Search results for: deep%20neural%20network

Authors: Noga Bregman

Abstract:

Accurate and efficient earthquake detection and phase picking are crucial for seismic hazard assessment and emergency response. This study introduces EQMamba, a deep-learning method that combines the strengths of the Earthquake Transformer and the Mamba model for simultaneous earthquake detection and phase picking. EQMamba leverages the computational efficiency of Mamba layers to process longer seismic sequences while maintaining a manageable model size. The proposed architecture integrates convolutional neural networks (CNNs), bidirectional long short-term memory (BiLSTM) networks, and Mamba blocks. The model employs an encoder composed of convolutional layers and max pooling operations, followed by residual CNN blocks for feature extraction. Mamba blocks are applied to the outputs of BiLSTM blocks, efficiently capturing long-range dependencies in seismic data. Separate decoders are used for earthquake detection, P-wave picking, and S-wave picking. We trained and evaluated EQMamba using a subset of the STEAD dataset, a comprehensive collection of labeled seismic waveforms. The model was trained using a weighted combination of binary cross-entropy loss functions for each task, with the Adam optimizer and a scheduled learning rate. Data augmentation techniques were employed to enhance the model's robustness. Performance comparisons were conducted between EQMamba and the EQTransformer over 20 epochs on this modest-sized STEAD subset. Results demonstrate that EQMamba achieves superior performance, with higher F1 scores and faster convergence compared to EQTransformer. EQMamba reached F1 scores of 0.8 by epoch 5 and maintained higher scores throughout training. The model also exhibited more stable validation performance, indicating good generalization capabilities. While both models showed lower accuracy in phase-picking tasks compared to detection, EQMamba's overall performance suggests significant potential for improving seismic data analysis. The rapid convergence and superior F1 scores of EQMamba, even on a modest-sized dataset, indicate promising scalability for larger datasets. This study contributes to the field of earthquake engineering by presenting a computationally efficient and accurate method for simultaneous earthquake detection and phase picking. Future work will focus on incorporating Mamba layers into the P and S pickers and further optimizing the architecture for seismic data specifics. The EQMamba method holds the potential for enhancing real-time earthquake monitoring systems and improving our understanding of seismic events.

Keywords: earthquake, detection, phase picking, s waves, p waves, transformer, deep learning, seismic waves

Procedia PDF Downloads 14

Authors: V. Vargas-Alejo, L. E. Montero-Moguel

Abstract:

Understanding the concept of function has been important in mathematics education for many years. In this study, the models built by a group of five business administration and accounting undergraduate students when carrying out a population growth activity are analyzed. The theoretical framework is the Models and Modeling Perspective. The results show how the students included tables, graphics, and algebraic representations in their models. Using technology was useful to interpret, describe, and predict the situation. The first model, the students built to describe the situation, was linear. After that, they modified and refined their ways of thinking; finally, they created exponential growth. Modeling the activity was useful to deep on mathematical concepts such as covariation, rate of change, and exponential function also to differentiate between linear and exponential growth.

Keywords: covariation reasoning, exponential function, modeling, representations

Procedia PDF Downloads 107

Authors: Mohsen Farshad

Abstract:

Information and force definition has been intertwined with the concept of entropy for many years. The displacement information of degrees of freedom with Brownian motions at a given temperature in space emerges as an entropic force between species. Here, we use this concept of entropy to understand the underlying physics behind the formation of attractive and repulsive forces by imagining that space is filled with free Brownian degrees of freedom. We incorporate the radius of bodies and the distance between them into entropic force relation systematically. Using this modified gravitational entropic force, we derive the attractive entropic force between bodies without considering their spin. We further hypothesize a possible mechanism for the formation of the repulsive force between two bodies. We visually elaborate that the repulsive entropic force will be manifested through the rotation of degrees of freedom around the spinning particles.

Keywords: entropy, information, force, Brownian Motions

Procedia PDF Downloads 58

Authors: Fathi Mohamed Abdrabbo, Khaled Elsayed Gaaver, Musab Musa Eldooma

Abstract:

This paper presents an attempt made to investigate the behavior of a single vertical steel hollow pile embedded in sand subjected to compressive inclined load at various inclination angles α through FEM package MIDAS GTS/NX 2019. The effect of the inclination angle and slenderness ratio on the performance of the pile was investigated. Inclined load caring capacity and pile stiffness, as well as lateral deformation profiles along with the pile, were presented. The global, vertical, and horizontal load displacements, as well as the deformation profiles along with the pile and the pile stiffness, are significantly affected by α. Whereas P-Y curves of the pile are independent of α., also the slenderness ratios are markedly affecting the behavior of the pile. In addition, there was a noticeable effect of the horizontal component on the vertical behavior of the pile, whereas there was no influence of the presence of vertical load on the horizontal behavior of the pile.

Keywords: deep foundations, piles, inclined load, pile deformations

Procedia PDF Downloads 156

Authors: Afaf Alharbi, Qianni Zhang

Abstract:

The identification of malignant tissue in histopathological slides holds significant importance in both clinical settings and pathology research. This paper introduces a methodology aimed at automatically categorizing cancerous tissue through the utilization of a multiple-instance learning framework. This framework is specifically developed to acquire knowledge of the Bernoulli distribution of the bag label probability by employing neural networks. Furthermore, we put forward a neural network based permutation-invariant aggregation operator, equivalent to attention mechanisms, which is applied to the multi-instance learning network. Through empirical evaluation of an openly available colon cancer histopathology dataset, we provide evidence that our approach surpasses various conventional deep learning methods.

Keywords: attention multiple instance learning, MIL and transfer learning, histopathological slides, cancer tissue classification

Procedia PDF Downloads 86

Authors: Marta Ribo, Ian D. Goodwin, Thomas Mortlock, Phil O’Brien

Abstract:

Coastal behavior is best investigated using a sediment budget approach, based on the identification of sediment sources and sinks. Grain size distribution over the New South Wales (NSW) continental shelf has been widely characterized since the 1970’s. Coarser sediment has generally accumulated on the outer shelf, and/or nearshore zones, with the latter related to the presence of nearshore reef and bedrocks. The central part of the NSW shelf is characterized by the presence of fine sediments distributed parallel to the coastline. This study presents new grain size distribution maps along the NSW continental shelf, built using all available NSW and Commonwealth Government holdings. All available seabed bathymetric data form prior projects, single and multibeam sonar, and aerial LiDAR surveys were integrated into a single bathymetric surface for the NSW continental shelf. Grain size information was extracted from the sediment sample data collected in more than 30 studies. The information extracted from the sediment collections varied between reports. Thus, given the inconsistency of the grain size data, a common grain size classification was her defined using the phi scale. The new sediment distribution maps produced, together with new detailed seabed bathymetric data enabled us to revise the delineation of sediment compartments to more accurately reflect the true nature of sediment movement on the inner shelf and nearshore. Accordingly, nine primary mega coastal compartments were delineated along the NSW coast and shelf. The sediment compartments are bounded by prominent nearshore headlands and reefs, and major river and estuarine inlets that act as sediment sources and/or sinks. The new sediment grain size distribution was used as an input in the morphological modelling to quantify the sediment transport patterns (and indicative rates of transport), used to investigate sand supply rates and processes from the lower shoreface to the NSW coast. The rate of sand supply to the NSW coast from deep water is a major uncertainty in projecting future coastal response to sea-level rise. Offshore transport of sand is generally expected as beaches respond to rising sea levels but an onshore supply from the lower shoreface has the potential to offset some of the impacts of sea-level rise, such as coastline recession. Sediment exchange between the lower shoreface and sub-aerial beach has been modelled across the south, central, mid-north and far-north coast of NSW. Our model approach is that high-energy storm events are the primary agents of sand transport in deep water, while non-storm conditions are responsible for re-distributing sand within the beach and surf zone.

Keywords: New South Wales coast, off-shore transport, sand supply, sediment distribution maps

Procedia PDF Downloads 216

Authors: Bingqing Chen

Abstract:

The Opium War unlocked the gate of China. Since then, modern western culture has been imported strongly and spread throughout this Asian country. The monologue of traditional Chinese culture in the past has been replaced by the hustle and bustle of multiculturalism. In the field of music, starting from school music, China, a country without the concept of composition, was deeply influenced by western culture and professional music composition, and entered the era of professional music composition. Recognizing the importance of national culture, a group of insightful artists began to try to add ‘China’ to musical composition. However, due to the special historical origin of Chinese professional musical composition and the three times of cultural nihilism in China, professional musical composition at this time failed to interpret the deep language structure of local culture within Chinese traditional culture, but only regarded Chinese traditional music as a ‘melody material library.’ At this time, the cross-cultural composition still takes Western music as its ‘norm,’ while our own music culture only exists as the sound of the contrast of Western music. However, after reading scores extensively, watching video performances, and interviewing several active composers, we found that at least in the past 30 years, China has created some works that can be called intercultural music. In these kinds of music, composers put Chinese and Western, traditional and modern in an almost equal position to have a dialogue based on their deep understanding and respect for the two cultures. This kind of music connects two music worlds, and links the two cultural and ideological worlds behind it, and communicates and grows together. This paper chose the works of three composers with different educational backgrounds, and pay attention to how composers can make a dialogue at the organizational structure level of their works. Based on the strategies adopted by composers in structuring their works, this paper expounds on how the composer's music procedure shows intercultural in terms of whole sound effects and cultural symbols. By actively participating in this intercultural practice, composers resorting to various musical and extra-musical procedures to arrive at the so-called ‘innovation within tradition.’ Through the dialogue, we can activate the space of creative thinking and explore the potential contained in culture. This interdisciplinary research promotes the rethinking of the possibility of innovation in contemporary Chinese intercultural music composition, spanning the fields of sound studies, dialogue theory, cultural research, music theory, and so on. Recently, China is calling for actively promoting 'the construction of Chinese music canonization,’ expecting to form a particular music style to show national-cultural identity. In the era of globalization, it is possible to form a brand-new Chinese music style through intercultural composition, but it is a question about talents, and the key lies in how composers do it. There is no recipe for the formation of the Chinese music style, only the composers constantly trying and tries to solve problems in their works.

Keywords: dialogism, intercultural music, national-cultural identity, organization/structure, sound

Procedia PDF Downloads 97

Authors: Tatyana Gavrilova, Vadim Onufriev

Abstract:

Knowledge-based systems’ design requires the developer’s owning the advanced analytical skills. The efficient development of that skills within university courses needs a deep understanding of main pitfalls and drawbacks, which students usually make during their analytical work in form of visual modeling. Thus, it was necessary to hold an analysis of 5-th year students’ learning exercises within courses of 'Intelligent systems' and 'Knowledge engineering' in Saint-Petersburg Polytechnic University. The analysis shows that both lack of system thinking skills and methodological mistakes in course design cause the errors that are discussed in the paper. The conclusion contains an exploration of the issues and topics necessary and sufficient for the implementation of the improved practices in educational design for future curricula of teaching programs.

Keywords: knowledge based systems, knowledge engineering, students’ errors, visual modeling

Procedia PDF Downloads 295

Authors: Jaeyoung Lee

Abstract:

Autonomous driving systems require high reliability to provide people with a safe and comfortable driving experience. However, despite the development of a number of vehicle sensors, it is difficult to always provide high perceived performance in driving environments that vary from time to season. The image segmentation method using deep learning, which has recently evolved rapidly, provides high recognition performance in various road environments stably. However, since the system controls a vehicle in real time, a highly complex deep learning network cannot be used due to time and memory constraints. Moreover, efficient networks are optimized for GPU environments, which degrade performance in embedded processor environments equipped simple hardware accelerators. In this paper, a semantic segmentation network, matrix multiplication accelerator network (MMANet), optimized for matrix multiplication accelerator (MMA) on Texas instrument digital signal processors (TI DSP) is proposed to improve the recognition performance of autonomous driving system. The proposed method is designed to maximize the number of layers that can be performed in a limited time to provide reliable driving environment information in real time. First, the number of channels in the activation map is fixed to fit the structure of MMA. By increasing the number of parallel branches, the lack of information caused by fixing the number of channels is resolved. Second, an efficient convolution is selected depending on the size of the activation. Since MMA is a fixed, it may be more efficient for normal convolution than depthwise separable convolution depending on memory access overhead. Thus, a convolution type is decided according to output stride to increase network depth. In addition, memory access time is minimized by processing operations only in L3 cache. Lastly, reliable contexts are extracted using the extended atrous spatial pyramid pooling (ASPP). The suggested method gets stable features from an extended path by increasing the kernel size and accessing consecutive data. In addition, it consists of two ASPPs to obtain high quality contexts using the restored shape without global average pooling paths since the layer uses MMA as a simple adder. To verify the proposed method, an experiment is conducted using perfsim, a timing simulator, and the Cityscapes validation sets. The proposed network can process an image with 640 x 480 resolution for 6.67 ms, so six cameras can be used to identify the surroundings of the vehicle as 20 frame per second (FPS). In addition, it achieves 73.1% mean intersection over union (mIoU) which is the highest recognition rate among embedded networks on the Cityscapes validation set.

Keywords: edge network, embedded network, MMA, matrix multiplication accelerator, semantic segmentation network

Procedia PDF Downloads 110

Authors: Mukesh Kumar, Onkar Singh, Naveen Kumar, Amar Deep

Abstract:

The waste cooking oil is taken as feedstock for biodiesel production. For this research, waste cooking oil is collected from many hotels and restaurants, and then biodiesel is prepared for experimentation purpose. The prepared biodiesel is mixed with mineral diesel in the proportion of 10%, 20%, and 30% to perform tests on a diesel engine. The experimental analysis is carried out at different load conditions to analyze the impact of the blending ratio on the performance and emission parameters. When the blending proportion of biodiesel is increased, then the highest pressure reduces due to the fall in the calorific value of the blended mixture. Experimental analysis shows a promising decrease in nitrogen oxides (NOx). A mixture of 20% biodiesel and mineral diesel is the best negotiation, mixing ratio, and beyond that, a remarkable reduction in the outcome of the performance has been observed.

Keywords: alternative sources, diesel engine, emissions, performance

Procedia PDF Downloads 159

Authors: Lali Khurtsia, Vano Tsertsvadze

Abstract:

The purpose of the study was to find out incentives and expectations, methods and ways, which are influential to students during working with their thesis. Research findings shows that the use of plagiarism has cultural links deep in the history - on the one hand, the tradition of sharing knowledge in the oral manner, with its different interpretations, and on the other hand the lack of fair and honest methods in the academic process. Research results allow us to determine general ideas about preventive policy to reduce the use of plagiarism. We conducted surveys in three different groups – we interviewed so-called diploma writers, students on bachelors and masters level and the focus group of lecturers. We found that the problem with plagiarism in Georgia has cultural-mental character. We think that nearest years’ main task should be breaking of barriers existed between lecturers and students and acknowledgement of honest principals of study process among students and pupils.

Keywords: education, Georgia, plagiarism, study process, school, university

Procedia PDF Downloads 214

Authors: Chen Xi, Lao Xuerui, Li Junjie, Jiang Yike, Wang Hanwei, Zeng Zihao

Abstract:

To further promote the development of smart cities, the microscopic "nerve endings" of the City Intelligent Model (CIM) are extended to be more sensitive. In this paper, we develop a pedestrian trajectory twin terminal based on the CIM and CNN technology. It also uses 5G networks, architectural and geoinformatics technologies, convolutional neural networks, combined with deep learning networks for human behaviour recognition models, to provide empirical data such as 'pedestrian flow data and human behavioural characteristics data', and ultimately form spatial performance evaluation criteria and spatial performance warning systems, to make the empirical data accurate and intelligent for prediction and decision making.

Keywords: urban planning, urban governance, CIM, artificial intelligence, convolutional neural network

Procedia PDF Downloads 101

Authors: Li Hui, Riyadh Hindi

Abstract:

Concrete structures frequently suffer from crack formation, a critical issue that can significantly reduce their lifespan by allowing damaging agents to enter. Traditional methods of crack detection depend on manual visual inspections, which heavily relies on the experience and expertise of inspectors using tools. In this study, a more efficient, computer vision-based approach is introduced by using the lightweight MobileNetV2 neural network. A dataset of 40,000 images was used to develop a specialized crack evaluation algorithm. The analysis indicates that MobileNetV2 matches the accuracy of traditional CNN methods but is more efficient due to its smaller size, making it well-suited for mobile device applications. The effectiveness and reliability of this new method were validated through experimental testing, highlighting its potential as an automated solution for crack detection in concrete structures.

Keywords: Concrete crack, computer vision, deep learning, MobileNetV2 neural network

Procedia PDF Downloads 46

Authors: Valentin Père, Mathieu Milhé, Fabien Baillon, Jean-Louis Dirion

Abstract:

Microgrids offer potential solutions to meet the need for local grid stability and increase isolated networks autonomy with the integration of intermittent renewable energy production and storage facilities. In such a context, sizing production and storage for a given network is a complex task, highly depending on input data such as power load profile and renewable resource availability. This work aims at developing an operating cost computation methodology for different microgrid designs based on the use of deep reinforcement learning (RL) algorithms to tackle the optimal operation problem in stochastic environments. RL is a data-based sequential decision control method based on Markov decision processes that enable the consideration of random variables for control at a chosen time scale. Agents trained via RL constitute a promising class of Energy Management Systems (EMS) for the operation of microgrids with energy storage. Microgrid sizing (or design) is generally performed by minimizing investment costs and operational costs arising from the EMS behavior. The latter might include economic aspects (power purchase, facilities aging), social aspects (load curtailment), and ecological aspects (carbon emissions). Sizing variables are related to major constraints on the optimal operation of the network by the EMS. In this work, an islanded mode microgrid is considered. Renewable generation is done with photovoltaic panels; an electrochemical battery ensures short-term electricity storage. The controllable unit is a hydrogen tank that is used as a long-term storage unit. The proposed approach focus on the transfer of agent learning for the near-optimal operating cost approximation with deep RL for each microgrid size. Like most data-based algorithms, the training step in RL leads to important computer time. The objective of this work is thus to study the potential of Batch-Constrained Q-learning (BCQ) for the optimal sizing of microgrids and especially to reduce the computation time of operating cost estimation in several microgrid configurations. BCQ is an off-line RL algorithm that is known to be data efficient and can learn better policies than on-line RL algorithms on the same buffer. The general idea is to use the learned policy of agents trained in similar environments to constitute a buffer. The latter is used to train BCQ, and thus the agent learning can be performed without update during interaction sampling. A comparison between online RL and the presented method is performed based on the score by environment and on the computation time.

Keywords: batch-constrained reinforcement learning, control, design, optimal

Procedia PDF Downloads 105

Authors: Hareem Kausar

Abstract:

Women of today’s world are energetic, enthusiastic and high-spirited. They tend to be the best in whatever they do and strive to accept and fulfil each challenge with utmost liveliness. The aim of the research was about studying the impact of Maternal Employment on the Child’s Behavioral Development. It was conducted as an initiative to study the impact factor in Pakistani culture and for deep insight to the subject using qualitative research methodology. The samples were interviewed through semi-structured interview method in three phases including two working mothers, two children and a day care center official and the data was collected and analyzed through content analysis. Further, it was linked with the literature from the west and the results show that children of working mothers tend to be sound mentally and physically but at some points they face the inner feeling of solitude. Overall, develop the mechanism in independence in their nature and behavior but maternal employment definitely affects the overall behavioral development of the children.

Keywords: maternal employment, child behavior- development, childhood, impact

Procedia PDF Downloads 535

Authors: Hammadi Larbi, Bergane Cheikh

Abstract:

Drilling for oil is done through many mechanisms. The goal is first to dig deep and then, after arriving at the oil source, to simply suck it up. And for this, it is important to know the role of oil-based drilling muds, which had many benefits for the drilling tool and for drilling generally, and also and essentially to know the rheological behavior of the emulsion system in particular water-in-oil inverse emulsions (Water/crude oil). This work contributes to the improvement of the stability and rheological properties of crude oil-based drilling mud by organophilic clay. Experimental data from steady-state flow measurements of crude oil-based drilling mud are classically analyzed by the Herschel-Bulkley model. The effects of organophilic clay type VG69 are studied. Microscopic observation showed that the addition of quantities of organophilic clay type VG69 less than or equal to 3 g leads to the stability of inverse Water/Oil emulsions; on the other hand, for quantities greater than 3g, the emulsions are destabilized.

Keywords: drilling, organophilic clay, crude oil, stability

Procedia PDF Downloads 107

Authors: Jiahe Liu, HongyangYu, Feng Qian, Miao Luo

Abstract:

This study aimed to address the problem of improving the confidence of key points by fusing multi-view information, thereby estimating human posture more accurately. We first obtained multi-view image information and then used the MvP algorithm to fuse this multi-view information together to obtain a set of high-confidence human key points. We used these as the input for the Spatio-Temporal Graph Convolution (ST-GCN). ST-GCN is a deep learning model used for processing spatio-temporal data, which can effectively capture spatio-temporal relationships in video sequences. By using the MvP algorithm to fuse multi-view information and inputting it into the spatio-temporal graph convolution model, this study provides an effective method to improve the accuracy of human posture estimation and provides strong support for further research and application in related fields.

Keywords: multi-view, pose estimation, ST-GCN, joint fusion

Procedia PDF Downloads 51

Authors: Nam Chol Yu, GyongIl Chu, HoJong Ri

Abstract:

Using DLTS (Deep-level transient spectroscopy) measurement techniques, we determined the dominant recombination center levels (defects of both A and B) in gold and platinum doped n-type silicon. Also, the injection and temperature dependence of the Shockley-Read-Hall (SRH) carrier lifetime was studied under low-level injection and high-level injection. Here measurements show that the dominant level under low-level injection located at EC-0.25 eV (A) correlated to the Pt+G1 and the dominant level under high-level injection located at EC-0.54 eV (B) correlated to the Au+G4. Finally, A and B are the same dominant levels for controlling the lifetime in gold-platinum doped n-silicon.

Keywords: recombination center level, lifetime, carrier lifetime control, Gold, Platinum, Silicon

Procedia PDF Downloads 45

Authors: Pawan Kumar, Dhananjoy Shaw, Manoj Kumar Rathi

Abstract:

Purpose of the study were to compare between (a) frequencies among the four quartiles of percentile norms of autonomic variables from power events and (b) frequencies among the four quartiles percentile norms of autonomic variables from aerobic events of Indian sportspersons withdrawn from competitive games and sports in regard to number of samples falling in each quartile. The study was conducted on 430 males of 30 to 35 years of age. Based on the nature of game/sports the retired sportspersons were classified into power events (throwers, judo players, wrestlers, short distance swimmers, cricket fast bowlers and power lifters) and aerobic events (long distance runners, long distance swimmers, water polo players). Date was collected using ECG polygraphs. Data were processed and extracted using frequency domain analysis and time domain analysis. Collected data were computed with frequency, percentage of each quartile and finally the frequencies were compared with the chi square analysis. The finding pertaining to norm reference comparison of frequencies among the four quartiles of Indian sportspersons withdrawn from competitive games and sports from (a) power events suggests that frequency distribution in four quartile namely Q1, Q2, Q3, and Q4 are significantly different at .05 level in regard to variables namely, SDNN, Total Power (Absolute Power), HF (Absolute Power), LF (Normalized Power), HF (Normalized Power), LF/HF ratio, deep breathing test, expiratory respiratory ratio, valsalva manoeuvre, hand grip test, cold pressor test and lying to standing test, whereas, insignificantly different at .05 level in regard to variables namely, SDSD, RMSSD, SDANN, NN50 Count, pNN50 Count, LF (Absolute Power) and 30: 15 Ratio (b) aerobic events suggests that frequency distribution in four quartile are significantly different at .05 level in regard to variables namely, SDNN, LF (Normalized Power), HF (Normalized Power), LF/HF ratio, deep breathing test, expiratory respiratory ratio, hand grip test, cold pressor test, lying to standing test and 30: 15 ratio, whereas, insignificantly different at .05 level in regard to variables namely, SDSD, RMSSD. SDANN, NN50 count, pNN50 count, Total Power (Absolute Power), LF(Absolute Power) HF(Absolute Power), and valsalva manoeuvre. The study concluded that comparison of frequencies among the four quartiles of Indian retired sportspersons from power events and aerobic events are different in four quartiles in regard to selected autonomic functions, hence the developed percentile norms are not homogenously distributed across the percentile scale; hence strengthen the percentage distribution towards normal distribution.

Keywords: power, aerobic, absolute power, normalized power

Procedia PDF Downloads 342

Authors: Burak Oksuzoglu, Savas Yildirim, Ferhat Kutlu

Abstract:

Opinion mining (OM) is one of the natural language processing (NLP) problems to determine the polarity of opinions, mostly represented on a positive-neutral-negative axis. The data for OM is usually collected from various social media platforms. In an era where social media has considerable control over companies’ futures, it’s worth understanding social media and taking actions accordingly. OM comes to the fore here as the scale of the discussion about companies increases, and it becomes unfeasible to gauge opinion on individual levels. Thus, the companies opt to automize this process by applying machine learning (ML) approaches to their data. For the last two decades, OM or sentiment analysis (SA) has been mainly performed by applying ML classification algorithms such as support vector machines (SVM) and Naïve Bayes to a bag of n-gram representations of textual data. With the advent of deep learning and its apparent success in NLP, traditional methods have become obsolete. Transfer learning paradigm that has been commonly used in computer vision (CV) problems started to shape NLP approaches and language models (LM) lately. This gave a sudden rise to the usage of the pretrained language model (PTM), which contains language representations that are obtained by training it on the large datasets using self-supervised learning objectives. The PTMs are further fine-tuned by a specialized downstream task dataset to produce efficient models for various NLP tasks such as OM, NER (Named-Entity Recognition), Question Answering (QA), and so forth. In this study, the traditional and modern NLP approaches have been evaluated for OM by using a sizable corpus belonging to a large private company containing about 76,000 comments in Turkish: SVM with a bag of n-grams, and two chosen pre-trained models, multilingual universal sentence encoder (MUSE) and bidirectional encoder representations from transformers (BERT). The MUSE model is a multilingual model that supports 16 languages, including Turkish, and it is based on convolutional neural networks. The BERT is a monolingual model in our case and transformers-based neural networks. It uses a masked language model and next sentence prediction tasks that allow the bidirectional training of the transformers. During the training phase of the architecture, pre-processing operations such as morphological parsing, stemming, and spelling correction was not used since the experiments showed that their contribution to the model performance was found insignificant even though Turkish is a highly agglutinative and inflective language. The results show that usage of deep learning methods with pre-trained models and fine-tuning achieve about 11% improvement over SVM for OM. The BERT model achieved around 94% prediction accuracy while the MUSE model achieved around 88% and SVM did around 83%. The MUSE multilingual model shows better results than SVM, but it still performs worse than the monolingual BERT model.

Keywords: BERT, MUSE, opinion mining, pretrained language model, SVM, Turkish

Procedia PDF Downloads 122

Authors: Stephen L. Green, Alexander N. Gorban, Ivan Y. Tyukin

Abstract:

Within the past decade, using Convolutional Neural Networks (CNN)’s to create Deep Learning systems capable of translating Sign Language into text has been a breakthrough in breaking the communication barrier for deaf-mute people. Conventional research on this subject has been concerned with training the network to recognize the fingerspelling gestures of a given language and produce their corresponding alphanumerics. One of the problems with the current developing technology is that images are scarce, with little variations in the gestures being presented to the recognition program, often skewed towards single skin tones and hand sizes that makes a percentage of the population’s fingerspelling harder to detect. Along with this, current gesture detection programs are only trained on one finger spelling language despite there being one hundred and forty-two known variants so far. All of this presents a limitation for traditional exploitation for the state of current technologies such as CNN’s, due to their large number of required parameters. This work aims to present a technology that aims to resolve this issue by combining a pretrained legacy AI system for a generic object recognition task with a corrector method to uptrain the legacy network. This is a computationally efficient procedure that does not require large volumes of data even when covering a broad range of sign languages such as American Sign Language, British Sign Language and Chinese Sign Language (Pinyin). Implementing recent results on method concentration, namely the stochastic separation theorem, an AI system is supposed as an operate mapping an input present in the set of images u ∈ U to an output that exists in a set of predicted class labels q ∈ Q of the alphanumeric that q represents and the language it comes from. These inputs and outputs, along with the interval variables z ∈ Z represent the system’s current state which implies a mapping that assigns an element x ∈ ℝⁿ to the triple (u, z, q). As all xi are i.i.d vectors drawn from a product mean distribution, over a period of time the AI generates a large set of measurements xi called S that are grouped into two categories: the correct predictions M and the incorrect predictions Y. Once the network has made its predictions, a corrector can then be applied through centering S and Y by subtracting their means. The data is then regularized by applying the Kaiser rule to the resulting eigenmatrix and then whitened before being split into pairwise, positively correlated clusters. Each of these clusters produces a unique hyperplane and if any element x falls outside the region bounded by these lines then it is reported as an error. As a result of this methodology, a self-correcting recognition process is created that can identify fingerspelling from a variety of sign language and successfully identify the corresponding alphanumeric and what language the gesture originates from which no other neural network has been able to replicate.

Keywords: convolutional neural networks, deep learning, shallow correctors, sign language

Procedia PDF Downloads 85

Authors: Saeed Delara, Kendra MacKay

Abstract:

Varying methodologies are considered for pile design for both Russian and Western approaches. Although both approaches rely on toe and side frictional resistances, different calculation methods are proposed to estimate pile capacity. The Western approach relies on compactness (internal friction angle) of soil for cohesionless soils and undrained shear strength for cohesive soils. The Russian approach relies on grain size for cohesionless soils and liquidity index for cohesive soils. Though most recommended methods in the Western approaches are relatively simple methods to predict pile settlement, the Russian approach provides a detailed method to estimate single pile and pile group settlement. Details to calculate pile axial capacity and settlement using the Russian and Western approaches are discussed and compared against field test results.

Keywords: pile capacity, pile settlement, Russian approach, western approach

Procedia PDF Downloads 151

Authors: Chen Xi, Liu Xuebing, Lao Xueru, Kuan Sinman, Jiang Yike, Wang Hanwei, Yang Xiaolang, Zhou Junjie, Xie Jinpeng

Abstract:

To further promote the development of smart cities, the microscopic "nerve endings" of the City Intelligent Model (CIM) are extended to be more sensitive. In this paper, we develop a pedestrian trajectory twin terminal based on the CIM and CNN technology. It also uses 5G networks, architectural and geoinformatics technologies, convolutional neural networks, combined with deep learning networks for human behavior recognition models, to provide empirical data such as 'pedestrian flow data and human behavioral characteristics data', and ultimately form spatial performance evaluation criteria and spatial performance warning systems, to make the empirical data accurate and intelligent for prediction and decision making.

Keywords: urban planning, urban governance, CIM, artificial intelligence, sustainable development

Procedia PDF Downloads 378

Authors: Thayalan Nall, Harry Poulos

Abstract:

Piling is the most widely used foundation method for heavy structures in poor soil conditions. The geotechnical engineer can choose among a variety of piling methods, but in most cases, driving piles by impact hammer is the most cost-effective alternative. Under unfavourable conditions, driving piles can cause environmental problems, such as noise, ground movements and vibrations, with the risk of ground disturbance leading to potential damage to proposed structures. In one of the project sites in which the authors were involved, three offshore container terminals, namely CT1, CT2 and CT3, were constructed over thick compressible marine mud. The seabed was around 6m deep and the soft clay thickness within the project site varied between 9m and 20m. CT2 and CT3 were connected together and rectangular in shape and were 2600mx800m in size. CT1 was 400m x 800m in size and was located on south opposite of CT2 towards its eastern end. CT1 was constructed first and due to time and environmental limitations, it was supported on a “forest” of large diameter driven piles. CT2 and CT3 are now under construction and are being carried out using a traditional dredging and reclamation approach with ground improvement by surcharging with vertical drains. A few months after the installation of the CT1 piles, a 2600m long sand bund to 2m above mean sea level was constructed along the southern perimeter of CT2 and CT3 to contain the dredged mud that was expected to be pumped. The sand bund was constructed by sand spraying and pumping using a dredging vessel. About 2000m length of the sand bund in the west section was constructed without any major stability issues or any noticeable distress. However, as the sand bund approached the section parallel to CT1, it underwent a series of deep seated failures leading the displaced soft clay materials to heave above the standing water level. The crest of the sand bund was about 100m away from the last row of piles. There were no plausible geological reasons to conclude that the marine mud only across the CT1 region was weaker than over the rest of the site. Hence it was suspected that the pile driving by impact hammer may have caused ground movements and vibrations, leading to generation of excess pore pressures and cyclic softening of the marine mud. This paper investigates the probable cause of failure by reviewing: (1) All ground investigation data within the region; (2) Soil displacement caused by pile driving, using theories similar to spherical cavity expansion; (3) Transfer of stresses and vibrations through the entire system, including vibrations transmitted from the hammer to the pile, and the dynamic properties of the soil; and (4) Generation of excess pore pressure due to ground vibration and resulting cyclic softening. The evidence suggests that the problems encountered at the site were primarily caused by the “side effects” of the pile driving operations.

Keywords: pile driving, ground vibration, excess pore pressure, cyclic softening

Procedia PDF Downloads 215

Authors: Eng Lye, Ooi

Abstract:

The Old Man and The Sea is a novella written by Ernest Hemingway that depicts an old fisherman’ journey out into the deep sea in his pursuit to catch a big fish. Through this novella, Hemingway creates a world for his protagonist, Santiago who is portrayed as an old man who has gone eighty-four days without catching a fish, at last hooks an eighteen-foot marlin, the largest he ever known. The old man endures pain and struggles to bring back to shore. Looking through the lens of gerotranscendence, we can see that the old man has his dreams, and goals in life. In his pursuit for happiness, he has fought tirelessly to ward off the shark attacks and finally he won even though only half of his fish is left. Hemingway has portrayed Santiago as an old man as a transcendent self leaping from the dimension of “The Self” to the cosmic dimension with the personal and social relationship dimension in tow. The Old Man and The Sea offers a glimpse of the struggles of an old man, who is old and gaunt but spiritually undefeated in his battle out in the sea. He is surprisingly strong and powerful despite his old age, he respects the sea, the birds. the turtles, the sharks and the fish. He can endure suffering and is focussed on achieving his goals. This is what Hemingway has portrayed Santiago to be a gerotranscendent in the eyes of the gerotranscendental approach in respect of the changes and development as seen in Santiago, the protagonist in this novella.

Keywords: gerotranscendence, gerotranscendenatal, old man, the sea, hemingway

Procedia PDF Downloads 16

Authors: Jae Soo Do, Kyoung-Seok Kim

Abstract:

Radical changes in the industrial environment, and spectacular developments of IT have changed the current of managements from people-centered to technology- or IT-centered. Interpersonal emotion exchanges have long become insipid and interactive services have also come as mechanical reactions. This study offers various concepts for the emotional labor based on traditional studies on emotional labor. Especially the present day, on which human emotions are subject to being served as machinized thing, is the time when the study on human emotions comes momentous. Precedent researches on emotional labors commonly and basically dealt with the relationship between the active group who performs actions and the passive group who is done with the action. This study focuses on the passive group and tries to offer a new perspective of 'liquid emotion' as a defence mechanism for the passive group from the external environment. Especially, this addresses a concrete discussion on directions of following studies on the liquid labor as a newly suggested perspective.

Keywords: emotion labor, surface acting, deep acting, liquid emotion

Procedia PDF Downloads 325

Authors: Asia Majeed

Abstract:

The learning of mathematics in science, engineering and social science programs can be enhanced through practical problem-solving techniques. The instructors can design their lessons with some strategies to improve students’ educational needs and accomplishments in mathematics classrooms. The use of technology in class problem solving and application sessions can enhance deep understanding of mathematics among students. As mathematician, we believe in subject specific and content-driven teaching methods. Through technology the relationship between the physical problems and the mathematical models can be analyzed. This paper is about selective use of technology in mathematics classrooms and helpful to others mathematics instructors who wishes to improve their traditional teaching techniques to improve students’ attitude in learning mathematics. These techniques corpus can be used in teaching large mathematics classes in science, technology, engineering, and social science.

Keywords: attitude in learning mathematics, mathematics, non-mathematics undergraduate students, technology

Procedia PDF Downloads 195

Authors: Ulas Vural, M. Ergun Okay, E. Mesut Yildiz

Abstract:

Telecommunication service providers demand accurate and precise prediction of customer churn probabilities to increase the effectiveness of their customer relation services. The large amount of customer data owned by the service providers is suitable for analysis by machine learning methods. In this study, expenditure data of customers are analyzed by using an artificial neural network (ANN). The ANN model is applied to the data of customers with different billing duration. The proposed model successfully predicts the churn probabilities at 83% accuracy for only three months expenditure data and the prediction accuracy increases up to 89% when the nine month data is used. The experiments also show that the accuracy of ANN model increases on an extended feature set with information of the changes on the bill amounts.

Keywords: customer relationship management, churn prediction, telecom industry, deep learning, artificial neural networks

Procedia PDF Downloads 128

Authors: Safwat E. Musa, Nuha E. Mohamed, Nuha A. Bagi

Abstract:

In this study, several crossplots of the P-impedance with the lithology logs (gamma ray, neutron porosity, deep resistivity, water saturation and Vp/Vs curves) were made in three available wells, which were drilled in central part of the Blue Nile basin in depths varies from 1460 m to 1600 m. These crossplots were successful to discriminate between sand and shale when using P-Impedance values, and between the wet sand and the pay sand when using both P-impedance and Vp/Vs together. Also, some impedance sections were converted to porosity sections using linear formula to characterize the reservoir in terms of porosity. The used crossplots were created on log resolution, while the seismic resolution can identify only the reservoir, unless a 3D seismic angle stacks were available; then it would be easier to identify the pay sand with great confidence; through high resolution seismic inversion and geostatistical approach when using P-impedance and Vp/Vs volumes.

Keywords: basin, Blue Nile, inversion, seismic

Procedia PDF Downloads 414

Authors: Surita Maini

Abstract:

There are many perceived advantages of microwave ablation have driven researchers to develop innovative antennas to effectively treat deep-seated, non-resectable hepatic tumors. In this paper a coaxial antenna with a miniaturized sleeve choke has been discussed for microwave interstitial ablation therapy, in order to reduce backward heating effects irrespective of the insertion depth into the tissue. Two dimensional Finite Element Method (FEM) is used to simulate and measure the results of miniaturized sleeve choke antenna. This paper emphasizes the importance of factors that can affect simulation accuracy, which include mesh resolution, surface heating and reflection coefficient. Quarter wavelength choke effectiveness has been discussed by comparing it with the unchoked antenna with same dimensions.

Keywords: microwave ablation, tumor, finite element method, coaxial slot antenna, coaxial dipole antenna

Procedia PDF Downloads 343