and [6]. In this thesis work, we choose four different splitting solutions to build four CRAN architecture models. Under one specific case scenario with the fixed number of LTE base stations, we calculate the transport capacity requirement for fronthaul and adopt three different fronthaul technology. The power consumption is calculated by adding up the power utilized by RRHs, fronthaul network and baseband processing. By comparing the numerical results, split 1 and 2 shows the best results while split 2 is more practical for dense cell area, since split 1 requires large fronthaul capacity. The fronthaul transport technology can be decided according to different density of base stations. TWDM-PON shows better energy performance as fronthaul network when the capacity requirement is high, compared to EPON. However, for larger number of BSs, mm-Wave fronthaul is a better solution in terms of energy efficiency, fiber saving and flexibility.
48
Singh, Jaswinder. "RNA Structure Prediction using Deep Neural Network Architectures and Improved Evolutionary Profiles." Thesis, Griffith University, 2022. http://hdl.handle.net/10072/414924.
Full textAbstract:
RNAs are important biological macro-molecules that play critical roles in many biological processes. The functionality of RNA depends on its three-dimensional (3D) structure, which further depends on its primary structure, i.e. the order of sequence of nucleotides in the RNA chain. Direct prediction of the 3D structure of an RNA from its sequence is a challenging task. Therefore, the 3D structure is further divided into two-dimensional (2D) properties such as secondary structure, contact maps and one-dimensional (1D) properties such as torsion angles and solvent accessibility. An accurate prediction of these 1D and 2D structural properties will increase the accuracy in predicting the 3D structure of the RNA. This thesis explores various deep learning algorithms and input features relevant to predicting the 1D and 2D structural properties of an RNA. Using these predicted 1D and 2D structural properties further as restraints, we have demonstrated an improvement in the prediction of the RNA 3D structure. There are four primary studies performed in this thesis for RNA structural properties prediction. The first study introduces two methods (SPOT-RNA and SPOT-RNA2) for RNA secondary structure prediction using an ensemble of Residual Con-volution and Bi-directional LSTM recurrent neural networks. This study shows that deep learning based methods can outperform existing dynamic programming based algorithms and achieve state-of-the-art performance using single-sequence and evolutionary information as input. The second study investigates the application of deep neural networks for predicting RNA backbone torsion and pseudotorsion angles. We have pioneered in predicting the backbone torsion and pseudotorsion angles using deep learning (SPOT-RNA-1D). The angles predicted using SPOT-RNA-1D could be used as 3D model quality indicators. The third study introduces a method (SPOT-RNA-2D) to predict RNA distance-based contact maps using an ensemble of deep neural networks and improved evolutionary profles from RNAcmap. This study shows that the use of predicted distance-based contact maps as restraints can signifcantly improve the performance of 3D structure prediction. The fourth study developed a fully automated pipeline (RNAcmap2) to generate aligned homologs. Here, we showed that using a combination of BLAST-N and iterative INFERNAL searches along with an expanded sequence database leads to multiple sequence alignments (MSA) comparable to those provided by Rfam MSAs according to secondary structure extracted from mutational coupling analysis and alignment accuracy when compared to structural alignment. This fully automatic tool (RNAcmap2) allows to search homolog, multiple sequence alignment, and mutational coupling analysis for any non-Rfam RNA sequences with Rfam-like performance. The improved RNA 1D and 2D structural properties predictions using deep learning along with improved homolog search collectively is expected to be useful in predicting RNA three-dimensional structure and better un-derstand its biological function.Thesis (PhD Doctorate)
Doctor of Philosophy (PhD)
School of Eng & Built Env
Science, Environment, Engineering and Technology
Full Text
49
Шурук, Андрій Сергійович. "Система аналізу людської активності на основі даних з носимих пристроїв." Master's thesis, КПІ ім. Ігоря Сікорського, 2020. https://ela.kpi.ua/handle/123456789/38278.
Full textAbstract:
Актуальність проблеми. Глобалізація та збільшення числа населення сприяють розвитку галузей пов’язаних зі спостереженням за людською активністю, а отже — появі нових засобів для спостереження за різноманітними показниками діяльності людини та способів аналізу цих показників. Враховуючи дані фактори, важливо в сучасному світі правильно використовувати такі об’єми даних в умовах ринку. Якщо мова йде про, наприклад, догляд за літніми людьми, то важливо правильно проаналізувати та спробувати розпізнати ту, чи іншу активність людини, це може допомогти збільшити тривалість життя літніх людей.
Зв’язок роботи з науковими програмами, планами, темами. Дипломну роботу (магістерського) рівня вищої освіти було виконано в Національному технічному університеті України «Київський політехнічний інститут імені Ігоря Сікорського» відповідно до планів науково-дослідних робіт кафедри обчислювальної техніки.
Мета і задачі дослідження. Завданням цієї роботи є дослідження можливості розпізнавання людської активності на основі даних носимих пристроїв. Метою є розроблення системи, побудованої на базі нейронної мережі, здатної розпізнати людську активність та надати їх користувачеві за допомогою кросплатформеного додатку.
Об’єкт дослідження. Процес розпізнавання людської активності з використанням елементів нейронної мережі.
Предмет дослідження. Методи аналізу та обробки даних отриманих з носимих пристроїв в реальному часі.
Новизна. Запропоновано новий спосіб розпізнавання людської активності на основі даних носимих пристроїв, який за рахунок використання нейронної мережі, дозволяє отримати результати розпізнавання в реальному часі з високою точністю.The urgency of the problem. Globalization and population growth are contributing to the development of areas related to human activity monitoring, and thus to the emergence of new tools for monitoring various human performance indicators and ways to analyze these indicators. Given these factors, it is important in today's world to properly use such volumes of data in market conditions. When it comes to, for example, caring for the elderly, it is important to properly analyze and try to recognize a particular human activity, it can help increase the life expectancy of the elderly. Relationship with working with scientific programs, plans, topics. Thesis of master's level of higher education was performed at the National Technical University of Ukraine "Kyiv Polytechnic Institute named after Igor Sikorsky" in accordance with the plans of research work of the Department of Computer Science. The purpose and objectives of the study. The aim of this work is to study the possibility of recognizing human activity based on the data of wearable devices. The aim is to develop a system built on a neural network capable of recognizing human activity and providing it to the user through a cross-platform application. Object of study. The process of recognizing human activity using elements of the neural network. Subject of study. Methods of analysis and processing of data obtained from wearable devices in real time. Novelty. A new method of recognizing human activity based on the data of wearable devices is proposed, which, due to the use of a neural network, allows to obtain real-time recognition results with high accuracy.
50
Rosell, Felicia. "Tracking a ball during bounce and roll using recurrent neural networks." Thesis, KTH, Skolan för elektroteknik och datavetenskap (EECS), 2018. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-239733.
Full textAbstract:
In many types of sports, on-screen graphics such as an reconstructed ball trajectory, can be displayed for spectators or players in order to increase understanding. One sub-problem of trajectory reconstruction is tracking of ball positions, which is a difficult problem due to the fast and often complex ball movement. Historically, physics based techniques have been used to track ball positions, but this thesis investigates using a recurrent neural network design, in the application of tracking bouncing golf balls. The network is trained and tested on synthetically created golf ball shots, created to imitate balls shot out from a golf driving range. It is found that the trained network succeeds in tracking golf balls during bounce and roll, with an error rate of under 11 %.Grafik visad på en skärm, så som en rekonstruerad bollbana, kan användas i många typer av sporter för att öka en åskådares eller spelares förståelse. För att lyckas rekonstruera bollbanor behöver man först lösa delproblemet att följa en bolls positioner. Följning av bollpositioner är ett svårt problem på grund av den snabba och ofta komplexa bollrörelsen. Tidigare har fysikbaserade tekniker använts för att följa bollpositioner, men i den här uppsatsen undersöks en metod baserad på återkopplande neurala nätverk, för att följa en studsande golfbolls bana. Nätverket tränas och testas på syntetiskt skapade golfslag, där bollbanorna är skapade för att imitera golfslag från en driving range. Efter träning lyckades nätverket följa golfbollar under studs och rull med ett fel på under 11 %.