Search results for: neural machine translation

Authors: Luisa Carrer, Lorenzo De Masi

Abstract:

In this paper, we exploit the massive amount of information from Facebook to build a measure of gender attitudes in Italy at a previously impossible resolution—the municipal level. We construct our index via a machine learning method to replicate a benchmark region-level measure. Interestingly, we find that most of the variation in our Gender Norms Index (GNI) is across towns within narrowly defined geographical areas rather than across regions or provinces. In a second step, we show how this local variation in norms can be leveraged for identification purposes. In particular, we use our index to investigate whether these differences in norms carry over to the policy activity of politicians elected in the Italian Parliament. We document that females are more likely to sit in parliamentary committees focused on gender-sensitive matters, labor, and social issues, but not if they come from a relatively conservative town. These effects are robust to conditioning the legislative term and electoral district, suggesting the importance of social norms in shaping legislators’ policy activity.

Keywords: gender equality, gender norms index, Facebook, machine learning, politics

Procedia PDF Downloads 61

Authors: Yasmine Abu Adla, Racha Soubra, Milana Kasab, Mohamad O. Diab, Aly Chkeir

Abstract:

Over the past several years, researchers have shown a great interest in assessing the mobility of elderly people to measure their functional status. Usually, such an assessment is done by conducting tests that require the subject to walk a certain distance, turn around, and finally sit back down. Consequently, this study aims to provide an at home monitoring system to assess the patient’s status continuously. Thus, we proposed a technique to automatically detect when a subject sits down while walking at home. In this study, we utilized a Doppler radar system to capture the motion of the subjects. More than 20 features were extracted from the radar signals, out of which 11 were chosen based on their intraclass correlation coefficient (ICC > 0.75). Accordingly, the sequential floating forward selection wrapper was applied to further narrow down the final feature vector. Finally, 5 features were introduced to the linear discriminant analysis classifier, and an accuracy of 93.75% was achieved as well as a precision and recall of 95% and 90%, respectively.

Keywords: Doppler radar system, stand-to-sit phase, TUG test, machine learning, classification

Procedia PDF Downloads 145

Authors: Cheng-yu Hsieh, Bo Zhang, Ahmed Hambaba

Abstract:

Stock market prediction has long been an area of interest for both expert analysts and investors, driven by its complexity and the noisy, volatile conditions it operates under. This research examines the efficacy of combining the Patch Time Series Transformer (PatchTST) with wavelet transforms, specifically focusing on Haar and Daubechies wavelets, in forecasting the adjusted closing price of the S&P 500 index for the following day. By comparing the performance of the augmented PatchTST models with traditional predictive models such as Recurrent Neural Networks (RNNs), Convolutional Neural Networks (CNNs), Long Short-Term Memory (LSTM) networks, and Transformers, this study highlights significant enhancements in prediction accuracy. The integration of the Daubechies wavelet with PatchTST notably excels, surpassing other configurations and conventional models in terms of Mean Absolute Error (MAE) and Mean Squared Error (MSE). The success of the PatchTST model paired with Daubechies wavelet is attributed to its superior capability in extracting detailed signal information and eliminating irrelevant noise, thus proving to be an effective approach for financial time series forecasting.

Keywords: deep learning, financial forecasting, stock market prediction, patch time series transformer, wavelet transform

Procedia PDF Downloads 25

Authors: Maryam Ramezankhani

Abstract:

The present research tries to study the possibility of formalizing the sense relation of hyponymy. It applied mathematical tools and also uses mathematical logic concepts especially those from propositional logic. In order to do so, firstly, it goes over the definitions of hyponymy presented in linguistic dictionaries and semantic textbooks. Then, it introduces a formal translation of the sense relation of hyponymy. Lastly, it examines the efficiency of the suggested formula by some examples of natural language.

Keywords: sense relations, hyponymy, formalizing, words’ sense relation, formalizing sense relations

Procedia PDF Downloads 220

Authors: Jessica Lopez Espejel, Mahaman Sanoussi Yahaya Alassan, Walid Dahhane, El Hassane Ettifouri

Abstract:

Pretrained transformer-based models have shown high performance in natural language generation tasks. However, a new wave of interest has surged: automatic programming language code generation. This task consists of translating natural language instructions to a source code. Despite the fact that well-known pre-trained models on language generation have achieved good performance in learning programming languages, effort is still needed in automatic code generation. In this paper, we introduce JaCoText, a model based on Transformer neural network. It aims to generate java source code from natural language text. JaCoText leverages the advantages of both natural language and code generation models. More specifically, we study some findings from state of the art and use them to (1) initialize our model from powerful pre-trained models, (2) explore additional pretraining on our java dataset, (3) lead experiments combining the unimodal and bimodal data in training, and (4) scale the input and output length during the fine-tuning of the model. Conducted experiments on CONCODE dataset show that JaCoText achieves new state-of-the-art results.

Keywords: java code generation, natural language processing, sequence-to-sequence models, transformer neural networks

Procedia PDF Downloads 255

Authors: Aadityaa Singla

Abstract:

The big five personality traits are as follows: openness, conscientiousness, extraversion, agreeableness, and neuroticism. In order to get an insight into their personality, many flocks to these categories, which each have different meanings/characteristics. This information is important not only to individuals but also to career professionals and psychologists who can use this information for candidate assessment or job recruitment. The links between AI and psychology have been well studied in cognitive science, but it is still a rather novel development. It is possible for various AI classification models to accurately predict a personality question via ten input questions. This would contrast with the hundred questions that normal humans have to answer to gain a complete picture of their five personality traits. In order to approach this problem, various AI classification models were used on a dataset to predict what a user may answer. From there, the model's prediction was compared to its actual response. Normally, there are five answer choices (a 20% chance of correct guess), and the models exceed that value to different degrees, proving their significance. By utilizing an MLP classifier, decision tree, linear model, and K-nearest neighbors, they were able to obtain a test accuracy of 86.643, 54.625, 47.875, and 52.125, respectively. These approaches display that there is potential in the future for more nuanced predictions to be made regarding personality.

Keywords: machine learning, personally, big five personality traits, cognitive science

Procedia PDF Downloads 134

Authors: Hayriye Anıl, Görkem Kar

Abstract:

In addition to being rich in renewable energy resources, Turkey is one of the countries that promise potential in geothermal energy production with its high installed power, cheapness, and sustainability. Increasing imbalance penalties become an economic burden for organizations since geothermal generation plants cannot maintain the balance of supply and demand due to the inadequacy of the production forecasts given in the day-ahead market. A better production forecast reduces the imbalance penalties of market participants and provides a better imbalance in the day ahead market. In this study, using machine learning, deep learning, and, time series methods, the total generation of the power plants belonging to Zorlu Natural Electricity Generation, which has a high installed capacity in terms of geothermal, was estimated for the first one and two weeks of March, then the imbalance penalties were calculated with these estimates and compared with the real values. These modeling operations were carried out on two datasets, the basic dataset and the dataset created by extracting new features from this dataset with the feature engineering method. According to the results, Support Vector Regression from traditional machine learning models outperformed other models and exhibited the best performance. In addition, the estimation results in the feature engineering dataset showed lower error rates than the basic dataset. It has been concluded that the estimated imbalance penalty calculated for the selected organization is lower than the actual imbalance penalty, optimum and profitable accounts.

Keywords: machine learning, deep learning, time series models, feature engineering, geothermal energy production forecasting

Procedia PDF Downloads 91

Authors: Bipasha Sen, Aditya Agarwal

Abstract:

Multilingual automatic speech recognition (ASR) system is a single entity capable of transcribing multiple languages sharing a common phone space. Performance of such a system is highly dependent on the compatibility of the languages. State of the art speech recognition systems are built using sequential architectures based on recurrent neural networks (RNN) limiting the computational parallelization in training. This poses a significant challenge in terms of time taken to bootstrap and validate the compatibility of multiple languages for building a robust multilingual system. Complex architectural choices based on self-attention networks are made to improve the parallelization thereby reducing the training time. In this work, we propose Reed, a simple system based on 1D convolutions which uses very short context to improve the training time. To improve the performance of our system, we use raw time-domain speech signals directly as input. This enables the convolutional layers to learn feature representations rather than relying on handcrafted features such as MFCC. We report improvement on training and inference times by atleast a factor of 4x and 7.4x respectively with comparable WERs against standard RNN based baseline systems on SpeechOcean's multilingual low resource dataset.

Keywords: convolutional neural networks, language compatibility, low resource languages, multilingual automatic speech recognition

Procedia PDF Downloads 104

Authors: Ronit Chakraborty, Sugata Banerji

Abstract:

There has been a marked increase in the reported prevalence of Autism Spectrum Disorder (ASD) among children in the US over the past two decades. This research has analyzed the growth in state-level ASD prevalence against 45 different potentially explanatory factors, including socio-economic, demographic, healthcare, public policy, and political factors. The goal was to understand if these factors have adequate predictive power in modeling the differential growth in ASD prevalence across various states and if they do, which factors are the most influential. The key findings of this study include (1) the confirmation that the chosen feature set has considerable power in predicting the growth in ASD prevalence, (2) the identification of the most influential predictive factors, (3) given the nature of the most influential predictive variables, an indication that a considerable portion of the reported ASD prevalence differentials across states could be attributable to over and under diagnosis, and (4) identification of Florida as a key outlier state pointing to a potential under-diagnosis of ASD there.

Keywords: autism spectrum disorder, clustering, machine learning, predictive modeling

Procedia PDF Downloads 84

Authors: Pei Yi Lin

Abstract:

Objective: The objective of this study is to use machine learning methods to build an early prediction classifier model for acute delirium to improve the quality of medical care for intensive care patients. Background: Delirium is a common acute and sudden disturbance of consciousness in critically ill patients. After the occurrence, it is easy to prolong the length of hospital stay and increase medical costs and mortality. In 2021, the incidence of delirium in the intensive care unit of internal medicine was as high as 59.78%, which indirectly prolonged the average length of hospital stay by 8.28 days, and the mortality rate is about 2.22% in the past three years. Therefore, it is expected to build a delirium prediction classifier through big data analysis and machine learning methods to detect delirium early. Method: This study is a retrospective study, using the artificial intelligence big data database to extract the characteristic factors related to delirium in intensive care unit patients and let the machine learn. The study included patients aged over 20 years old who were admitted to the intensive care unit between May 1, 2022, and December 31, 2022, excluding GCS assessment <4 points, admission to ICU for less than 24 hours, and CAM-ICU evaluation. The CAMICU delirium assessment results every 8 hours within 30 days of hospitalization are regarded as an event, and the cumulative data from ICU admission to the prediction time point are extracted to predict the possibility of delirium occurring in the next 8 hours, and collect a total of 63,754 research case data, extract 12 feature selections to train the model, including age, sex, average ICU stay hours, visual and auditory abnormalities, RASS assessment score, APACHE-II Score score, number of invasive catheters indwelling, restraint and sedative and hypnotic drugs. Through feature data cleaning, processing and KNN interpolation method supplementation, a total of 54595 research case events were extracted to provide machine learning model analysis, using the research events from May 01 to November 30, 2022, as the model training data, 80% of which is the training set for model training, and 20% for the internal verification of the verification set, and then from December 01 to December 2022 The CU research event on the 31st is an external verification set data, and finally the model inference and performance evaluation are performed, and then the model has trained again by adjusting the model parameters. Results: In this study, XG Boost, Random Forest, Logistic Regression, and Decision Tree were used to analyze and compare four machine learning models. The average accuracy rate of internal verification was highest in Random Forest (AUC=0.86), and the average accuracy rate of external verification was in Random Forest and XG Boost was the highest, AUC was 0.86, and the average accuracy of cross-validation was the highest in Random Forest (ACC=0.77). Conclusion: Clinically, medical staff usually conduct CAM-ICU assessments at the bedside of critically ill patients in clinical practice, but there is a lack of machine learning classification methods to assist ICU patients in real-time assessment, resulting in the inability to provide more objective and continuous monitoring data to assist Clinical staff can more accurately identify and predict the occurrence of delirium in patients. It is hoped that the development and construction of predictive models through machine learning can predict delirium early and immediately, make clinical decisions at the best time, and cooperate with PADIS delirium care measures to provide individualized non-drug interventional care measures to maintain patient safety, and then Improve the quality of care.

Keywords: critically ill patients, machine learning methods, delirium prediction, classifier model

Procedia PDF Downloads 51

Authors: Humeyra Bolakar Tosun

Abstract:

The transport sector has an undisputed place in human life. People need transport access to continuous increase day by day with growing population. The number of rail network, urban transport planning, infrastructure improvements, transportation management and other related areas is a key factor affecting our country made it quite necessary to improve the work of transportation. In this context, it plays an important role in domestic rail freight demand planning. Alternatives that the increase in the transportation field and has made it mandatory requirements such as the demand for improving transport quality. In this study generally is known and used in studies by the definition, rail freight transport, railway line length, population, energy consumption. In this study, Iron Road Load Net Demand was modeled by multiple regression and ANN methods. In this study, model dependent variable (Output) is Iron Road Load Net demand and 6 entries variable was determined. These outcome values extracted from the model using ANN and regression model results. In the regression model, some parameters are considered as determinative parameters, and the coefficients of the determinants give meaningful results. As a result, ANN model has been shown to be more successful than traditional regression model.

Keywords: railway load carrying, neural network, modelling transport, transportation

Procedia PDF Downloads 134

Authors: Sholeh Motaghian, Pekka Toivanen, Keiji Haataja

Abstract:

The swift expansion of the transportation industry and its associated emissions have captured the focus of policymakers who are dedicated to upholding ecological sustainability. As a result, understanding the key contributors to transportation emissions is of utmost significance. Amidst the escalating transportation emissions, the significance of electric vehicles cannot be overstated. Electric vehicles play a critical role in steering us towards a low-carbon economy and a sustainable ecological setting. The effective integration of electric vehicles hinges on the development of energy consumption models capable of accurately and efficiently predicting energy usage. Enhancing the energy efficiency of electric vehicles will play a pivotal role in reducing driver concerns and establishing a vital framework for the efficient operation, planning, and management of charging infrastructure. In this article, the works done in this field are reviewed, and the advantages and disadvantages of each are stated.

Keywords: deep learning, electrical vehicle, energy consumption, machine learning, smart grid

Procedia PDF Downloads 59

Authors: Richard Draper

Abstract:

A supercomputer can be constructed from identical building blocks which are small parallel processors connected by a network referred to as the local network. The routers have unused ports which are used to interconnect the building blocks. These connections are referred to as the global network. The address space has a global and a local component (g, l). The conventional way to connect the building blocks is to connect (g, l) to (g’,l). If there are K blocks, this requires K global ports in each router. If a block is of size M, the result is a machine with KM routers having diameter two. To increase the size of the machine to 2K blocks, each router connects to only half of the other blocks. The result is a larger machine but also one with greater diameter. This is a crude description of how the network of the CRAY XC® is designed. In this paper, a family of interconnection networks using routers with K global and M local ports is defined. Coordinates are (c,d, p) and the global connections are (c,d,p)↔(c’,p,d) which swaps p and d. The network is denoted D3(K,M) and is called a Swapped Dragonfly. D3(K,M) has KM2 routers and has diameter three, regardless of the size of K. To produce a network of size KM2 conventionally, diameter would be an increasing function of K. The family of Swapped Dragonflies has other desirable properties: 1) D3(K,M) scales linearly in K and quadratically in M. 2) If L < K, D3(K,M) contains many copies of D3(L,M). 3) If L < M, D3(K,M) contains many copies of D3(K,L). 4) D3(K,M) can perform an all-to-all exchange in KM2+KM time which is only slightly more than the time to do a one-to-all. This paper makes several contributions. It is the first time that a swap has been used to define a linearly scalable family of networks. Structural properties of this new family of networks are thoroughly examined. A synchronizing packet header is introduced. It specifies the path to be followed and it makes it possible to define highly parallel communication algorithm on the network. Among these is an all-to-all exchange in time KM2+KM. To demonstrate the effectiveness of the swap properties of the network of the CRAY XC® and D3(K,16) are compared.

Keywords: all-to-all exchange, CRAY XC®, Dragonfly, interconnection network, packet switching, swapped network, topology

Procedia PDF Downloads 108

Authors: Ronak Khosravi, Mahmood Abbasi Layegh, Siamak Haghipour, Avin Esmaili

Abstract:

In this paper, a study on classification of musical woodwind instruments using a small set of features selected from a broad range of extracted ones by the sequential forward selection method was carried out. Firstly, we extract 42 features for each record in the music database of 402 sound files belonging to five different groups of Flutes (end blown and internal duct), Single –reed, Double –reed (exposed and capped), Triple reed and Quadruple reed. Then, the sequential forward selection method is adopted to choose the best feature set in order to achieve very high classification accuracy. Two different classification techniques of support vector machines and relevance vector machines have been tested out and an accuracy of up to 96% can be achieved by using 21 time, frequency and coefficient features and relevance vector machine with the Gaussian kernel function.

Keywords: coefficient features, relevance vector machines, spectral features, support vector machines, temporal features

Procedia PDF Downloads 302

Authors: Mehmet Akif Bütüner, İlhan Koşalay

Abstract:

Hydroelectric power plants (HEPPs) are renewable energy power plants with the highest installed power in the world. While the control systems operating in these power plants ensure that the system operates at the desired operating point, it is also responsible for stopping the relevant unit safely in case of any malfunction. While these control systems are expected not to miss signals that require stopping, on the other hand, it is desired not to cause unnecessary stops. In traditional control systems including modern systems with SCADA infrastructure, alarm conditions to create warnings or trip conditions to put relevant unit out of service automatically are usually generated with predefined limits regardless of different operating conditions. This approach results in alarm/trip conditions to be less likely to detect minimal changes which may result in serious malfunction scenarios in near future. With the methods proposed in this research, routine behavior of the oil circulation of hydraulic governor of a HEPP will be modeled with machine learning methods using historical data obtained from SCADA system. Using the created model and recently gathered data from control system, oil pressure of hydraulic accumulators will be estimated. Comparison of this estimation with the measurements made and recorded instantly by the SCADA system will help to foresee failure before becoming worse and determine remaining useful life. By using model outputs, maintenance works will be made more planned, so that undesired stops are prevented, and in case of any malfunction, the system will be stopped or several alarms are triggered before the problem grows.

Keywords: hydroelectric, governor, anomaly detection, machine learning, regression

Procedia PDF Downloads 75

Authors: Yeonjeong Lee, Kyoung-jae Kim, Youngtae Kim

Abstract:

This study proposes a novel recommender system that uses data mining and multi-model ensemble techniques to enhance the recommendation performance through reflecting the precise user’s preference. The proposed model consists of two steps. In the first step, this study uses logistic regression, decision trees, and artificial neural networks to predict customers who have high likelihood to purchase products in each product group. Then, this study combines the results of each predictor using the multi-model ensemble techniques such as bagging and bumping. In the second step, this study uses the market basket analysis to extract association rules for co-purchased products. Finally, the system selects customers who have high likelihood to purchase products in each product group and recommends proper products from same or different product groups to them through above two steps. We test the usability of the proposed system by using prototype and real-world transaction and profile data. In addition, we survey about user satisfaction for the recommended product list from the proposed system and the randomly selected product lists. The results also show that the proposed system may be useful in real-world online shopping store.

Keywords: product recommender system, ensemble technique, association rules, decision tree, artificial neural networks

Procedia PDF Downloads 279

Authors: P. Wlodarczak, J. Soar, M. Ally

Abstract:

The dramatic rise in the use of Social Media (SM) platforms such as Facebook and Twitter provide access to an unprecedented amount of user data. Users may post reviews on products and services they bought, write about their interests, share ideas or give their opinions and views on political issues. There is a growing interest in the analysis of SM data from organisations for detecting new trends, obtaining user opinions on their products and services or finding out about their online reputations. A recent research trend in SM analysis is making predictions based on sentiment analysis of SM. Often indicators of historic SM data are represented as time series and correlated with a variety of real world phenomena like the outcome of elections, the development of financial indicators, box office revenue and disease outbreaks. This paper examines the current state of research in the area of SM mining and predictive analysis and gives an overview of the analysis methods using opinion mining and machine learning techniques.

Keywords: social media, text mining, knowledge discovery, predictive analysis, machine learning

Procedia PDF Downloads 411

Authors: Ahmed Al-Awthan

Abstract:

This study addresses the prevalence of English terminology in published Arabic documentation on social media. Although the problem of using English terms in translation instead of existing native ones has been addressed in general by researchers around the world, to the best of the author’s knowledge the attitude of the translators as professionals to this phenomenon in Qatar and Yemen has not received a detailed study. This study examines the impact of the use of English, social media terms in the Arab world on aspiring and professional translators; it explores the benefits and drawbacks of linguistic borrowing as identified by the translators and investigates whether translators consider any means of resisting linguistic borrowing and prioritizing Arabic. It also aims to answer the following questions: i. Is there any prevalence of English, social media terms in Arabic translation? Why or why not? ii. Do Arabic translators prefer using English, social media terms to their equivalents in Arabic? If so, why? iii. Which measures could be adopted to help reduce the frequently observed borrowing of English terms? In particular, how do translators see the role of the Arabic Language Academies in preserving Arabic? iv. This research is descriptive, comparative and analytical in nature. It is both qualitative and quantitative. To validate the problem, the researcher will analyze articles published by Al-Jazeera in 2016-2018 that refer to the use of social media in diplomacy. It will be examined whether the increased international discussion of political events in social media increased the amount of transliterated English terminology referring to this mode of communication.To investigate whether the translators recognize the phenomenon of borrowing, the researcher proposes to use a survey. This survey will use multiple choice questions. It will target 20 aspiring translators from Yemen and 20 participants from Qatar. It will offer 15 English, social media terms used in discourse in 15 sentences. For each sentence, the researcher will provide three different translations and will ask the translators to rate them and offer their rendition. After collecting all the answers online, the researcher will analyze the data. The results are expected to confirm whether there is a prevalence of English terms in translating into Arabic. It is also expected to show what measures the translators used to render the English, social media terms, and it raises awareness of borrowing English terms. It will guide the translator toward using Arabicization methods in order to contribute to preserving Arabic.

Keywords: Arabicization, trans lingual borrowing, social media terms, terminology

Procedia PDF Downloads 139

Authors: Tahmidul Islam

Abstract:

Propensity Score Matching (PSM) technique has been widely used for estimating causal effect of treatment in observational studies. One major step of implementing PSM is estimating the propensity score (PS). Logistic regression model with additive linear terms of covariates is most used technique in many studies. Logistics regression model is also used with cubic splines for retaining flexibility in the model. However, choosing the functional form of the logistic regression model has been a question since the effectiveness of PSM depends on how accurately the PS been estimated. In many situations, the linearity assumption of linear logistic regression may not hold and non-linear relation between the logit and the covariates may be appropriate. One can estimate PS using machine learning techniques such as random forest, neural network etc for more accuracy in non-linear situation. In this study, an attempt has been made to compare the efficacy of Generalized Additive Model (GAM) in various linear and non-linear settings and compare its performance with usual logistic regression. GAM is a non-parametric technique where functional form of the covariates can be unspecified and a flexible regression model can be fitted. In this study various simple and complex models have been considered for treatment under several situations (small/large sample, low/high number of treatment units) and examined which method leads to more covariate balance in the matched dataset. It is found that logistic regression model is impressively robust against inclusion quadratic and interaction terms and reduces mean difference in treatment and control set equally efficiently as GAM does. GAM provided no significantly better covariate balance than logistic regression in both simple and complex models. The analysis also suggests that larger proportion of controls than treatment units leads to better balance for both of the methods.

Keywords: accuracy, covariate balances, generalized additive model, logistic regression, non-linearity, propensity score matching

Procedia PDF Downloads 354

Authors: Amit Verma, Pardeep Kaur

Abstract:

In software development organizations, Component based Software engineering (CBSE) is emerging paradigm for software development and gained wide acceptance as it often results in increase quality of software product within development time and budget. In component reusability, main challenges are the right component identification from large repositories at right time. The major objective of this work is to provide efficient algorithm for storage and effective retrieval of components using neural network and parameters based on user choice through clustering. This research paper aims to propose an algorithm that provides error free and automatic process (for retrieval of the components) while reuse of the component. In this algorithm, keywords (or components) are extracted from software document, after by applying k mean clustering algorithm. Then weights assigned to those keywords based on their frequency and after assigning weights, ANN predicts whether correct weight is assigned to keywords (or components) or not, otherwise it back propagates in to initial step (re-assign the weights). In last, store those all keywords into repositories for effective retrieval. Proposed algorithm is very effective in the error correction and detection with user base choice while choice of component for reusability for efficient retrieval is there.

Keywords: component based development, clustering, back propagation algorithm, keyword based retrieval

Procedia PDF Downloads 366

Authors: M. Draman, S. Z. Muhamad Yassin, M. S. Alias, Z. Lambak, M. I. Zulkifli, S. N. Padhi, K. N. Baharim, F. Maskuriy, A. I. A. Rahim

Abstract:

The performance of Malay automatic speech recognition (ASR) system for the call centre environment is presented. The system utilizes Kaldi toolkit as the platform to the entire library and algorithm used in performing the ASR task. The acoustic model implemented in this system uses a deep neural network (DNN) method to model the acoustic signal and the standard (n-gram) model for language modelling. With 80 hours of training data from the call centre recordings, the ASR system can achieve 72% of accuracy that corresponds to 28% of word error rate (WER). The testing was done using 20 hours of audio data. Despite the implementation of DNN, the system shows a low accuracy owing to the varieties of noises, accent and dialect that typically occurs in Malaysian call centre environment. This significant variation of speakers is reflected by the large standard deviation of the average word error rate (WERav) (i.e., ~ 10%). It is observed that the lowest WER (13.8%) was obtained from recording sample with a standard Malay dialect (central Malaysia) of native speaker as compared to 49% of the sample with the highest WER that contains conversation of the speaker that uses non-standard Malay dialect.

Keywords: conversational speech recognition, deep neural network, Malay language, speech recognition

Procedia PDF Downloads 309

Authors: Jan F. Talts, Amit Saxena, Kåre Engkilde

Abstract:

Chemotherapy-induced peripheral neuropathy (CIPN) is one of the most frequent side effects caused by anti-neoplastic agents, with a prevalence from 19 % to 85 %. Clinically, CIPN is a mostly sensory neuropathy leading to pain and to motor and autonomic changes. Due to its high prevalence among cancer patients, CIPN constitutes a major problem for both cancer patients and survivors, especially because currently, there is no single effective method of preventing CIPN. Hearing loss is the most common form of sensory impairment in humans and can be caused by ototoxic chemical compounds such as chemotherapy (platinum-based antineoplastic agents).In rodents, single or repeated cisplatin injections induce peripheral neuropathy and hearing impairment mimicking human disorder, allowing studying the efficacy of new pharmacological candidates in chemotherapy-induced hearing loss and peripheral neuropathy. RNA sequencing data from full term amniotic fluid (TAF) mesenchymal stemcell (MSC) clones was used to identify neural-specific markers present on TAF-MSC. Several prospective neural markers were tested by flow cytometry on cultured TAF-MSC. One of these markers was used for cell-sorting using Tyto MACSQuant cell sorter, and the neural marker positive cell population was expanded for several passages to the final therapeutic product stage. Peripheral neuropathy and hearing loss was induced in mice by administration of cisplatin in three week-long cycles. The efficacy of neural-specific TAF-MSC in treating hearing loss and pain perception was evaluated by administration of three injections of 3 million cells/kg by intravenous route or three injections of 3 million cells/kg by intra-arterial route after each cisplatin cycle treatment. Auditory brainstem responses (ABR) are electric potentials recorded from scalp electrodes, and the first ABR wave represents the summed activity of the auditory nerve fibers contacting the inner hair cells. For ABR studies, mice were anesthetized, then earphones were placed in the left ear of each mouse, an active electrode was placed in the vertex of the skull, a reference electrode under the skin of the mastoid bone, and a ground electrode in the neck skin. The stimuli consisted of tone pips of five frequencies (2, 4, 6, 12, 16, and 24 kHz) at various sound levels (from 0 to 90 dB) ranging to cover the mouse auditory frequency range. The von Frey test was used to assess the onset and maintenance of mechanical allodynia over time. Mice were placed in clear plexiglass cages on an elevated mesh floor and tested after 30 min of habituation. Mechanical paw withdrawal threshold was examined using an electronic von Frey anesthesiometer. Cisplatin groups treated with three injections of 3 million cells/kg by intravenous route and three injections of 3 million cells/kg by intra-arterial route after each cisplatin cycle treatment presented, a significant increase of hearing acuity characterized by a decrease of ABR threshold and a decrease of neuropathic pain characterized by an increase of von Frey paw withdrawal threshold compared to controls only receiving cisplatin. This study shows that treatment with MSCselected for neural specificity presents significant positive efficacy on the chemotherapy-induced neuropathic pain and the chemotherapy-induced hearing loss.

Keywords: mesenchymal stem cell, peripheral neuropathy, amniotic fluid, regenerative medicine

Procedia PDF Downloads 149

Authors: Shuen-Tai Wang, Yu-Ching Lin, Hsi-Ya Chang

Abstract:

Virtualization technologies are experiencing a renewed interest as a way to improve system reliability, and availability, reduce costs, and provide flexibility. This paper presents the development on leverage existing cloud infrastructure and virtualization tools. We adopted some virtualization technologies which improve portability, manageability and compatibility of applications by encapsulating them from the underlying operating system on which they are executed. Given the development of application virtualization, it allows shifting the user’s applications from the traditional PC environment to the virtualized environment, which is stored on a remote virtual machine rather than locally. This proposed effort has the potential to positively provide an efficient, resilience and elastic environment for online cloud service. Users no longer need to burden the platform maintenance and drastically reduces the overall cost of hardware and software licenses. Moreover, this flexible and web-based application virtualization service represent the next significant step to the mobile workplace, and it lets user executes their applications from virtually anywhere.

Keywords: cloud service, application virtualization, virtual machine, elastic environment

Procedia PDF Downloads 269

Authors: Mina Beheshti, Vida Shams, Mojtaba Esfandiari, Farzaneh Abdollahi, Abdolreza Ohadi

Abstract:

Most of the people who survive a stroke need rehabilitation tools to regain their mobility. The core function of these devices is a brake actuator. The goal of this study is to design and control a magnetorheological brake which can be used as a rehabilitation tool. In fact, the fluid used in this brake is called magnetorheological fluid or MR that properties can change by variation of the magnetic field. The braking properties can be set as control by using this feature of the fluid. In this research, different MR brake designs are first introduced in each design, and the dimensions of the brake have been determined based on the required torque for foot movement. To calculate the brake dimensions, it is assumed that the shear stress distribution in the fluid is uniform and the fluid is in its saturated state. After designing the rehabilitation brake, the mathematical model of the healthy movement of a healthy person is extracted. Due to the nonlinear nature of the system and its variability, various adaptive controllers, neural networks, and robust have been implemented to estimate the parameters and control the system. After calculating torque and control current, the best type of controller in terms of error and control current has been selected. Finally, this controller is implemented on the experimental data of the patient's movements, and the control current is calculated to achieve the desired torque and motion.

Keywords: rehabilitation, magnetorheological fluid, knee, brake, adaptive control, robust control, neural network control, torque control

Procedia PDF Downloads 135

Authors: Rachnavy Pornthep, Khaothin Thawichai

Abstract:

The purpose of this paper was to investigate the effects of rowing ergometer exercise on older persons health. The subjects were divided into two groups. Group 1 was control group (10 male and 10 female) Group 2 was experimental group (10 male and 10 female). The time for study was 12 week. Group 1 engage in normal daily activities Group 2 Training with rowing machine for 20 minutes three days a week. The average age of the experimental group was 73.7 years old, mean weight 55.4 kg, height 154.8 cm in the control group, mean age was 74.95 years, mean weight 48.6 kg, mean height 153.85 cm. Physical fitness test composted of body size, flexibility, Strength, muscle endurance and cardiovascular endurance. The comparison between the experimental and control groups before training showed that body weight, body mass index and waist to hip ratio were significantly different. The flexibility, strength, cardiovascular endurance was not significantly different. The comparison between the control group and the experimental group after training showed that body weight, body mass index and cardiovascular endurance were significantly different. The ratio of waist to hips, flexibility and muscular strength were not significantly different. Comparison of physical fitness before training and after training of the control group showed that body weight, flexibility (Sit and reach) and muscular strength (30 – Second chair stand) were significantly different. Body mass index, waist to hip ratio, muscles flexible (Shoulder girdle flexibility), muscle strength (30 – Second arm curl) and the cardiovascular endurance were not significantly difference. Comparison of physical fitness before training and after training the experimental group showed that waist to hip ratio, flexibility (sit and reach) muscle strength (30 – Second chair stand), cardiovascular endurance (Standing leg raises - up to 2 minutes) were significantly different. The Body mass index and the flexibility (Shoulder girdle flexibility) no significantly difference. The study found that exercising with rowing machine can improve the physical fitness of the elderly, especially the cardiovascular endurance, corresponding with the past research on the effects of exercise in the elderly with different exercise such as cycling, treadmill, walking on the elliptical machine. Therefore, we can conclude that exercise by using rowing machine can improve cardiovascular system and flexibility in the elderly.

Keywords: effect, rowing, exercise, elderly

Procedia PDF Downloads 482

Authors: Su Mon Saw, Joe Rothnagel

Abstract:

Short open reading frames (sORFs) located in 5’UTR of mRNAs are known as uORFs. Characterization of uORF-encoded peptides (uPEPs) i.e., a subset of short open reading frame encoded peptides (sPEPs) and their translation regulation lead to understanding of causes of genetic disease, proteome complexity and development of treatments. Existence of uORFs within cellular proteome could be detected by LC-MS/MS. The ability of uORF to be translated into uPEP and achievement of uPEP identification will allow uPEP’s characterization, structures, functions, subcellular localization, evolutionary maintenance (conservation in human and other species) and abundance in cells. It is hypothesized that a subset of sORFs are translatable and that their encoded sPEPs are functional and are endogenously expressed contributing to the eukaryotic cellular proteome complexity. This project aimed to investigate whether sORFs encode functional peptides. Liquid chromatography-mass spectrometry (LC-MS) and bioinformatics were thus employed. Due to probable low abundance of sPEPs and small in sizes, the need for efficient peptide enrichment strategies for enriching small proteins and depleting the sub-proteome of large and abundant proteins is crucial for identifying sPEPs. Low molecular weight proteins were extracted using SDS-PAGE from Human Embryonic Kidney (HEK293) cells and Strong Cation Exchange Chromatography (SCX) from secreted HEK293 cells. Extracted proteins were digested by trypsin to peptides, which were detected by LC-MS/MS. The MS/MS data obtained was searched against Swiss-Prot using MASCOT version 2.4 to filter out known proteins, and all unmatched spectra were re-searched against human RefSeq database. ProteinPilot v5.0.1 was used to identify sPEPs by searching against human RefSeq, Vanderperre and Human Alternative Open Reading Frame (HaltORF) databases. Potential sPEPs were analyzed by bioinformatics. Since SDS PAGE electrophoresis could not separate proteins <20kDa, this could not identify sPEPs. All MASCOT-identified peptide fragments were parts of main open reading frame (mORF) by ORF Finder search and blastp search. No sPEP was detected and existence of sPEPs could not be identified in this study. 13 translated sORFs in HEK293 cells by mass spectrometry in previous studies were characterized by bioinformatics. Identified sPEPs from previous studies were <100 amino acids and <15 kDa. Bioinformatics results showed that sORFs are translated to sPEPs and contribute to proteome complexity. uPEP translated from uORF of SLC35A4 was strongly conserved in human and mouse while uPEP translated from uORF of MKKS was strongly conserved in human and Rhesus monkey. Cross-species conserved uORFs in association with protein translation strongly suggest evolutionary maintenance of coding sequence and indicate probable functional expression of peptides encoded within these uORFs. Translation of sORFs was confirmed by mass spectrometry and sPEPs were characterized with bioinformatics.

Keywords: bioinformatics, HEK293 cells, liquid chromatography-mass spectrometry, ProteinPilot, Strong Cation Exchange Chromatography, SDS-PAGE, sPEPs

Procedia PDF Downloads 174

Authors: Orestis Κ. Efthymiou, Stavros T. Ponis

Abstract:

In the last decade, a new industrial revolution seems to be emerging, supported -once again- by the rapid advancements of Information Technology in the areas of Machine-to-Machine (M2M) communication permitting large numbers of intelligent devices, e.g. sensors to communicate with each other and take decisions without any or minimum indirect human intervention. The advent of these technologies have triggered the emergence of a new category of hybrid (cyber-physical) manufacturing systems, combining advanced manufacturing techniques with innovative M2M applications based on the Internet of Things (IoT), under the umbrella term Industry 4.0. Even though the topic of Industry 4.0 has attracted much attention during the last few years, the attempts of providing a systematic literature review of the subject are scarce. In this paper, we present the authors’ initial study of the field with a special focus on the use and applications of Industry 4.0 principles in material handling automations and in-house logistics. Research shows that despite the vivid discussion and attractiveness of the subject, there are still many challenges and issues that have to be addressed before Industry 4.0 becomes standardized and widely applicable.

Keywords: Industry 4.0, internet of things, manufacturing systems, material handling, logistics

Procedia PDF Downloads 114

Authors: Oumaima Lahmar

Abstract:

This paper aims to define a structured topography for finance researchers seeking to navigate the body of knowledge in their extrapolation of finance phenomena. To make sense of the body of knowledge in finance, a probabilistic topic modeling approach is applied on 6000 abstracts of academic articles published in three top journals in finance between 1976 and 2020. This approach combines both machine learning techniques and natural language processing to statistically identify the conjunctions between research articles and their shared topics described each by relevant keywords. The topic modeling analysis reveals 35 coherent topics that can well depict finance literature and provide a comprehensive structure for the ongoing research themes. Comparing the extracted topics to the Journal of Economic Literature (JEL) classification system, a significant similarity was highlighted between the characterizing keywords. On the other hand, we identify other topics that do not match the JEL classification despite being relevant in the finance literature.

Keywords: finance literature, textual analysis, topic modeling, perplexity

Procedia PDF Downloads 149

Authors: Sidharth Talan, Sanjay Kumar Sharma, Eoin Joseph Wallace, Nikita Agrawal

Abstract:

Underground mining is at the core of rapidly evolving metals and minerals sector due to the increasing mineral consumption globally. Even though the surface mines are still more abundant on earth, the scales of industry are slowly tipping towards underground mining due to rising depth and complexities of orebodies. Thus, the efficient and productive functioning of underground operations depends significantly on the synchronized performance of key elements such as operating site, mining equipment, manpower and mine services. Production drilling is the process of conducting long hole drilling for the purpose of charging and blasting these holes for the production of ore in underground metal mines. Thus, production drilling is the crucial segment in the underground metal mining value chain. This paper presents the process mapping tool to evaluate the production drilling process in the underground metal mining operation by dividing the given process into three segments namely Input, Process and Output. The three segments are further segregated into factors and sub-factors. As per the study, the major input factors crucial for the efficient functioning of production drilling process are power, drilling water, geotechnical support of the drilling site, skilled drilling operators, services installation crew, oils and drill accessories for drilling machine, survey markings at drill site, proper housekeeping, regular maintenance of drill machine, suitable transportation for reaching the drilling site and finally proper ventilation. The major outputs for the production drilling process are ore, waste as a result of dilution, timely reporting and investigation of unsafe practices, optimized process time and finally well fragmented blasted material within specifications set by the mining company. The paper also exhibits the drilling loss matrix, which is utilized to appraise the loss in planned production meters per day in a mine on account of availability loss in the machine due to breakdowns, underutilization of the machine and productivity loss in the machine measured in drilling meters per unit of percussion hour with respect to its planned productivity for the day. The given three losses would be essential to detect the bottlenecks in the process map of production drilling operation so as to instigate the action plan to suppress or prevent the causes leading to the operational performance deficiency. The given tool is beneficial to mine management to focus on the critical factors negatively impacting the production drilling operation and design necessary operational and maintenance strategies to mitigate them. 

Keywords: process map, drilling loss matrix, SIPOC, productivity, percussion rate

Procedia PDF Downloads 197

Authors: Louise Moles, Steve Riley, Sarah D. Lichenstein, Marzieh Babaeianjelodar, Robert Kohler, Annie Cheng, Corey Horien Abigail Greene, Wenjing Luo, Jonathan Ahern, Bohan Xu, Yize Zhao, Chun Chieh Fan, R. Todd Constable, Sarah W. Yip

Abstract:

Background: Risks for depression are myriad and include both genetic and brain-based factors. However, relationships between these systems are poorly understood, limiting understanding of disease etiology, particularly at the developmental level. Methods: We use a data-driven machine learning approach connectome-based predictive modeling (CPM) to identify functional connectivity signatures associated with polygenic risk scores for depression (DEP-PRS) among youth from the Adolescent Brain and Cognitive Development (ABCD) study across diverse brain states, i.e., during resting state, during affective working memory, during response inhibition, during reward processing. Results: Using 10-fold cross-validation with 100 iterations and permutation testing, CPM identified connectivity signatures of DEP-PRS across all examined brain states (rho’s=0.20-0.27, p’s<.001). Across brain states, DEP-PRS was positively predicted by increased connectivity between frontoparietal and salience networks, increased motor-sensory network connectivity, decreased salience to subcortical connectivity, and decreased subcortical to motor-sensory connectivity. Subsampling analyses demonstrated that model accuracies were robust across random subsamples of N’s=1,000, N’s=500, and N’s=250 but became unstable at N’s=100. Conclusions: These data, for the first time, identify neural networks of polygenic depression risk in a large sample of youth before the onset of significant clinical impairment. Identified networks may be considered potential treatment targets or vulnerability markers for depression risk.

Keywords: genetics, functional connectivity, pre-adolescents, depression

Procedia PDF Downloads 38