Search results for: residual LSTM

Authors: Muhammad Zubair Khan, Yugyung Lee

Abstract:

Deep learning has recently achieved enormous response in semantic image segmentation. The previously developed U-Net inspired architectures operate with continuous stride and pooling operations, leading to spatial data loss. Also, the methods lack establishing long-term pixels connection to preserve context knowledge and reduce spatial loss in prediction. This article developed encoder-decoder architecture with bi-directional LSTM embedded in long skip-connections and densely connected convolution blocks. The network non-linearly combines the feature maps across encoder-decoder paths for finding dependency and correlation between image pixels. Additionally, the densely connected convolutional blocks are kept in the final encoding layer to reuse features and prevent redundant data sharing. The method applied batch-normalization for reducing internal covariate shift in data distributions. The empirical evidence shows a promising response to our method compared with other semantic segmentation techniques.

Keywords: deep learning, semantic segmentation, image analysis, pixels connection, convolution neural network

Procedia PDF Downloads 79

Authors: Parisa Golbayani, Dan Wang, Ionut¸ Florescu

Abstract:

In this work we implement machine learning techniques to financial statement reports in order to asses company’s credit rating. Specifically, the work analyzes the performance of four neural network architectures (MLP, CNN, CNN2D, LSTM) in predicting corporate credit rating as issued by Standard and Poor’s. The paper focuses on companies from the energy, financial, and healthcare sectors in the US. The goal of this analysis is to improve application of machine learning algorithms to credit assessment. To accomplish this, the study investigates three questions. First, we investigate if the algorithms perform better when using a selected subset of important features or whether better performance is obtained by allowing the algorithms to select features themselves. Second, we address the temporal aspect inherent in financial data and study whether it is important for the results obtained by a machine learning algorithm. Third, we aim to answer if one of the four particular neural network architectures considered consistently outperforms the others, and if so under which conditions. This work frames the problem as several case studies to answer these questions and analyze the results using ANOVA and multiple comparison testing procedures.

Keywords: convolutional neural network, long short term memory, multilayer perceptron, credit rating

Procedia PDF Downloads 209

Authors: Faisal Guru Rashid, Syed Nisar, Mohammad Hussain Mir, Ulfat Ara, Richa Tripathi

Abstract:

Background: Biosimilar pegylated L-asparaginase is a potential alternative to the innovator version for treating acute lymphoblastic leukaemia (ALL) in Indian children, addressing issues of availability and cost. Biosimilar offers a viable solution, ensuring wider access to essential treatment in resource-limited settings like India. With this in mind, we conducted a study to assess the efficacy and toxicity of Biosimilar Pegaspargase (Asviia) in patients with Acute Leukaemia at our centre. Materials and methods: A retrospective study was conducted to assess the efficacy and safety of biosimilar PEG-asparaginase (Asviia) in newly diagnosed paediatric acute lymphoblastic leukaemia patients at the Paediatric Oncology unit of Department of Medical Oncology at Sher-I-Kashmir Institute of Medical Sciences, SKIMS Srinagar. The study included patients of ALL treated at our centre between January 2021- and December 2023. Each patient received 2 induction doses of pegaspargenase. Results: 45 patients (16 females and 29 males) were included in the study who received biosimilar PEG-asparaginase (Asviia) as a part of the treatment protocol. The age range of patients was between 1 and 16 years with a median age was 7.5 years. Median PEG Asparaginase dose received was 1175 IU (1125-3750 IU). The majority of patients were Pre-B ALL. There was considerable improvement in the haematological parameters, like haemoglobin levels rising by 1.39 and platelet counts rising by 30,402 after the patients received the first dose of Peg-ASP. Biosimilar Pegaspargase in Acute Leukaemia patients showed a tolerable safety profile with no life-threatening events. 13% of patients exhibited allergic reactions, and 17% had sepsis. Two patients (4.4%) had pancreatitis and Transaminitis events. At the end of induction, out of 45 patients, 40 (88.89%) patients had complete remission with Minimal Residual Disease (MRD) negativity, while 5 patients were MRD positive. Conclusion: Biosimilar PEG-Asparaginase (Asviia) demonstrated a tolerable safety profile and good efficacy, with nearly 90% of patients having complete Remission with MRD negativity.

Keywords: acute lymphoblastic leukaemia, biosimilar, PEG-asparaginase, minimal residual disease, remission

Procedia PDF Downloads 23

Authors: Jacqueline Rose T. Alipo-on, Francesca Isabelle F. Escobar, Myles Joshua T. Tan, Hezerul Abdul Karim, Nouar Al Dahoul

Abstract:

Electrocardiogram (ECG) signal analysis and processing are crucial in the diagnosis of cardiovascular diseases, which are considered one of the leading causes of mortality worldwide. However, the traditional rule-based analysis of large volumes of ECG data is time-consuming, labor-intensive, and prone to human errors. With the advancement of the programming paradigm, algorithms such as machine learning have been increasingly used to perform an analysis of ECG signals. In this paper, various deep learning algorithms were adapted to classify five classes of heartbeat types. The dataset used in this work is the synthetic MIT-BIH Arrhythmia dataset produced from generative adversarial networks (GANs). Various deep learning models such as ResNet-50 convolutional neural network (CNN), 1-D CNN, and long short-term memory (LSTM) were evaluated and compared. ResNet-50 was found to outperform other models in terms of recall and F1 score using a five-fold average score of 98.88% and 98.87%, respectively. 1-D CNN, on the other hand, was found to have the highest average precision of 98.93%.

Keywords: heartbeat classification, convolutional neural network, electrocardiogram signals, generative adversarial networks, long short-term memory, ResNet-50

Procedia PDF Downloads 93

Authors: Christina Adly, Meena Abdelmeseeh, Tamer Basha

Abstract:

This paper proposes a system for hand movement recognition using multichannel surface EMG(sEMG) signals obtained from 40 subjects using 40 different exercises, which are available on the Ninapro(Non-Invasive Adaptive Prosthetics) database. First, we applied processing methods to the raw sEMG signals to convert them to their amplitudes. Second, we used deep learning methods to solve our problem by passing the preprocessed signals to Fully connected neural networks(FCNN) and recurrent neural networks(RNN) with Long Short Term Memory(LSTM). Using intrasubject evaluation, The accuracy using the FCNN is 72%, with a processing time for training around 76 minutes, and for RNN's accuracy is 79.9%, with 8 minutes and 22 seconds processing time. Third, we applied some postprocessing methods to improve the accuracy, like majority voting(MV) and Movement Error Rate(MER). The accuracy after applying MV is 75% and 86% for FCNN and RNN, respectively. The MER value has an inverse relationship with the prediction delay while varying the window length for measuring the MV. The different part uses the RNN with the intersubject evaluation. The experimental results showed that to get a good accuracy for testing with reasonable processing time, we should use around 20 subjects.

Keywords: hand movement recognition, recurrent neural network, movement error rate, intrasubject evaluation, intersubject evaluation

Procedia PDF Downloads 111

Authors: Markus Remm, Sebastian Dienert

Abstract:

Vehicle interiors are not only changing in terms of design and functionality but also due to new driving situations in which, for example, autonomous operating modes are possible. Flexible seating positions are changing the requirements for passive safety system behavior and location in the interior of a vehicle. With fully autonomous driving, the driver can, for example, leave the position behind the steering wheel and take a seated position facing backward. Since autonomous and non-autonomous vehicles will share the same road network for the foreseeable future, accidents cannot be avoided, which makes the use of passive safety systems indispensable. With JENOPTIK-VOTAN® A technology, the trend towards flexible predetermined airbag weakening lines is enabled. With the help of laser beams, the predetermined weakening lines are introduced from the backside of the components so that they are absolutely invisible. This machining process is sensor-controlled and guarantees that a small residual wall thickness remains for the best quality and reliability for airbag weakening lines. Due to the wide processing range of the laser, the processing of almost all materials is possible. A CO₂ laser is used for many plastics, natural fiber materials, foams, foils and material composites. A femtosecond laser is used for natural materials and textiles that are very heat-sensitive. This laser type has extremely short laser pulses with very high energy densities. Supported by a high-precision and fast movement of the laser beam by a laser scanner system, the so-called cold ablation is enabled to predetermine weakening lines layer by layer until the desired residual wall thickness remains. In that way, for example, genuine leather can be processed in a material-friendly and process-reliable manner without design implications to the components A-Side. Passive safety in the vehicle is increased through the interaction of modern airbag technology and high-precision laser airbag weakening. The JENOPTIK-VOTAN® A product family has been representing this for more than 25 years and is pointing the way to the future with new and innovative technologies.

Keywords: design freedom, interior material processing, laser technology, passive safety

Procedia PDF Downloads 86

Authors: K. O. Ogbedeh, T. T. Epidi, E. U. Onweremadu, E. E. Ihem

Abstract:

Apart from the devastating nature of termites as pest of cassava, nearly all termite species have been implicated in soil fertility modifications. Elasticity of soil fertility indicators and pH in termite infested cassava field as influenced by tillage and organic manure sources in Owerri, Southeast, Nigeria was investigated in this study. Three years of of field trials were conducted in 2007, 2008 and 2009 cropping seasons respectively at the Teaching and Research Farm of the Federal University of Technology, Owerri. The experiments were laid out in a 3x6 split-plot factorial arrangement fitted into a randomized complete block design (RCBD) with three replications. The TMS 4 (2)1425 was the cassava cultivar used. Treatments consists three tillage methods (zero, flat and mound), two rates of municipal waste (1.5 and 3.0tonnes/ha), two rates of Azadirachta indica (neem) leaves (20 and 30tonnes/ha), control (0.0 tonnes/ha) and a unit dose of carbofuran (chemical check). Data were collected on pre-planting soil physical and chemical properties, post-harvest soil pH (both in water and KCl) and residual total exchangeable bases (Ca, K, Mg and Na). These were analyzed using a Mixed-model procedure of Statistical Analysis Software (SAS). Means were separated using Least Significant Difference (LSD.) at 5% level of probability. Result shows that the native soil fertility status of the experimental site was poor. However soil pH increased substantially in plots where mounds, A.indica leaves at 30t/ha and municipal waste (1.5 and 3.0t/ha) were treated especially in 2008 and 2009. In 2007 trial, highest soil pH was maintained with flat (5.41 in water and 4.97 in KCl). Control on the other hand, recorded least soil pH especially in 2009 with values of 5.18 and 4.63 in water and KCl respectively. Equally, mound, A. indica leaves at 30t/ha and municipal waste at 3.0t/ha consistently increased organic matter content of the soil than other treatments. Finally, mound and A. indica leaves at 30t/ha linearly and consistently increased residual total exchangeable bases of the soil.

Keywords: elasticity, fertility, indicators, termites, tillage, cassava and manure sources

Procedia PDF Downloads 275

Authors: Haolun Sun, Ping Wang, Mei Wu, Meng Zhang, Bin Hou, Ling Yang, Xiaohua Ma, Yue Hao

Abstract:

Gallium nitride (GaN) is an attractive material for next-generation power devices. It is noted that the performance of GaN-based high electron mobility transistors (HEMTs) is always limited by the self-heating effect. In response to the problem, integrating devices with polycrystalline diamond (PCD) has been demonstrated to be an efficient way to alleviate the self-heating issue of the GaN-based HEMTs. Among all the heat-spreading schemes, using PCD to cap the epitaxial layer before the HEMTs process is one of the most effective schemes. Now, the mainstream method of fabricating the PCD-capped HEMTs is to deposit the diamond heat-spreading layer on the AlGaN surface, which is covered by a thin nucleation dielectric/passivation layer. To achieve the pattern etching of the diamond heat spreader and device preparation, we selected SiN as the hard mask for diamond etching, which was deposited by plasma-enhanced chemical vapor deposition (PECVD). The conventional diamond etching method first uses F-based etching to remove the SiN from the special window region, followed by using O₂/Ar plasma to etch the diamond. However, the results of the scanning electron microscope (SEM) and focused ion beam microscopy (FIB) show that there are lots of diamond pillars on the etched diamond surface. Through our study, we found that it was caused by the high roughness of the diamond surface and the existence of the overlap between the diamond grains, which makes the etching of the SiN hard mask insufficient and leaves micro-masks on the diamond surface. Thus, a cyclic etching method was proposed to solve the problem of the residual SiN, which was left in the F-based etching. We used F-based etching during the first step to remove the SiN hard mask in the specific region; then, the O₂/Ar plasma was introduced to etch the diamond in the corresponding region. These two etching steps were set as one cycle. After the first cycle, we further used cyclic etching to clear the pillars, in which the F-based etching was used to remove the residual SiN, and then the O₂/Ar plasma was used to etch the diamond. Whether to take the next cyclic etching depends on whether there are still SiN micro-masks left. By using this method, we eventually achieved the self-terminated etching of the diamond and the smooth surface after the etching. These results demonstrate that the cyclic etching method can be successfully applied to the integrated preparation of polycrystalline diamond thin films and GaN HEMTs.

Keywords: AlGaN/GaN heterojunction, O₂/Ar plasma, cyclic etching, polycrystalline diamond

Procedia PDF Downloads 87

Authors: Lan Zhang, Cailing Zhang

Abstract:

In recent years, many face anti-spoofing algorithms have high detection accuracy when detecting 2D face anti-spoofing or 3D mask face anti-spoofing alone in the field of face anti-spoofing, but their detection performance is greatly reduced in multidimensional and cross-datasets tests. The rPPG method used for face anti-spoofing uses the unique vital information of real face to judge real faces and face anti-spoofing, so rPPG method has strong stability compared with other methods, but its detection rate of 2D face anti-spoofing needs to be improved. Therefore, in this paper, we improve an rPPG(Remote Photoplethysmography) method(MrPPG) for face anti-spoofing which through color space fusion, using the correlation of pulse signals between real face regions and background regions, and introducing the cyclic neural network (LSTM) method to improve accuracy in 2D face anti-spoofing. Meanwhile, the MrPPG also has high accuracy and good stability in face anti-spoofing of multi-dimensional and cross-data datasets. The improved method was validated on Replay-Attack, CASIA-FASD, Siw and HKBU_MARs_V2 datasets, the experimental results show that the performance and stability of the improved algorithm proposed in this paper is superior to many advanced algorithms.

Keywords: face anti-spoofing, face presentation attack detection, remote photoplethysmography, MrPPG

Procedia PDF Downloads 149

Authors: Mehrzad Mohabbi Yadollahi, Pouneh Abdollahifard, Behzad Mokhtare, Majid Atashafrazeh

Abstract:

Geopolymers are believed to provide good fire resistance. The effects of elevated temperatures on mechanical and microstructural properties of pumice-based geopolymer were investigated in this study. Pumice based geopolymer was exposed to elevated temperatures of 200, 400, 600, and 800 ºC for 3 hours. The residual strength of these specimens was determined after cooling at room temperature and microstructures of these samples were investigated by FTIR and SEM analyses. Specimens which were initially grey turned reddish accompanied by the appearance of cracks as temperatures increased to 600 and 800 ºC.

Keywords: geopolymer, pumice, elevated temperature, SEM, FTIR

Procedia PDF Downloads 417

Authors: Surajit Chakrabarty, Piyush Chauhan, Subhasis Panda, Sujoy Bhattacharya

Abstract:

A large proportion of roads in India are not well maintained as per the laid down public safety guidelines leading to loss of direction control and fatal accidents. We propose a technique to detect speed breakers and potholes using mobile sensor data captured from multiple vehicles and provide a profile of the road. This would, in turn, help in monitoring roads and revolutionize digital maps. Incorporating randomness in the model formulation for detection of speed breakers and potholes is crucial due to substantial heterogeneity observed in data obtained using a mobile application from multiple vehicles driven by different drivers. This is accomplished with Hidden Markov Models, whose hidden state sequence is found for each time step given the observables sequence, and are then fed as input to LSTM network with peephole connections. A precision score of 0.96 and 0.63 is obtained for classifying bumps and potholes, respectively, a significant improvement from the machine learning based models. Further visualization of bumps/potholes is done by converting time series to images using Markov Transition Fields where a significant demarcation among bump/potholes is observed.

Keywords: deep learning, hidden Markov model, pothole, speed breaker

Procedia PDF Downloads 111

Authors: Shigeyuki Haruyama, Oke Oktavianty, Zefry Darmawan, Tadayuki Kyoutani, Ken Kaminishi

Abstract:

Cab’s frame strength is considered as an important factor in excavator’s operator safety, especially during roll-over. In this study, we use a model of cab frame with different thicknesses and perform elastoplastic numerical analysis by using Finite Element Method (FEM). Deformation mode and energy absorption's of cab’s frame part are investigated on two conditions, with wrinkle and without wrinkle. The occurrence of wrinkle when deforming cab frame can reduce energy absorption, and among 4 parts with wrinkle, the energy absorption significantly decreases in part C. Residual stress that generated upon the bending process of part C is analyzed to confirm it possibility in increasing the energy absorption.

Keywords: ROPS, FEM, hydraulic excavator, cab frame

Procedia PDF Downloads 408

Authors: I. Dangeard, S. Meineri, M. Dupré

Abstract:

large body of research on energy consumption reveals that regular information on energy consumption produces a positive effect on behavior. The present research aims to test this feedback paradigm on waste management. A small-scale experiment on residual household waste was performed in a large french urban area, in partnership with local authorities, as part of the development of larger-scale project. A two-step door-to-door recruitment scheme led to 85 households answering a questionnaire. Among them, 54 accepted to participate in a study on waste (second step). Participants were then randomly assigned to one of the 3 experimental conditions : self-reported feedback on curbside waste, external feedback on waste weight based on information technologies, and no feedback for the control group. An additional control group was added, including households who were not requested to answer the questionnaire. Household residual waste was collected every week, and tags on curbside bins fed a database with waste weight of households. The feedback period lasted 14 weeks (february-may 2014). Quantitative data on waste weight were analysed, including these 14 weeks and the 7 previous weeks. Households were then contacted by phone in order to confirm the quantitative results. Regarding the recruitment questionnaire, results revealed high pro-environmental attitude on the NEP scale, high recycling behavior level and moderate level of source reduction behavior on the adapted 3R scale, but no statistical difference between the 3 experimental groups. Regarding the feedback manipulation paradigm, waste weight reveals important differences between households, but doesn't prove any statistical difference between the experimental conditions. Qualitative phone interviews confirm that recycling is a current practice among participants, whereas source reduction of waste is not, and mainly appears as a producer problem of packaging limitation. We conclude that triggering waste prevention behaviors among recycling households involves long-term feedback and should promote benchmarking, in order to clearly set waste reduction as an objective to be managed through feedback figures.

Keywords: eco-feedback, household waste, waste reduction, experimental research

Procedia PDF Downloads 370

Authors: Jesús M. Martín, Francisco Corona, Dolores Hidalgo

Abstract:

Landfills present long-term threats to soil, air, groundwater and surface water due to the formation of greenhouse gases (methane gas and carbon dioxide) and leachate from decomposing garbage. The composition of leachate differs from site to site and also within the landfill. The leachates alter with time (from weeks to years) since the landfilled waste is biologically highly active and their composition varies. Mainly, the composition of the leachate depends on factors such as characteristics of the waste, the moisture content, climatic conditions, degree of compaction and the age of the landfill. Therefore, the leachate composition cannot be generalized and the traditional treatment models should be adapted in each case. Although leachate composition is highly variable, what different leachates have in common is hazardous constituents and their potential eco-toxicological effects on human health and on terrestrial ecosystems. Since leachate has distinct compositions, each landfill or dumping site would represent a different type of risk on its environment. Nevertheless, leachates consist always of high organic concentration, conductivity, heavy metals and ammonia nitrogen. Leachate could affect the current and future quality of water bodies due to uncontrolled infiltrations. Therefore, control and treatment of leachate is one of the biggest issues in urban solid waste treatment plants and landfills design and management. This work presents a treatment model that will be carried out "in-situ" using a cost-effective novel technology that combines solar evaporation/condensation plus forward osmosis. The plant is powered by renewable energies (solar energy, biomass and residual heat), which will minimize the carbon footprint of the process. The final effluent quality is very high, allowing reuse (preferred) or discharge into watercourses. In the particular case of this work, the final effluents will be reused for cleaning and gardening purposes. A minority semi-solid residual stream is also generated in the process. Due to its special composition (rich in metals and inorganic elements), this stream will be valorized in ceramic industries to improve the final products characteristics.

Keywords: forward osmosis, landfills, leachate valorization, solar evaporation

Procedia PDF Downloads 178

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

Abstract:

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

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

Procedia PDF Downloads 14

Authors: Khwaja Naweed Seddiqi, Shigeo Honma

Abstract:

Relative permeability is the important parameter controlling the immiscible displacement of multiphase fluids flow in porous medium. The relative permeability for immiscible displacement of two-phase fluids flow (oil and water) in porous medium has been measured in this paper. As a result of the experiment, irreducible water saturation, Swi, residual oil saturation, Sor, and relative permeability curves for Kerosene, Heavy oil and Lubricant oil were determined successfully.

Keywords: relative permeability, two-phase flow, immiscible displacement, porous medium

Procedia PDF Downloads 282

Authors: Tesfaye Alamirew Dessie

Abstract:

Our study focuses on developing a Model Predictive Controller (MPC) and evaluating it against a traditional PID for a pasteurization process. Utilizing system identification from the experimental data, the dynamics of the pasteurization process were calculated. Using best fit with data validation, residual, and stability analysis, the quality of several model architectures was evaluated. The validation data fit the auto-regressive with exogenous input (ARX322) model of the pasteurization process by roughly 80.37 percent. The ARX322 model structure was used to create MPC and PID control techniques. After comparing controller performance based on settling time, overshoot percentage, and stability analysis, it was found that MPC controllers outperform PID for those parameters.

Keywords: MPC, PID, ARX, pasteurization

Procedia PDF Downloads 131

Authors: David Adade-Boateng, Benard Fei Baffoe, Colin A. Booth, Michael A. Fullen

Abstract:

Remediation of diesel fuel, oil and grease in contaminated soils obtained from a mine site in Ghana are explored using rhizoremediation technology with different levels of nutrient amendments (i.e. N (nitrogen) in Compost (0.2, 0.5 and 0.8%), Urea (0.2, 0.5 and 0.8%) and Topsoil (0.2, 0.5 and 0.8%)) for a native species. A Ghanaian native grass species, Paspalum spp. from the Poaceae family, indicative across Sub-Saharan Africa, was selected following the development of essential and desirable growth criteria. Vegetative parts of the species were subjected to ten treatments in a Randomized Complete Block Design (RCBD) in three replicates. The plant-associated microbial community was examined in Paspalum spp. An assessment of the influence of Paspalum spp on the abundance and activity of micro-organisms in the rhizosphere revealed a build-up of microbial communities over a three month period. This was assessed using the MPN method, which showed rhizospheric samples from the treatments were significantly different (P <0.05). Multiple comparisons showed how microbial populations built-up in the rhizosphere for the different treatments. Treatments G (0.2% compost), H (0.5% compost) and I (0.8% compost) performed significantly better done other treatments, while treatments D (0.2% topsoil) and F (0.8% topsoil) were insignificant. Furthermore, treatment A (0.2% urea), B (0.5% urea), C (0.8% urea) and E (0.5% topsoil) also performed the same. Residual diesel and oil concentrations (as total petroleum hydrocarbons, TPH and oil and grease) were measured using infra-red spectroscopy and gravimetric methods, respectively. The presence of single species successfully enhanced the removal of hydrocarbons from soil. Paspalum spp. subjected to compost levels (0.5% and 0.8%) and topsoil levels (0.5% and 0.8%) showed significantly lower residual hydrocarbon concentrations compared to those treated with Urea. A strong relationship (p<0.001) between the abundance of hydrocarbon degrading micro-organisms in the rhizosphere and hydrocarbon biodegradation was demonstrated for rhizospheric samples with treatment G (0.2% compost), H (0.5% compost) and I (0.8% compost) (P <0.001). The same level of amendment with 0.8% compost (N-level) can improve the application effectiveness. These findings have wide-reaching implications for the environmental management of soils contaminated by hydrocarbons in Sub-Saharan Africa. However, it is necessary to further investigate the in situ rhizoremediation potential of Paspalum spp. at the field scale.

Keywords: rhizoremediation, microbial population, rhizospheric sample, treatments

Procedia PDF Downloads 283

Authors: Victor Breux, Jérôme Boutet, Alain Goret, Viviane Cattin

Abstract:

Early detection of anomalies in data centers is important to reduce downtimes and the costs of periodic maintenance. However, there is little research on this topic and even fewer on the fusion of sensor data for the detection of abnormal events. The goal of this paper is to propose a method for anomaly detection in data centers by combining sensor data (temperature, humidity, power) and deep learning models. The model described in the paper uses one autoencoder per sensor to reconstruct the inputs. The auto-encoders contain Long-Short Term Memory (LSTM) layers and are trained using the normal samples of the relevant sensors selected by correlation analysis. The difference signal between the input and its reconstruction is then used to classify the samples using feature extraction and a random forest classifier. The data measured by the sensors of a data center between January 2019 and May 2020 are used to train the model, while the data between June 2020 and May 2021 are used to assess it. Performances of the model are assessed a posteriori through F1-score by comparing detected anomalies with the data center’s history. The proposed model outperforms the state-of-the-art reconstruction method, which uses only one autoencoder taking multivariate sequences and detects an anomaly with a threshold on the reconstruction error, with an F1-score of 83.60% compared to 24.16%.

Keywords: anomaly detection, autoencoder, data centers, deep learning

Procedia PDF Downloads 164

Authors: YingWei Tan, XueFeng Ding

Abstract:

Attention-based voice activity detection models have gained significant attention in recent years due to their fast training speed and ability to capture a wide contextual range. The inclusion of multi-head style and position embedding in the attention architecture are crucial. Having multiple attention heads allows for differential focus on different parts of the sequence, while position embedding provides guidance for modeling dependencies between elements at various positions in the input sequence. In this work, we propose an approach by considering each head as a node, enabling the application of graph neural networks (GNN) to identify correlations among the different nodes. In addition, we adopt an implementation named rotary position embedding (RoPE), which encodes absolute positional information into the input sequence by a rotation matrix, and naturally incorporates explicit relative position information into a self-attention module. We evaluate the effectiveness of our method on a synthetic dataset, and the results demonstrate its superiority over the baseline CRNN in scenarios with low signal-to-noise ratio and noise, while also exhibiting robustness across different noise types. In summary, our proposed framework effectively combines the strengths of CNN and RNN (LSTM), and further enhances detection performance through the integration of graph neural networks and rotary position embedding.

Keywords: voice activity detection, CRNN, graph neural networks, rotary position embedding

Procedia PDF Downloads 36

Authors: Arpit Rai

Abstract:

In this work, we propose PatchMix, a semi-supervised method for pre-training visual representations. PatchMix mixes patches of two images and then solves an auxiliary task of predicting the label of each patch in the mixed image. Our experiments on the CIFAR-10, 100 and the SVHN dataset show that the representations learned by this method encodes useful information for transfer to new tasks and outperform the baseline Residual Network encoders by on CIFAR 10 by 12% on ResNet 101 and 2% on ResNet-56, by 4% on CIFAR-100 on ResNet101 and by 6% on SVHN dataset on the ResNet-101 baseline model.

Keywords: self-supervised learning, representation learning, computer vision, generalization

Procedia PDF Downloads 63

Authors: Khalid Ali, Manar Jammal

Abstract:

In order to meet the stringent latency and reliability requirements of the upcoming 5G networks, Open Radio Access Networks (O-RAN) have been proposed. The virtualization of O-RAN has allowed it to be treated as a Network Function Virtualization (NFV) architecture, while its components are considered Virtualized Network Functions (VNFs). Hence, intelligent Machine Learning (ML) based solutions can be utilized to apply different resource management and allocation techniques on O-RAN. However, intelligently allocating resources for O-RAN VNFs can prove challenging due to the dynamicity of traffic in mobile networks. Network providers need to dynamically scale the allocated resources in response to the incoming traffic. Elastically allocating resources can provide a higher level of flexibility in the network in addition to reducing the OPerational EXpenditure (OPEX) and increasing the resources utilization. Most of the existing elastic solutions are reactive in nature, despite the fact that proactive approaches are more agile since they scale instances ahead of time by predicting the incoming traffic. In this work, we propose and evaluate traffic forecasting models based on the ML algorithm. The algorithms aim at predicting future O-RAN traffic by using previous traffic data. Detailed analysis of the traffic data was carried out to validate the quality and applicability of the traffic dataset. Hence, two ML models were proposed and evaluated based on their prediction capabilities.

Keywords: O-RAN, traffic forecasting, NFV, ARIMA, LSTM, elasticity

Procedia PDF Downloads 186

Authors: Supakorn Tantisriyanurak, Hussaya Maneesuwan, Thanyalak Chaisuwan, Sujitra Wongkasemjit

Abstract:

TUD-1 is a siliceous mesoporous material with a three-dimensional amorphous structure of random, interconnecting pores, large pore size, high surface area (400-1000 m2/g), hydrothermal stability, and tunable porosity. However, the significant disadvantage of the mesoporous silicates is few catalytic active sites. In this work, a series of bimetallic Fe and Ti incorporated into TUD-1 framework is successfully synthesized by sol–gel method. The synthesized Fe,Ti-TUD-1 is characterized by various techniques. To study the catalytic activity of Fe, Ti–TUD-1, phenol hydroxylation was selected as a model reaction. The amounts of residual phenol and oxidation products were determined by high performance liquid chromatography coupled with UV-detector (HPLC-UV).

Keywords: iron, phenol hydroxylation, titanium, TUD-1

Procedia PDF Downloads 232

Authors: Leah Lee

Abstract:

This paper examines the correlation between deep learning models and emotions with noises to see whether or not noises mask emotions. The deep learning models used are plain convolutional neural networks (CNN), auto-encoder, long short-term memory (LSTM), and Visual Geometry Group-16 (VGG-16). Emotion datasets used are Ryerson Audio-Visual Database of Emotional Speech and Song (RAVDESS), Crowd-sourced Emotional Multimodal Actors Dataset (CREMA-D), Toronto Emotional Speech Set (TESS), and Surrey Audio-Visual Expressed Emotion (SAVEE). To make it four times bigger, audio set files, stretch, and pitch augmentations are utilized. From the augmented datasets, five different features are extracted for inputs of the models. There are eight different emotions to be classified. Noise variations are white noise, dog barking, and cough sounds. The variation in the signal-to-noise ratio (SNR) is 0, 20, and 40. In summation, per a deep learning model, nine different sets with noise and SNR variations and just augmented audio files without any noises will be used in the experiment. To compare the results of the deep learning models, the accuracy and receiver operating characteristic (ROC) are checked.

Keywords: auto-encoder, convolutional neural networks, long short-term memory, speech emotion recognition, visual geometry group-16

Procedia PDF Downloads 47

Authors: M. Ndaw, G. Mendy, S. Ouya

Abstract:

Information system risk management helps to reduce or eliminate risk by implementing appropriate controls. In this paper, we propose a quantification model of controls impact on information system risks by automatizing the residual criticality estimation step of FMECA which is based on a inductive reasoning. For this, we defined three equations based on type and maturity of controls. For testing, the values obtained with the model were compared to estimated values given by interlocutors during different working sessions and the result is satisfactory. This model allows an optimal assessment of controls maturity and facilitates risk analysis of information system.

Keywords: information system, risk, control, FMECA method

Procedia PDF Downloads 333

Authors: Jimeng Shi, Mahek Jain, Giri Narasimhan

Abstract:

Time Series Forecasting (TSF) is used to predict the target variables at a future time point based on the learning from previous time points. To keep the problem tractable, learning methods use data from a fixed-length window in the past as an explicit input. In this paper, we study how the performance of predictive models changes as a function of different look-back window sizes and different amounts of time to predict the future. We also consider the performance of the recent attention-based Transformer models, which have had good success in the image processing and natural language processing domains. In all, we compare four different deep learning methods (RNN, LSTM, GRU, and Transformer) along with a baseline method. The dataset (hourly) we used is the Beijing Air Quality Dataset from the UCI website, which includes a multivariate time series of many factors measured on an hourly basis for a period of 5 years (2010-14). For each model, we also report on the relationship between the performance and the look-back window sizes and the number of predicted time points into the future. Our experiments suggest that Transformer models have the best performance with the lowest Mean Average Errors (MAE = 14.599, 23.273) and Root Mean Square Errors (RSME = 23.573, 38.131) for most of our single-step and multi-steps predictions. The best size for the look-back window to predict 1 hour into the future appears to be one day, while 2 or 4 days perform the best to predict 3 hours into the future.

Keywords: air quality prediction, deep learning algorithms, time series forecasting, look-back window

Procedia PDF Downloads 134

Authors: Niklas Panten, Eberhard Abele

Abstract:

This paper presents a novel approach for real-time and near-optimal control of industrial smart grids by deep reinforcement learning (DRL). To achieve highly energy-efficient factory systems, the energetic linkage of machines, technical building equipment and the building itself is desirable. However, the increased complexity of the interacting sub-systems, multiple time-variant target values and stochastic influences by the production environment, weather and energy markets make it difficult to efficiently control the energy production, storage and consumption in the hybrid industrial smart grids. The studied deep reinforcement learning approach allows to explore the solution space for proper control policies which minimize a cost function. The deep neural network of the DRL agent is based on a multilayer perceptron (MLP), Long Short-Term Memory (LSTM) and convolutional layers. The agent is trained within multiple Modelica-based factory simulation environments by the Advantage Actor Critic algorithm (A2C). The DRL controller is evaluated by means of the simulation and then compared to a conventional, rule-based approach. Finally, the results indicate that the DRL approach is able to improve the control performance and significantly reduce energy respectively operating costs of industrial smart grids.

Keywords: industrial smart grids, energy efficiency, deep reinforcement learning, optimal control

Procedia PDF Downloads 169

Authors: Muhammad S. El-Feky, Mohamed I. Serag, Ahmed M. Yasien, Hala Elkady

Abstract:

Reinforced concrete requires steel bars in order to provide the tensile strength that is needed in structural concrete. However, when steel bars corrode, a loss in bond between the concrete and the steel bars occurs due to the formation of rust on the bars surface. Permeability of concrete is a fundamental property in perspective of the durability of concrete as it represents the ease with which water or other fluids can move through concrete, subsequently transporting corrosive agents. Nanotechnology is a standout amongst active research zones that envelops varies disciplines including construction materials. The application of nanotechnology in the corrosion protection of metal has lately gained momentum as nano scale particles have ultimate physical, chemical and physicochemical properties, which may enhance the corrosion protection in comparison to large size materials. The presented research aims to study the bond performance of concrete containing relatively high volume nano silica (up to 4.5%) exposed to corrosive conditions. This was extensively studied through tensile, bond strengths as well as the permeability of nano silica concrete. In addition micro-structural analysis was performed in order to evaluate the effect of nano silica on the properties of concrete at both; the micro and nano levels. The results revealed that by the addition of nano silica, the permeability of concrete mixes decreased significantly to reach about 50% of the control mix by the addition of 4.5% nano silica. As for the corrosion resistance, the nano silica concrete is comparatively higher resistance than ordinary concrete. Increasing Nano Silica percentage increased significantly the critical time corresponding to a metal loss (equal to 50 ϻm) which usually corresponding to the first concrete cracking due to the corrosion of reinforcement to reach about 49 years instead of 40 years as for the normal concrete. Finally, increasing nano Silica percentage increased significantly the residual bond strength of concrete after being subjected to corrosive environment. After being subjected to corrosive environment, the pullout behavior was observed for the bars embedded in all of the mixes instead of the splitting behavior that was observed before being corroded. Adding 4.5% nano silica in concrete increased the residual bond strength to reach 79% instead of 27% only as compared to control mix (0%W) before the subjection of the corrosive environment. From the conducted study we can conclude that the Nano silica proved to be a significant pore blocker material.

Keywords: bond strength, concrete, corrosion resistance, nano silica, permeability

Procedia PDF Downloads 284

Authors: Raymond Xu, Cindy Jingru Wang

Abstract:

Social media has become a globally connected and influencing platform. Social media data, such as tweets, can help predict the spread of pandemics and provide individuals and healthcare providers early warnings. Public psychological reactions and opinions can be efficiently monitored by AI models on the progression of dominant topics on Twitter. However, statistics show that as the coronavirus spreads, so does an infodemic of misinformation due to pandemic-related factors such as unemployment and lockdowns. Social media algorithms are often biased toward outrage by promoting content that people have an emotional reaction to and are likely to engage with. This can influence users’ attitudes and cause confusion. Therefore, social media is a double-edged sword. Combating fake news and biased content has become one of the essential tasks. This research analyzes the variety of methods used for fake news detection covering random forest, logistic regression, support vector machines, decision tree, naive Bayes, BoW, TF-IDF, LDA, CNN, RNN, LSTM, DeepFake, and hierarchical attention network. The performance of each method is analyzed. Based on these models’ achievements and limitations, a multi-dimensional AI framework is proposed to achieve higher accuracy in infodemic detection, especially pandemic-related news. The model is trained on contextual content, images, and news metadata.

Keywords: artificial intelligence, fake news detection, infodemic detection, image recognition, sentiment analysis

Procedia PDF Downloads 213

Authors: Ashish Jaiswal, Ashwin Ramesh Babu, Mohammad Zaki Zadeh, Fillia Makedon, Glenn Wylie

Abstract:

Functional magnetic resonance imaging (fMRI) is a neuroimaging technique that records neural activations in the brain by capturing the blood oxygen level in different regions based on the task performed by a subject. Given fMRI data, the problem of predicting the state of cognitive fatigue in a person has not been investigated to its full extent. This paper proposes tackling this issue as a multi-class classification problem by dividing the state of cognitive fatigue into six different levels, ranging from no-fatigue to extreme fatigue conditions. We built a spatio-temporal model that uses convolutional neural networks (CNN) for spatial feature extraction and a long short-term memory (LSTM) network for temporal modeling of 4D fMRI scans. We also applied a self-supervised method called MoCo (Momentum Contrast) to pre-train our model on a public dataset BOLD5000 and fine-tuned it on our labeled dataset to predict cognitive fatigue. Our novel dataset contains fMRI scans from Traumatic Brain Injury (TBI) patients and healthy controls (HCs) while performing a series of N-back cognitive tasks. This method establishes a state-of-the-art technique to analyze cognitive fatigue from fMRI data and beats previous approaches to solve this problem.

Keywords: fMRI, brain imaging, deep learning, self-supervised learning, contrastive learning, cognitive fatigue

Procedia PDF Downloads 162