Search results for: bidirectional LSTM

Authors: Esther Chiramal, Kelvin Soh Boon Kai

Abstract:

Explainable AI is a strong strategy implemented to understand complex black-box model predictions in a human-interpretable language. It provides the evidence required to execute the use of trustworthy and reliable AI systems. On the other hand, however, it also opens the door to locating possible vulnerabilities in an AI model. Traditional adversarial text attack uses word substitution, data augmentation techniques, and gradient-based attacks on powerful pre-trained Bidirectional Encoder Representations from Transformers (BERT) variants to generate adversarial sentences. These attacks are generally white-box in nature and not practical as they can be easily detected by humans e.g., Changing the word from “Poor” to “Rich”. We proposed a simple yet effective Grey-box cum Black-box approach that does not require the knowledge of the model while using a set of surrogate Transformer/BERT models to perform the attack using Explainable AI techniques. As Transformers are the current state-of-the-art models for almost all Natural Language Processing (NLP) tasks, an attack generated from BERT1 is transferable to BERT2. This transferability is made possible due to the attention mechanism in the transformer that allows the model to capture long-range dependencies in a sequence. Using the power of BERT generalisation via attention, we attempt to exploit how transformers learn by attacking a few surrogate transformer variants which are all based on a different architecture. We demonstrate that this approach is highly effective to generate semantically good sentences by changing as little as one word that is not detectable by humans while still fooling other BERT models.

Keywords: BERT, explainable AI, Grey-box text attack, transformer

Procedia PDF Downloads 117

Authors: Amber Brittain-Hale

Abstract:

This research investigates the evolving role of social media in crisis situations by employing discourse analysis methodology and honing in on the Russo-Ukrainian War, particularly Ukraine's use of LinkedIn. The study posits that social media platforms, such as LinkedIn, play a crucial role in shaping communication, disseminating information, and influencing geopolitical strategies during conflicts. Focusing on Ukraine's official state account on LinkedIn and analyzing its posts and interactions, the research aims to unveil discourse dynamics in high-stakes scenarios and provide valuable insights for leaders navigating complex global challenges. A comprehensive analysis of the data will contribute to a deeper understanding of the tactics adopted by political leaders in managing communication, the bidirectional nature of discourse provided by online social networks, and the rapid advancement of technology that has led to the growing significance of social media platforms in crisis situations. Through this approach, the geopolitical factors that influenced the country's social media strategy during the Russo-Ukrainian War will be illuminated, offering a broader perspective on the role of social media in such challenging times. Ultimately, the study seeks to uncover lessons that can be drawn from Ukraine's LinkedIn approach, informing future strategies for utilizing social media during crises and advancing the understanding of how social media can be harnessed to address intricate global issues.

Keywords: russo-ukrainian war, social media, crisis, discourse analysis

Procedia PDF Downloads 89

Authors: Joseph George, Anne Kotteswara Roa

Abstract:

Skin disease is one of the most common and popular kinds of health issues faced by people nowadays. Skin cancer (SC) is one among them, and its detection relies on the skin biopsy outputs and the expertise of the doctors, but it consumes more time and some inaccurate results. At the early stage, skin cancer detection is a challenging task, and it easily spreads to the whole body and leads to an increase in the mortality rate. Skin cancer is curable when it is detected at an early stage. In order to classify correct and accurate skin cancer, the critical task is skin cancer identification and classification, and it is more based on the cancer disease features such as shape, size, color, symmetry and etc. More similar characteristics are present in many skin diseases; hence it makes it a challenging issue to select important features from a skin cancer dataset images. Hence, the skin cancer diagnostic accuracy is improved by requiring an automated skin cancer detection and classification framework; thereby, the human expert’s scarcity is handled. Recently, the deep learning techniques like Convolutional neural network (CNN), Deep belief neural network (DBN), Artificial neural network (ANN), Recurrent neural network (RNN), and Long and short term memory (LSTM) have been widely used for the identification and classification of skin cancers. This survey reviews different DL techniques for skin cancer identification and classification. The performance metrics such as precision, recall, accuracy, sensitivity, specificity, and F-measures are used to evaluate the effectiveness of SC identification using DL techniques. By using these DL techniques, the classification accuracy increases along with the mitigation of computational complexities and time consumption.

Keywords: skin cancer, deep learning, performance measures, accuracy, datasets

Procedia PDF Downloads 103

Authors: Soheila Sadeghi

Abstract:

Accurate forecasting of project performance metrics is crucial for successfully managing and delivering urban road reconstruction projects. Traditional methods often rely on static baseline plans and fail to consider the dynamic nature of project progress and external factors. This research proposes a machine learning-based approach to forecast project performance metrics, such as cost variance and earned value, for each Work Breakdown Structure (WBS) category in an urban road reconstruction project. The proposed model utilizes time series forecasting techniques, including Autoregressive Integrated Moving Average (ARIMA) and Long Short-Term Memory (LSTM) networks, to predict future performance based on historical data and project progress. The model also incorporates external factors, such as weather patterns and resource availability, as features to enhance the accuracy of forecasts. By applying the predictive power of machine learning, the performance forecasting model enables proactive identification of potential deviations from the baseline plan, which allows project managers to take timely corrective actions. The research aims to validate the effectiveness of the proposed approach using a case study of an urban road reconstruction project, comparing the model's forecasts with actual project performance data. The findings of this research contribute to the advancement of project management practices in the construction industry, offering a data-driven solution for improving project performance monitoring and control.

Keywords: project performance forecasting, machine learning, time series forecasting, cost variance, earned value management

Procedia PDF Downloads 8

Authors: Jinhoa Lee

Abstract:

The relationships between environmental quality, energy use and economic output have created growing attention over the past decades among researchers and policy makers. Focusing on the empirical aspects of the role of carbon dioxide (CO2) emissions and energy use in affecting the economic output, this paper is an effort to fulfill the gap in a comprehensive case study at a country level using modern econometric techniques. To achieve the goal, this country-specific study examines the short-run and long-run relationships among energy consumption (using disaggregated energy sources: crude oil, coal, natural gas, and electricity), CO2 emissions and gross domestic product (GDP) for Turkey using time series analysis from the year 1980-2010. To investigate the relationships between the variables, this paper employs the Augmented Dickey-Fuller (ADF) test for stationarity, Johansen’s maximum likelihood method for cointegration and a Vector Error Correction Model (VECM) for both short- and long-run causality among the research variables for the sample. The long-run equilibrium in the VECM suggests no effects of the CO2 emissions and energy use on the GDP in Turkey. There exists a short-run bidirectional relationship between the electricity and natural gas consumption, and also there is a negative unidirectional causality running from the GDP to electricity use. Overall, the results partly support arguments that there are relationships between energy use and economic output; however, the effects may differ due to the source of energy such as in the case of Turkey for the period of 1980-2010. However, there is no significant relationship between the CO2 emissions and the GDP and between the CO2 emissions and the energy use both in the short term and long term.

Keywords: CO2 emissions, energy consumption, GDP, Turkey, time series analysis

Procedia PDF Downloads 489

Authors: Pius Marthin, Nihal Ata Tutkun

Abstract:

Survival analysis has become one of the paramount procedures in the modeling of time-to-event data. When we encounter complex survival problems, the traditional approach remains limited in accounting for the complex correlational structure between the covariates and the outcome due to the strong assumptions that limit the inference and prediction ability of the resulting models. Several studies exist on the deep learning approach to survival modeling; moreover, the application for the case of complex survival problems still needs to be improved. In addition, the existing models need to address the data structure's complexity fully and are subject to noise and redundant information. In this study, we design a deep learning technique (CmpXRnnSurv_AE) that obliterates the limitations imposed by traditional approaches and addresses the above issues to jointly predict the risk-specific probabilities and survival function for recurrent events with competing risks. We introduce the component termed Risks Information Weights (RIW) as an attention mechanism to compute the weighted cumulative incidence function (WCIF) and an external auto-encoder (ExternalAE) as a feature selector to extract complex characteristics among the set of covariates responsible for the cause-specific events. We train our model using synthetic and real data sets and employ the appropriate metrics for complex survival models for evaluation. As benchmarks, we selected both traditional and machine learning models and our model demonstrates better performance across all datasets.

Keywords: cumulative incidence function (CIF), risk information weight (RIW), autoencoders (AE), survival analysis, recurrent events with competing risks, recurrent neural networks (RNN), long short-term memory (LSTM), self-attention, multilayers perceptrons (MLPs)

Procedia PDF Downloads 68

Authors: Dake Xiong, Ben Hankamer, Ian Ross

Abstract:

The most urgent challenge of our time is to replace the depleting resources of fossil fuels by sustainable environmentally friendly alternatives. Hydrogen is a promising CO2-neutral fuel for a more sustainable future especially when produced photo-biologically. Hydrogen can be photosynthetically produced in unicellular green alga like Chlamydomonas reinhardtii, catalysed by the inducible highly active and bidirectional [FeFe]-hydrogenase enzymes (HydA). However, evolutionary and physiological constraints severely restrict the hydrogen yield of algae for industrial scale-up, mainly due to its competition among other metabolic pathways on photosynthetic electrons. Among them, a major challenge to be resolved is the inferior competitiveness of hydrogen production (catalysed by HydA) with NADPH production (catalysed by ferredoxin-NADP+-reductase (FNR)), which is essential for cell growth and takes up ~95% of photosynthetic electrons. In this work, the in vivo hydrogen production efficiency of mutants with ferredoxin-hydrogenase (Fd*-HydA1*) fusion protein construct, where the electron donor ferredoxin (Fd*) is fused to HydA1* and expressed in the model organism C. reinhardtii was investigated. Once Fd*-HydA1* fusion gene is expressed in algal cells, the fusion enzyme is able to draw the redistributed photosynthetic electrons and use them for efficient hydrogen production. From preliminary data, mutants with Fd*-HydA1* transgene showed a ~2-fold increase in the photosynthetic hydrogen production rate compared with its parental strain, which only possesses the native HydA in vivo. Therefore, a solid method of having more efficient hydrogen production in microalgae can be achieved through the expression of the synthetic enzymes.

Keywords: Chlamydomonas reinhardtii, ferredoxin, fusion protein, hydrogen production, hydrogenase

Procedia PDF Downloads 238

Authors: Javad Jamali Khouei, Mohammadreza Khoshravan

Abstract:

Considering that application of functionally graded material is increasing in most industries, it seems necessary to present a methodology for designing optimal profile of structures such as plate under mechanical loading which is highly consumed in industries. Therefore, volume fraction variation profile of functionally graded circular plate which has been considered two-directional is optimized so that stress of structure is minimized. For this purpose, equilibrium equations of two-directional functionally graded circular plate are solved by applying semi analytical-numerical method under mechanical loading and support conditions. By solving equilibrium equations, deflections and stresses are obtained in terms of control variables of volume fraction variation profile. As a result, the problem formula can be defined as an optimization problem by aiming at minimization of critical von-mises stress under constraints of deflections, stress and a physical constraint relating to structure of material. Then, the related problem can be solved with help of one of the metaheuristic algorithms such as genetic algorithm. Results of optimization for the applied model under constraints and loadings and boundary conditions show that functionally graded plate should be graded only in radial direction and there is no need for volume fraction variation of the constituent particles in thickness direction. For validating results, optimal values of the obtained design variables are graphically evaluated.

Keywords: two-directional functionally graded material, single objective optimization, semi analytical-numerical solution, genetic algorithm, graphical solution with contour

Procedia PDF Downloads 259

Authors: Soheila Sadeghi

Abstract:

In construction projects, predicting project performance metrics accurately is essential for effective management and successful delivery. However, conventional methods often depend on fixed baseline plans, disregarding the evolving nature of project progress and external influences. To address this issue, we introduce a distinct approach based on machine learning to forecast key performance indicators, such as cost variance and earned value, for each Work Breakdown Structure (WBS) category within an urban road reconstruction project. Our proposed model leverages time series forecasting techniques, namely Autoregressive Integrated Moving Average (ARIMA) and Long Short-Term Memory (LSTM) networks, to predict future performance by analyzing historical data and project progress. Additionally, the model incorporates external factors, including weather patterns and resource availability, as features to improve forecast accuracy. By harnessing the predictive capabilities of machine learning, our performance forecasting model enables project managers to proactively identify potential deviations from the baseline plan and take timely corrective measures. To validate the effectiveness of the proposed approach, we conduct a case study on an urban road reconstruction project, comparing the model's predictions with actual project performance data. The outcomes of this research contribute to the advancement of project management practices in the construction industry by providing a data-driven solution for enhancing project performance monitoring and control.

Keywords: project performance forecasting, machine learning, time series forecasting, cost variance, schedule variance, earned value management

Procedia PDF Downloads 9

Authors: Bahareh Golchin, Nooshin Riahi

Abstract:

One of the most significant issues as attended a lot in recent years is that of recognizing the sentiments and emotions in social media texts. The analysis of sentiments and emotions is intended to recognize the conceptual information such as the opinions, feelings, attitudes and emotions of people towards the products, services, organizations, people, topics, events and features in the written text. These indicate the greatness of the problem space. In the real world, businesses and organizations are always looking for tools to gather ideas, emotions, and directions of people about their products, services, or events related to their own. This article uses the Twitter social network, one of the most popular social networks with about 420 million active users, to extract data. Using this social network, users can share their information and opinions about personal issues, policies, products, events, etc. It can be used with appropriate classification of emotional states due to the availability of its data. In this study, supervised learning and deep neural network algorithms are used to classify the emotional states of Twitter users. The use of deep learning methods to increase the learning capacity of the model is an advantage due to the large amount of available data. Tweets collected on various topics are classified into four classes using a combination of two Bidirectional Long Short Term Memory network and a Convolutional network. The results obtained from this study with an average accuracy of 93%, show good results extracted from the proposed framework and improved accuracy compared to previous work.

Keywords: emotion classification, sentiment analysis, social networks, deep neural networks

Procedia PDF Downloads 119

Authors: Sangita Pokhrel, Nalinda Somasiri, Rebecca Jeyavadhanam, Swathi Ganesan

Abstract:

Tourism is a booming industry with huge future potential for global wealth and employment. There are countless data generated over social media sites every day, creating numerous opportunities to bring more insights to decision-makers. The integration of Big Data Technology into the tourism industry will allow companies to conclude where their customers have been and what they like. This information can then be used by businesses, such as those in charge of managing visitor centers or hotels, etc., and the tourist can get a clear idea of places before visiting. The technical perspective of natural language is processed by analysing the sentiment features of online reviews from tourists, and we then supply an enhanced long short-term memory (LSTM) framework for sentiment feature extraction of travel reviews. We have constructed a web review database using a crawler and web scraping technique for experimental validation to evaluate the effectiveness of our methodology. The text form of sentences was first classified through Vader and Roberta model to get the polarity of the reviews. In this paper, we have conducted study methods for feature extraction, such as Count Vectorization and TFIDF Vectorization, and implemented Convolutional Neural Network (CNN) classifier algorithm for the sentiment analysis to decide the tourist’s attitude towards the destinations is positive, negative, or simply neutral based on the review text that they posted online. The results demonstrated that from the CNN algorithm, after pre-processing and cleaning the dataset, we received an accuracy of 96.12% for the positive and negative sentiment analysis.

Keywords: counter vectorization, convolutional neural network, crawler, data technology, long short-term memory, web scraping, sentiment analysis

Procedia PDF Downloads 67

Authors: P. Karthick, K. Mahesh

Abstract:

Video has become an increasingly significant component of our digital everyday contact. With the advancement of greater contents and shows of the resolution, its significant volume poses serious obstacles to the objective of receiving, distributing, compressing, and revealing video content of high quality. In this paper, we propose the primary beginning to complete a deep video compression model that jointly upgrades all video compression components. The video compression method involves splitting the video into frames, comparing the images using convolutional neural networks (CNN) to remove duplicates, repeating the single image instead of the duplicate images by recognizing and detecting minute changes using generative adversarial network (GAN) and recorded with long short-term memory (LSTM). Instead of the complete image, the small changes generated using GAN are substituted, which helps in frame level compression. Pixel wise comparison is performed using K-nearest neighbours (KNN) over the frame, clustered with K-means, and singular value decomposition (SVD) is applied for each and every frame in the video for all three color channels [Red, Green, Blue] to decrease the dimension of the utility matrix [R, G, B] by extracting its latent factors. Video frames are packed with parameters with the aid of a codec and converted to video format, and the results are compared with the original video. Repeated experiments on several videos with different sizes, duration, frames per second (FPS), and quality results demonstrate a significant resampling rate. On average, the result produced had approximately a 10% deviation in quality and more than 50% in size when compared with the original video.

Keywords: video compression, K-means clustering, convolutional neural network, generative adversarial network, singular value decomposition, pixel visualization, stochastic gradient descent, frame per second extraction, RGB channel extraction, self-detection and deciding system

Procedia PDF Downloads 165

Authors: Runlin Liu, Shilong Li

Abstract:

China is a mountainous country, with two-thirds of its territory covered by plateaus, hills, and mountains, and nearly half of the cities and towns are distributed in mountainous areas. Compared with the environmental constraints in the development path of cities and towns in the plains, there are heterogeneities in aspects such as spatial characteristics, growth mode, and ecological protection and so on for mountain village and townlets, and the development path of mountain village and townlets has a bidirectional relationship between mountain space and economic growth. Based on classical growth theory, this article explores the two-dimensional coupling relation between space and economy in mountain village and townlets under background of rural rejuvenation. GIS technology is adopted in the study to analyze spatial trends and patterns, economical spatial differentiation characteristics of village and townlets. This powerful tool can also help differentiate and analyze limiting factors and assessment systems in the economic growth of village and townlets under spatial dimension of mountainous space. To make the research more specific, this article selects mountain village and townlets in Southwest China as the object of study; this provides good cases for analyzing parallel coupling mechanism of the duality structure system between economic growth and spatial expansion and discussing the path selection of spatial economic growth of mountain village and towns with multiple constraints. The research results can provide quantitative references for the spatial and economic development paths of mountain villages and towns, which is helpful in realizing efficient and high-quality development mode with equal emphasis on spatial and economic benefits for these type of towns.

Keywords: coupling mechanism, geographic information technology, mountainous town, synergetic development, spatial economy

Procedia PDF Downloads 127

Authors: Mohammed Abu Shquier

Abstract:

This paper introduces a word segmentation method using the novel BP-LSTM-CRF architecture for processing semantic output training. The objective of web morphological analysis tools is to link a formal morpho-syntactic description to a lemma, along with morpho-syntactic information, a vocalized form, a vocalized analysis with morpho-syntactic information, and a list of paradigms. A key objective is to continuously enhance the proposed system through an inductive learning approach that considers semantic influences. The system is currently under construction and development based on data-driven learning. To evaluate the tool, an experiment on homograph analysis was conducted. The tool also encompasses the assumption of deep binary segmentation hypotheses, the arbitrary choice of trigram or n-gram continuation probabilities, language limitations, and morphology for both Modern Standard Arabic (MSA) and Dialectal Arabic (DA), which provide justification for updating this system. Most Arabic word analysis systems are based on the phonotactic morpho-syntactic analysis of a word transmitted using lexical rules, which are mainly used in MENA language technology tools, without taking into account contextual or semantic morphological implications. Therefore, it is necessary to have an automatic analysis tool taking into account the word sense and not only the morpho-syntactic category. Moreover, they are also based on statistical/stochastic models. These stochastic models, such as HMMs, have shown their effectiveness in different NLP applications: part-of-speech tagging, machine translation, speech recognition, etc. As an extension, we focus on language modeling using Recurrent Neural Network (RNN); given that morphological analysis coverage was very low in dialectal Arabic, it is significantly important to investigate deeply how the dialect data influence the accuracy of these approaches by developing dialectal morphological processing tools to show that dialectal variability can support to improve analysis.

Keywords: NLP, DL, ML, analyser, MSA, RNN, CNN

Procedia PDF Downloads 19

Authors: Konstantinos G. Kostinakis, Manthos K. Papadopoulos, Asimina M. Athanatopoulou

Abstract:

An important area of research in seismic risk analysis is the evaluation of expected seismic damage of structures under a specific earthquake ground motion. Several conventional intensity measures of ground motion have been used to estimate their damage potential to structures. Yet, none of them was proved to be able to predict adequately the seismic damage of any structural system. Therefore, alternative advanced intensity measures which take into account not only ground motion characteristics but also structural information have been proposed. The adequacy of a number of advanced earthquake intensity measures in prediction of structural damage of 3D R/C buildings under seismic excitation which attacks the building with arbitrary incident angle is investigated in the present paper. To achieve this purpose, a symmetric in plan and an asymmetric 5-story R/C building are studied. The two buildings are subjected to 20 bidirectional earthquake ground motions. The two horizontal accelerograms of each ground motion are applied along horizontal orthogonal axes forming 72 different angles with the structural axes. The response is computed by non-linear time history analysis. The structural damage is expressed in terms of the maximum interstory drift as well as the overall structural damage index. The values of the aforementioned seismic damage measures determined for incident angle 0° as well as their maximum values over all seismic incident angles are correlated with 9 structure-specific ground motion intensity measures. The research identified certain intensity measures which exhibited strong correlation with the seismic damage of the two buildings. However, their adequacy for estimation of the structural damage depends on the response parameter adopted. Furthermore, it was confirmed that the widely used spectral acceleration at the fundamental period of the structure is a good indicator of the expected earthquake damage level.

Keywords: damage indices, non-linear response, seismic excitation angle, structure-specific intensity measures

Procedia PDF Downloads 482

Authors: Li-Ting Huang, Yu-Hsiang Shen, Cing-Ciao Ke, Sheng-Pin Tseng, Fan-Pin Tseng, Yu-Ching Ni, Chia-Yu Lin

Abstract:

Chest radiography is the most common technique for the diagnosis and follow-up of pulmonary diseases. However, the lesions superimposed with normal structures are difficult to be detected in chest radiography. Chest tomosynthesis is a relatively new technique to obtain 3D section images from a set of low-dose projections acquired over a limited angular range. However, there are some limitations with chest tomosynthesis. Patients undergoing tomosynthesis have to be able to hold their breath firmly for 10 seconds. A digital tomosynthesis system with advanced reconstruction algorithm and high-stability motion mechanism was developed by our research group. The potential for the system to perform a bidirectional chest scan within 10 seconds is expected. The purpose of this study is to realize the influences of the scan speed on the image quality for our digital tomosynthesis system. The major factors that lead image blurring are the motion of the X-ray source and the patient. For the fore one, an experiment of imaging a chest phantom with three different scan speeds, which are 6 cm/s, 8 cm/s, and 15 cm/s, was proceeded to understand the scan speed influences on the image quality. For the rear factor, a normal SD (Sprague-Dawley) rat was imaged with it alive and sacrificed to assess the impact on the image quality due to breath motion. In both experiments, the profile of the ROIs (region of interest) and the CNRs (contrast-to-noise ratio) of the ROIs to the normal tissue of the reconstructed images was examined to realize the degradations of the qualities of the images. The preliminary results show that no obvious degradation of the image quality was observed with increasing scan speed, possibly due to the advanced designs for the hardware and software of the system. It implies that higher speed (15 cm/s) than that of the commercialized tomosynthesis system (12 cm/s) for the proposed system is achieved, and therefore a complete chest scan within 10 seconds is expected.

Keywords: chest radiography, digital tomosynthesis, image quality, scan speed

Procedia PDF Downloads 308

Authors: Liwiusz Wojciechowski

Abstract:

Research background: The current knowledge does not give a clear answer to the question of the impact of FDI on productivity. Results of the empirical studies are still inconclusive, no matter how extensive and diverse in terms of research approaches or groups of countries analyzed they are. It should also take into account the possibility that FDI and productivity are linked and that there is a bidirectional relationship between them. This issue is particularly important because on one hand FDI can contribute to changes in productivity in the host country, but on the other hand its level and dynamics may imply that FDI should be undertaken in a given country. As already mentioned, a two-way relationship between the presence of foreign capital and productivity in the host country should be assumed, taking into consideration the endogenous nature of FDI. Purpose of the article: The overall objective of this study is to determine the causality between foreign direct investment and total factor productivity in host county in terms of different relative absorptive capacity across countries. In the classic sense causality among variables is not always obvious and requires for testing, which would facilitate proper specification of FDI models. The aim of this article is to study endogeneity of selected macroeconomic variables commonly being used in FDI models in case of Visegrad countries: main recipients of FDI in CEE. The findings may be helpful in determining the structure of the actual relationship between variables, in appropriate models estimation and in forecasting as well as economic policymaking. Methodology/methods: Panel and time-series data techniques including GMM estimator, VEC models and causality tests were utilized in this study. Findings & Value added: The obtained results allow to confirm the hypothesis states the bi-directional causality between FDI and total factor productivity. Although results differ from among countries and data level of aggregation implications may be useful for policymakers in case of providing foreign capital attracting policy.

Keywords: endogeneity, foreign direct investment, multi-equation models, total factor productivity

Procedia PDF Downloads 182

Authors: Thiago Spilborghs Bueno Meyer, Plinio Thomaz Aquino Junior

Abstract:

Through speech, which privileges the functional and interactive nature of the text, it is possible to ascertain the spatiotemporal circumstances, the conditions of production and reception of the discourse, the explicit purposes such as informing, explaining, convincing, etc. These conditions allow bringing the interaction between humans closer to the human-robot interaction, making it natural and sensitive to information. However, it is not enough to understand what is said; it is necessary to recognize emotions for the desired interaction. The validity of the use of neural networks for feature selection and emotion recognition was verified. For this purpose, it is proposed the use of neural networks and comparison of models, such as recurrent neural networks and deep neural networks, in order to carry out the classification of emotions through speech signals to verify the quality of recognition. It is expected to enable the implementation of robots in a domestic environment, such as the HERA robot from the RoboFEI@Home team, which focuses on autonomous service robots for the domestic environment. Tests were performed using only the Mel-Frequency Cepstral Coefficients, as well as tests with several characteristics of Delta-MFCC, spectral contrast, and the Mel spectrogram. To carry out the training, validation and testing of the neural networks, the eNTERFACE’05 database was used, which has 42 speakers from 14 different nationalities speaking the English language. The data from the chosen database are videos that, for use in neural networks, were converted into audios. It was found as a result, a classification of 51,969% of correct answers when using the deep neural network, when the use of the recurrent neural network was verified, with the classification with accuracy equal to 44.09%. The results are more accurate when only the Mel-Frequency Cepstral Coefficients are used for the classification, using the classifier with the deep neural network, and in only one case, it is possible to observe a greater accuracy by the recurrent neural network, which occurs in the use of various features and setting 73 for batch size and 100 training epochs.

Keywords: emotion recognition, speech, deep learning, human-robot interaction, neural networks

Procedia PDF Downloads 137

Authors: Liliia N. Butymova, Vladimir Y. Modorsky, Nikolai A. Shevelev

Abstract:

Prediction of vibration processes in gas transmittal units (GTU) is an urgent problem. Despite numerous scientific publications on the problem of vibrations in general, there are not enough works concerning FSI-modeling interaction processes between several deformable blades in gas-dynamic flow. Since it is very difficult to solve the problem in full scope, with all factors considered, a unidirectional dynamic coupled 1FSI model is suggested for use at the first stage, which would include, from symmetry considerations, two blades, which might be considered as the first stage of solving more general bidirectional problem. ANSYS CFX programmed multi-processor was chosen as a numerical computation tool. The problem was solved on PNRPU high-capacity computer complex. At the first stage of the study, blades were believed oscillating with the same frequency, although oscillation phases could be equal and could be different. At that non-stationary gas-dynamic forces distribution over the blades surfaces is calculated in run of simulation experiment. Oscillations in the “gas — structure” dynamic system are assumed to increase if the resultant of these gas-dynamic forces is in-phase with blade oscillation, and phase shift (φ=0). Provided these oscillation occur with phase shift, then oscillations might increase or decrease, depending on the phase shift value. The most important results are as follows: the angle of phase shift in inter-blade oscillation and the gas-dynamic force depends on the flow velocity, the specific inter-blade gap, and the shaft rotation speed; a phase shift in oscillation of adjacent blades does not always correspond to phase shift of gas-dynamic forces affecting the blades. Thus, it was discovered, that asynchronous oscillation of blades might cause either attenuation or intensification of oscillation. It was revealed that clocking effect might depend not only on the mutual circumferential displacement of blade rows and the gap between the blades, but also on the blade dynamic deformation nature.

Keywords: aeroelasticity, ANSYS CFX, oscillation, phase shift, clocking effect, vibrations

Procedia PDF Downloads 248

Authors: Bliss Singhal

Abstract:

Epilepsy is a prevalent neurological disorder affecting approximately 50 million individuals worldwide and 1.2 million Americans. There exist millions of pediatric patients with intractable epilepsy, a condition in which seizures fail to come under control. The occurrence of seizures can result in physical injury, disorientation, unconsciousness, and additional symptoms that could impede children's ability to participate in everyday tasks. Predicting seizures can help parents and healthcare providers take precautions, prevent risky situations, and mentally prepare children to minimize anxiety and nervousness associated with the uncertainty of a seizure. This research proposes a comprehensive framework to predict seizures in pediatric patients by evaluating machine learning algorithms on unimodal neuroimaging data consisting of electroencephalogram signals. The bandpass filtering and independent component analysis proved to be effective in reducing the noise and artifacts from the dataset. Various machine learning algorithms’ performance is evaluated on important metrics such as accuracy, precision, specificity, sensitivity, F1 score and MCC. The results show that the deep learning algorithms are more successful in predicting seizures than logistic Regression, and k nearest neighbors. The recurrent neural network (RNN) gave the highest precision and F1 Score, long short-term memory (LSTM) outperformed RNN in accuracy and convolutional neural network (CNN) resulted in the highest Specificity. This research has significant implications for healthcare providers in proactively managing seizure occurrence in pediatric patients, potentially transforming clinical practices, and improving pediatric care.

Keywords: intractable epilepsy, seizure, deep learning, prediction, electroencephalogram channels

Procedia PDF Downloads 59

Authors: Andrea Vega-Tinoco, Ana I. Gil-Lacruz, Marta Gil-Lacruz

Abstract:

The main objective of this research is to acknowledge whether civic participation affects the well-being of the elderly, thus being a key activity of active aging. It is also of interest to recognize any differences among genders, generational cohorts or country of residence. If a positive relationship is found between civic participation and well-being, the actions that promote this participation will benefit the quality of life of senior citizens. Otherwise, independent action must be taken in the improvement of social and human capital. The sample consists of approximately 50.000 individuals from the European Social Survey (2002-2016). Only individuals born before 1965 in 15 European countries were considered. The sample was distributed according to gender, year of birth, country, level of studies and ESS wave to form pseudo-panel data cohorts, leaving a total of 1.318 observations. The data were analyzed through a Cross-Lagged Model using Fixed-Effects. A bidirectional association is observed between the civic participation and well-being variables. However, participating in the past seems to have a higher impact on today’s health, happiness and life satisfaction than the other way around. Furthermore, 26% of the respondents expressed to be satisfied with their life, 27% to be happy and 57% to have good health. On the other hand, 49% have participated civically in the last year, being the most common activities: signing petitions, boycotting products and volunteer work in non-political organizations. A slight trend of BabyBoomers and men towards greater participation can be observed, as well as a higher impact of this participation on their well-being. In addition, international differences exhibit a stronger relation for Nordic, East European and Mediterranean countries. The given results support the hypothesis that civic participation is a promoter of well-being for the elderly. This paper positively highlights the activity of involving in political and non-political organizations, as well as wearing badges. At any rate, almost all forms of civic participation show a positive relationship with well-being and should therefore be promoted, although differences between countries must be taken into consideration.

Keywords: active aging, civic participation, Europe, well-being

Procedia PDF Downloads 65

Authors: Mpho Mokoatle, Darlington Mapiye, James Mashiyane, Stephanie Muller, Gciniwe Dlamini

Abstract:

Great advances in high-throughput sequencing technologies have resulted in availability of huge amounts of sequencing data in public and private repositories, enabling a holistic understanding of complex biological phenomena. Sequence data are used for a wide range of applications such as gene annotations, expression studies, personalized treatment and precision medicine. However, this rapid growth in sequence data poses a great challenge which calls for novel data processing and analytic methods, as well as huge computing resources. In this work, a machine and statistical learning approach for DNA sequence classification based on $k$-mer representation of sequence data is proposed. The approach is tested using whole genome sequences of Mycobacterium tuberculosis (MTB) isolates to (i) reduce the size of genomic sequence data, (ii) identify an optimum size of k-mers and utilize it to build classification models, (iii) predict the phenotype from whole genome sequence data of a given bacterial isolate, and (iv) demonstrate computing challenges associated with the analysis of whole genome sequence data in producing interpretable and explainable insights. The classification models were trained on 104 whole genome sequences of MTB isoloates. Cluster analysis showed that k-mers maybe used to discriminate phenotypes and the discrimination becomes more concise as the size of k-mers increase. The best performing classification model had a k-mer size of 10 (longest k-mer) an accuracy, recall, precision, specificity, and Matthews Correlation coeffient of 72.0%, 80.5%, 80.5%, 63.6%, and 0.4 respectively. This study provides a comprehensive approach for resampling whole genome sequencing data, objectively selecting a k-mer size, and performing classification for phenotype prediction. The analysis also highlights the importance of increasing the k-mer size to produce more biological explainable results, which brings to the fore the interplay that exists amongst accuracy, computing resources and explainability of classification results. However, the analysis provides a new way to elucidate genetic information from genomic data, and identify phenotype relationships which are important especially in explaining complex biological mechanisms.

Keywords: AWD-LSTM, bootstrapping, k-mers, next generation sequencing

Procedia PDF Downloads 141

Authors: Darlington Mapiye, Mpho Mokoatle, James Mashiyane, Stephanie Muller, Gciniwe Dlamini

Abstract:

Great advances in high-throughput sequencing technologies have resulted in availability of huge amounts of sequencing data in public and private repositories, enabling a holistic understanding of complex biological phenomena. Sequence data are used for a wide range of applications such as gene annotations, expression studies, personalized treatment and precision medicine. However, this rapid growth in sequence data poses a great challenge which calls for novel data processing and analytic methods, as well as huge computing resources. In this work, a machine and statistical learning approach for DNA sequence classification based on k-mer representation of sequence data is proposed. The approach is tested using whole genome sequences of Mycobacterium tuberculosis (MTB) isolates to (i) reduce the size of genomic sequence data, (ii) identify an optimum size of k-mers and utilize it to build classification models, (iii) predict the phenotype from whole genome sequence data of a given bacterial isolate, and (iv) demonstrate computing challenges associated with the analysis of whole genome sequence data in producing interpretable and explainable insights. The classification models were trained on 104 whole genome sequences of MTB isoloates. Cluster analysis showed that k-mers maybe used to discriminate phenotypes and the discrimination becomes more concise as the size of k-mers increase. The best performing classification model had a k-mer size of 10 (longest k-mer) an accuracy, recall, precision, specificity, and Matthews Correlation coeffient of 72.0 %, 80.5 %, 80.5 %, 63.6 %, and 0.4 respectively. This study provides a comprehensive approach for resampling whole genome sequencing data, objectively selecting a k-mer size, and performing classification for phenotype prediction. The analysis also highlights the importance of increasing the k-mer size to produce more biological explainable results, which brings to the fore the interplay that exists amongst accuracy, computing resources and explainability of classification results. However, the analysis provides a new way to elucidate genetic information from genomic data, and identify phenotype relationships which are important especially in explaining complex biological mechanisms

Keywords: AWD-LSTM, bootstrapping, k-mers, next generation sequencing

Procedia PDF Downloads 129

Authors: Jinhoa Lee

Abstract:

The relationships between environmental quality, energy use and economic output have created growing attention over the past decades among researchers and policy makers. Focusing on the empirical aspects of the role of CO2 emissions and energy use in affecting the economic output, this paper is an effort to fulfill the gap in a comprehensive case study at a country level using modern econometric techniques. To achieve the goal, this country-specific study examines the short-run and long-run relationships among energy consumption (using disaggregated energy sources: crude oil, coal, natural gas, electricity), carbon dioxide (CO2) emissions and gross domestic product (GDP) for Turkey using time series analysis from the year 1980-2010. To investigate the relationships between the variables, this paper employs the Phillips–Perron (PP) test for stationarity, Johansen maximum likelihood method for cointegration and a Vector Error Correction Model (VECM) for both short- and long-run causality among the research variables for the sample. All the variables in this study show very strong significant effects on GDP in the country for the long term. The long-run equilibrium in the VECM suggests negative long-run causalities from consumption of petroleum products and the direct combustion of crude oil, coal and natural gas to GDP. Conversely, positive impacts of CO2 emissions and electricity consumption on GDP are found to be significant in Turkey during the period. There exists a short-run bidirectional relationship between electricity consumption and natural gas consumption. There exists a positive unidirectional causality running from electricity consumption to natural gas consumption, while there exists a negative unidirectional causality running from natural gas consumption to electricity consumption. Moreover, GDP has a negative effect on electricity consumption in Turkey in the short run. Overall, the results support arguments that there are relationships among environmental quality, energy use and economic output but the associations can to be differed by the sources of energy in the case of Turkey over of period 1980-2010.

Keywords: CO2 emissions, energy consumption, GDP, Turkey, time series analysis

Procedia PDF Downloads 493

Authors: Mao Huasong, Tang Siqi, Cheng Yu

Abstract:

The cognition and practice of Chinese landscapes have distinct uniqueness. The intergenerational inheritance of urban and rural landscapes is a common objective fact which has created a unique type of heritage in China - scenic heritage. The current generalization of the concept of scenic heritage has affected the lack of innovation in corresponding protection practices. Therefore, clarifying the concepts and connotations of scenery and scenic heritage, clarifying the protection objects of scenic heritage and the methods and approaches in intergenerational inheritance can provide theoretical support for the practice of Chinese scenic heritage and contribute Chinese wisdom to the transformation of world heritage sites. Taking ancient Shaoxing, which has a long time span and rich descriptions of scenic types and quantities, as the research object and using local chronicles as the basic research material, based on text analysis, word frequency analysis, case statistics, and historical, geographical spatial annotation methods, this study traces back to ancient scenic practices and conducts in-depth descriptions in both text and space. it have constructed a scenic heritage identification method based on the basic connotation characteristics and morphological representation characteristics of natural and cultural correlations, combined with the intergenerational and representative characteristics of scenic heritage; Summarized the bidirectional integration of "scenic spots" and "form scenic spots", "outstanding people" and "local spirits" in the formation process of scenic heritage; In inheritance, guided by Confucian values of education; In communication, the cultural interpretation constructed by scenery and the way of landscape life are used to strengthen the intergenerational inheritance of natural, artificial material elements, and intangible spirits. As a unique type of heritage in China, scenic heritage should improve its standards, values, and connotations in current protection practices and actively absorb historical experience.

Keywords: scenic heritage, heritage protection, cultural landscape, shaoxing, chinese landscape

Procedia PDF Downloads 44

Authors: Bagher Bayat, Wouter Verhoef, Behnaz Arabi, Christiaan Van der Tol

Abstract:

The aim of this study was to follow the canopy reflectance patterns in response to soil water deficit and to detect trends of changes in biophysical and biochemical parameters of grass (Poa pratensis species). We used visual interpretation, imaging spectroscopy and radiative transfer model inversion to monitor the gradual manifestation of water stress effects in a laboratory setting. Plots of 21 cm x 14.5 cm surface area with Poa pratensis plants that formed a closed canopy were subjected to water stress for 50 days. In a regular weekly schedule, canopy reflectance was measured. In addition, Leaf Area Index (LAI), Chlorophyll (a+b) content (Cab) and Leaf Water Content (Cw) were measured at regular time intervals. The 1-D bidirectional canopy reflectance model SAIL, coupled with the leaf optical properties model PROSPECT, was inverted using hyperspectral measurements by means of an iterative optimization method to retrieve vegetation biophysical and biochemical parameters. The relationships between retrieved LAI, Cab, Cw, and Cs (Senescent material) with soil moisture content were established in two separated groups; stress and non-stressed. To differentiate the water stress condition from the non-stressed condition, a threshold was defined that was based on the laboratory produced Soil Water Characteristic (SWC) curve. All parameters retrieved by model inversion using canopy spectral data showed good correlation with soil water content in the water stress condition. These parameters co-varied with soil moisture content under the stress condition (Chl: R2= 0.91, Cw: R2= 0.97, Cs: R2= 0.88 and LAI: R2=0.48) at the canopy level. To validate the results, the relationship between vegetation parameters that were measured in the laboratory and soil moisture content was established. The results were totally in agreement with the modeling outputs and confirmed the results produced by radiative transfer model inversion and spectroscopy. Since water stress changes all parts of the spectrum, we concluded that analysis of the reflectance spectrum in the VIS-NIR-MIR region is a promising tool for monitoring water stress impacts on vegetation.

Keywords: hyperspectral remote sensing, model inversion, vegetation responses, water stress

Procedia PDF Downloads 201

Authors: Jianzhou Ni, Hua Fan

Abstract:

It has been determined that supervisor support has a major influence on postgraduate students' academic emotions and is considered a method of successfully anticipating postgraduates' good psychological well-being levels. As a result, by assessing the mediating influence upon academic emotions for contemporary postgraduates in China, this study investigated the tight reciprocal relationship between psychological empowerment and positive mental well-being among postgraduates. To that end, a help enables a theoretical analysis of role clarity, academic emotion, and positive psychological health was developed, and its validity and reliability were demonstrated for the first time using the normalized postgrad relationship with supervisor scale, academic emotion scale, and positive mental scale, as well as questionnaire data from Chinese postgraduate students. This study used the cross-lagged (ARCL) panel model data to longitudinally measure 798 valid data from two survey questions polls taken in 2019 (T1) and 2021 (T2) to investigate the link between supervisor support and positive graduate student mental well-being in a bidirectional relationship of influence. The study discovered that mentor assistance could have a considerable beneficial impact on graduate students' academic emotions and, as a result, indirectly help learners attain positive mental health development. This verifies the theoretical premise that academic emotions partially mediate the effect of mentor support on positive mental health development and argues for the coexistence of the two. The outcomes of this study can help researchers gain a better knowledge of the dynamic interplay among three different research variables: supervisor support, academic emotions, and positive mental health, as well as fill gaps in previous research. In this regard, the study indicated that mentor assistance directly stimulates students' academic drive and assists graduate students in developing good academic emotions, which contributes to the development of positive mental health. However, given the restricted measurement time in this study's cross-lagged panel data and the potential effect of moderating effects other than academic mood on graduate students' good mental health, the results of this study need to be more fully understood and validated.

Keywords: supervisor support, academic emotions, positive mental health, interaction effects, longitudinal cross-lagged measurements

Procedia PDF Downloads 69

Authors: Marcio Freire Cruz, Naoaki Ono, Shigehiko Kanaya, Carlos Arthur Mattos Teixeira Cavalcante

Abstract:

The systems that record patient care information, known as Electronic Medical Record (EMR) and those that monitor vital signs of patients, such as heart rate, body temperature, and blood pressure have been extremely valuable for the effectiveness of the patient’s treatment. Several kinds of research have been using data from EMRs and vital signs of patients to predict illnesses. Among them, we highlight those that intend to predict, classify, or, at least identify patterns, of sepsis illness in patients under vital signs monitoring. Sepsis is an organic dysfunction caused by a dysregulated patient's response to an infection that affects millions of people worldwide. Early detection of sepsis is expected to provide a significant improvement in its treatment. Preceding works usually combined medical, statistical, mathematical and computational models to develop detection methods for early prediction, getting higher accuracies, and using the smallest number of variables. Among other techniques, we could find researches using survival analysis, specialist systems, machine learning and deep learning that reached great results. In our research, patients are modeled as points moving each hour in an n-dimensional space where n is the number of vital signs (variables). These points can reach a sepsis target point after some time. For now, the sepsis target point was calculated using the median of all patients’ variables on the sepsis onset. From these points, we calculate for each hour the position vector, the first derivative (velocity vector) and the second derivative (acceleration vector) of the variables to evaluate their behavior. And we construct a prediction model based on a Long Short-Term Memory (LSTM) Network, including these derivatives as explanatory variables. The accuracy of the prediction 6 hours before the time of sepsis, considering only the vital signs reached 83.24% and by including the vectors position, speed, and acceleration, we obtained 94.96%. The data are being collected from Medical Information Mart for Intensive Care (MIMIC) Database, a public database that contains vital signs, laboratory test results, observations, notes, and so on, from more than 60.000 patients.

Keywords: dynamic analysis, long short-term memory, prediction, sepsis

Procedia PDF Downloads 101

Authors: Harris Niavis, Dimitra Politaki

Abstract:

The manufacturing industry was collecting vast amounts of data for monitoring product quality thanks to the advances in the ICT sector and dedicated IoT infrastructure is deployed to track and trace the production line. However, industries have not yet managed to unleash the full potential of these data due to defective data collection methods and untrusted data storage and sharing. Blockchain is gaining increasing ground as a key technology enabler for Industry 4.0 and the smart manufacturing domain, as it enables the secure storage and exchange of data between stakeholders. On the other hand, AI techniques are more and more used to detect anomalies in batch and time-series data that enable the identification of unusual behaviors. The proposed scheme is based on smart contracts to enable automation and transparency in the data exchange, coupled with anomaly detection algorithms to enable reliable data ingestion in the system. Before sensor measurements are fed to the blockchain component and the smart contracts, the anomaly detection mechanism uniquely combines artificial intelligence models to effectively detect unusual values such as outliers and extreme deviations in data coming from them. Specifically, Autoregressive integrated moving average, Long short-term memory (LSTM) and Dense-based autoencoders, as well as Generative adversarial networks (GAN) models, are used to detect both point and collective anomalies. Towards the goal of preserving the privacy of industries' information, the smart contracts employ techniques to ensure that only anonymized pointers to the actual data are stored on the ledger while sensitive information remains off-chain. In the same spirit, blockchain technology guarantees the security of the data storage through strong cryptography as well as the integrity of the data through the decentralization of the network and the execution of the smart contracts by the majority of the blockchain network actors. The blockchain component of the Data Traceability Software is based on the Hyperledger Fabric framework, which lays the ground for the deployment of smart contracts and APIs to expose the functionality to the end-users. The results of this work demonstrate that such a system can increase the quality of the end-products and the trustworthiness of the monitoring process in the smart manufacturing domain. The proposed AI-enabled data traceability software can be employed by industries to accurately trace and verify records about quality through the entire production chain and take advantage of the multitude of monitoring records in their databases.

Keywords: blockchain, data quality, industry4.0, product quality

Procedia PDF Downloads 161

Authors: Man Fung Ho, Lap So, Jiaqi Zhang, Yuheng Zhao, Huiyang Lu, Tat Shing Choi, K. Y. Michael Wong

Abstract:

Nowadays, a tremendous amount of data is available in the transportation system, enabling the development of various machine learning approaches to make short-term latency predictions. A natural question is then the choice of relevant information to enable accurate predictions. Using traffic data collected from the Taiwan Freeway System, we consider the prediction of short-term latency of a freeway segment with a length of 17 km covering 5 measurement points, each collecting vehicle-by-vehicle data through the electronic toll collection system. The processed data include the past latencies of the freeway segment with different time lags, the traffic conditions of the individual segments (the accumulations, the traffic fluxes, the entrance and exit rates), the total accumulations, and the weekday latency profiles obtained by Gaussian process regression of past data. We arrive at several important conclusions about how data should be refined to obtain accurate predictions, which have implications for future system-wide latency predictions. (1) We find that the prediction of median latency is much more accurate and meaningful than the prediction of average latency, as the latter is plagued by outliers. This is verified by machine-learning prediction using XGBoost that yields a 35% improvement in the mean square error of the 5-minute averaged latencies. (2) We find that the median latency of the segment 15 minutes ago is a very good baseline for performance comparison, and we have evidence that further improvement is achieved by machine learning approaches such as XGBoost and Long Short-Term Memory (LSTM). (3) By analyzing the feature importance score in XGBoost and calculating the mutual information between the inputs and the latencies to be predicted, we identify a sequence of inputs ranked in importance. It confirms that the past latencies are most informative of the predicted latencies, followed by the total accumulation, whereas inputs such as the entrance and exit rates are uninformative. It also confirms that the inputs are much less informative of the average latencies than the median latencies. (4) For predicting the latencies of segments composed of two or three sub-segments, summing up the predicted latencies of each sub-segment is more accurate than the one-step prediction of the whole segment, especially with the latency prediction of the downstream sub-segments trained to anticipate latencies several minutes ahead. The duration of the anticipation time is an increasing function of the traveling time of the upstream segment. The above findings have important implications to predicting the full set of latencies among the various locations in the freeway system.

Keywords: data refinement, machine learning, mutual information, short-term latency prediction

Procedia PDF Downloads 151