Search results for: models synthesis

Authors: Esther Chiramal, Kelvin Soh Boon Kai

Abstract:

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

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

Procedia PDF Downloads 137

Authors: Sunidhi Syal, Rasangi Tennakoon, Patrick O'Donoghue

Abstract:

Polyglutamine (polyQ) diseases are a group of diseases related to neurodegeneration caused by repeats of the amino acid glutamine (Q) in the DNA, which translates into an elongated polyQ tract in the protein. The pathological explanation is that the polyQ tract forms cytotoxic aggregates in the neurons, leading to their degeneration. There are no cures or preventative efforts established for these diseases as of today, although the symptoms of these diseases can be relieved. This study specifically focuses on Huntington's disease, which is a type of polyQ disease in which aggregation is caused by the extended cytosine, adenine, guanine (CUG) codon repeats in the huntingtin (HTT) gene, which encodes for the huntingtin protein. Using this principle, we attempted to create six models, which included mutating wildtype tRNA alanine variant tRNA-AGC-8-1 to have glutamine anticodons CUG and UUG so serine is incorporated at glutamine sites in poly Q tracts. In the process, we were successful in obtaining tAla-8-1 CUG mutant clones in the HTTexon1 plasmids with a polyQ tract of 23Q (non-pathogenic model) and 74Q (disease model). These plasmids were transfected into mouse neuroblastoma cells to characterize protein synthesis and aggregation in normal and mistranslating cells and to investigate the effects of glutamines replaced with alanines on the disease phenotype. Notably, we observed no noteworthy differences in mean fluorescence between the CUG mutants for 23Q or 74Q; however, the Triton X-100 assay revealed a significant reduction in insoluble 74Q aggregates. We were unable to create a tAla-8-1 UUG mutant clone, and determining the difference in the effects of the two glutamine anticodons may enrich our understanding of the disease phenotype. In conclusion, by generating structural disruption with the amino acid alanine, it may be possible to find ways to minimize the toxicity of Huntington's disease caused by these polyQ aggregates. Further research is needed to advance knowledge in this field by identifying the cellular and biochemical impact of specific tRNA variants found naturally in human genomes.

Keywords: Huntington's disease, polyQ, tRNA, anticodon, clone, overlap PCR

Procedia PDF Downloads 41

Authors: Supriya Gupta, Paresh Chokshi

Abstract:

Polymer-particle interactions can be effectively utilized to produce composites that possess physicochemical properties superior to that of neat polymer. The incorporation of fillers with dimensions comparable to polymer chain size produces composites with extra-ordinary properties owing to very high surface to volume ratio. The dispersion of nanoparticles is achieved by inducing steric repulsion realized by grafting particles with polymeric chains. A comprehensive understanding of the interparticle interaction between these functionalized nanoparticles plays an important role in the synthesis of a stable polymer nanocomposite. With the focus on incorporation of clay sheets in a polymer matrix, we theoretically construct the polymer mediated interparticle potential for two nanosheets grafted with polymeric chains. The self-consistent field theory (SCFT) is employed to obtain the inhomogeneous composition field under equilibrium. Unlike the continuum models, SCFT is built from the microscopic description taking in to account the molecular interactions contributed by both intra- and inter-chain potentials. We present the results of SCFT calculations of the interaction potential curve for two grafted nanosheets immersed in the matrix of polymeric chains of dissimilar chemistry to that of the grafted chains. The interaction potential is repulsive at short separation and shows depletion attraction for moderate separations induced by high grafting density. It is found that the strength of attraction well can be tuned by altering the compatibility between the grafted and the mobile chains. Further, we construct the interaction potential between two nanosheets grafted with diblock copolymers with one of the blocks being chemically identical to the free polymeric chains. The interplay between the enthalpic interaction between the dissimilar species and the entropy of the free chains gives rise to a rich behavior in interaction potential curve obtained for two separate cases of free chains being chemically similar to either the grafted block or the free block of the grafted diblock chains.

Keywords: clay nanosheets, polymer brush, polymer nanocomposites, self-consistent field theory

Procedia PDF Downloads 252

Authors: Mohammed Sghir Taleb

Abstract:

Forest ecosystems in Morocco are subject increasingly to natural and human pressures. Conscious of this problem, Morocco set a strategy that focuses on programs of in-situ and ex-situ biodiversity conservation. This study is the result of a synthesis of various existing studies on biodiversity and forest ecosystems. It gives an overview of Moroccan mountain forest ecosystems and flora diversity. It also focuses on the efforts made by Morocco to conserve and sustainably manage biodiversity.

Keywords: mountain, ecosystems, conservation, Morocco

Procedia PDF Downloads 582

Authors: Seyed-Ali Sadegh-Zadeh, Kaveh Kavianpour, Hamed Atashbar, Elham Heidari, Saeed Shiry Ghidary, Amir M. Hajiyavand

Abstract:

Evaluation metrics play a critical role in assessing the performance of machine learning models. In this review paper, we provide a comprehensive overview of performance measurement approaches for machine learning models. For each category, we discuss the most widely used metrics, including their mathematical formulations and interpretation. Additionally, we provide a comparative analysis of performance measurement approaches for metric combinations. Our review paper aims to provide researchers and practitioners with a better understanding of performance measurement approaches and to aid in the selection of appropriate evaluation metrics for their specific applications.

Keywords: evaluation metrics, performance measurement, supervised learning, unsupervised learning, reinforcement learning, model robustness and stability, comparative analysis

Procedia PDF Downloads 73

Authors: Chun-Fang Hsueh, Chen-Kang Chang

Abstract:

Introduction: Branched-chain amino acids (BCAA) could reduce cerebral uptake of tryptophan, leading to decreased synthesis of serotonin in the brain. Arginine and citrulline could reduce exercise-induced hyperammonemia by increasing nitric oxide synthesis and the urea cycle. The combination of these supplements could reduce exercise-induced central fatigue. The purpose of this study was to examine the effect of BCAA, arginine, and citrulline supplementation on repeated swimming performance in teenage athletes. Methods: Eight male and eight female high school swimmers ingested 0.085 g/kg BCAA, 0.05 g/kg arginine and 0.05 g/kg citrulline (AA trial) or placebo (PL trial) in a randomized cross-over design. One hour after the ingestion, the subjects performed a 50 m sprint with their best style every 2 min for 8 times in an indoor 25 m pool. The subjects were asked to swim with their maximal effort each time. The time, stroke frequency and stroke length in each sprint were recorded. Venous blood samples were collected before and after the exercise. The time for each sprint was analyzed by 2-way analysis of variance with repeated measurement. Results: When all subjects were pooled together, total time for the AA trial was significantly faster than the PL trial (AA: 244.02 ± 22.94 s; PL: 247.55 ± 24.17 s, p < .001). Individual sprint time showed significant trial (p= .001) and trial x time (p= .004) effects. The post-hoc analysis revealed that the AA trial was significantly faster than the PL trial in the 2nd, 5th, and 6th sprint. In female subjects, there is a significant trial effect (p= .004) with the AA trial being faster in the 1st, 2nd, and 5th sprint. On the other hand, the trial effect was not significant (p= .072) in male subjects. Conclusions: The combined supplementation could improve 8 x 50 m performance in high school swimmers. The blood parameters including BCAA, tryptophan, NH₃, nitric oxide, and urea, as well as the stroke frequency and length in each sprint, are being analyzed. The results will be presented in the conference.

Keywords: central fatigue, hyperammonemia, tryptophan, urea

Procedia PDF Downloads 184

Authors: Sara ElElimy, Samir Moustafa

Abstract:

Mobile network operators start to face many challenges in the digital era, especially with high demands from customers. Since mobile network operators are considered a source of big data, traditional techniques are not effective with new era of big data, Internet of things (IoT) and 5G; as a result, handling effectively different big datasets becomes a vital task for operators with the continuous growth of data and moving from long term evolution (LTE) to 5G. So, there is an urgent need for effective Big data analytics to predict future demands, traffic, and network performance to full fill the requirements of the fifth generation of mobile network technology. In this paper, we introduce data science techniques using machine learning and deep learning algorithms: the autoregressive integrated moving average (ARIMA), Bayesian-based curve fitting, and recurrent neural network (RNN) are employed for a data-driven application to mobile network operators. The main framework included in models are identification parameters of each model, estimation, prediction, and final data-driven application of this prediction from business and network performance applications. These models are applied to Telecom Italia Big Data challenge call detail records (CDRs) datasets. The performance of these models is found out using a specific well-known evaluation criteria shows that ARIMA (machine learning-based model) is more accurate as a predictive model in such a dataset than the RNN (deep learning model).

Keywords: big data analytics, machine learning, CDRs, 5G

Procedia PDF Downloads 139

Authors: Ankit Sinha, Soham Banerjee, Pratik Chattopadhyay

Abstract:

Automated product recognition in retail stores is an important real-world application in the domain of Computer Vision and Pattern Recognition. In this paper, we consider the problem of automatically identifying the classes of the products placed on racks in retail stores from an image of the rack and information about the query/product images. We improve upon the existing approaches in terms of effectiveness and memory requirement by developing a two-stage object detection and recognition pipeline comprising of a Faster-RCNN-based object localizer that detects the object regions in the rack image and a ResNet-18-based image encoder that classifies the detected regions into the appropriate classes. Each of the models is fine-tuned using appropriate data sets for better prediction and data augmentation is performed on each query image to prepare an extensive gallery set for fine-tuning the ResNet-18-based product recognition model. This encoder is trained using a triplet loss function following the strategy of online-hard-negative-mining for improved prediction. The proposed models are lightweight and can be connected in an end-to-end manner during deployment to automatically identify each product object placed in a rack image. Extensive experiments using Grozi-32k and GP-180 data sets verify the effectiveness of the proposed model.

Keywords: retail stores, faster-RCNN, object localization, ResNet-18, triplet loss, data augmentation, product recognition

Procedia PDF Downloads 156

Authors: Sujit Kumar Pradhan, Anil Kumar, Vijay Kumar

Abstract:

The testing process of the software during the software development time is a crucial step as it makes the software more efficient and dependable. To estimate software’s reliability through the mean value function, many software reliability growth models (SRGMs) were developed under the assumption that operating and testing environments are the same. Practically, it is not true because when the software works in a natural field environment, the reliability of the software differs. This article discussed an SRGM comprising change-point and imperfect debugging in a simulated testing environment. Later on, we extended it in a multi-release direction. Initially, the software was released to the market with few features. According to the market’s demand, the software company upgraded the current version by adding new features as time passed. Therefore, we have proposed a generalized multi-release SRGM where change-point and imperfect debugging concepts have been addressed in a simulated testing environment. The failure-increasing rate concept has been adopted to determine the change point for each software release. Based on nine goodness-of-fit criteria, the proposed model is validated on two real datasets. The results demonstrate that the proposed model fits the datasets better. We have also discussed the optimal release time of the software through a cost model by assuming that the testing and debugging costs are time-dependent.

Keywords: software reliability growth models, non-homogeneous Poisson process, multi-release software, mean value function, change-point, environmental factors

Procedia PDF Downloads 74

Authors: Sutapa Mondal Roy, Suban K. Sahoo

Abstract:

The primary aim of this work is to synthesis silver nanoparticles (AgNPs) through environmentally benign routes to avoid any chemical toxicity related undesired side effects. The nanoparticles were stabilized with drug ciprofloxacin (Cp) and were studied for their effectiveness as drug delivery agent. Targeted drug delivery improves the therapeutic potential of drugs at the diseased site as well as lowers the overall dose and undesired side effects. The small size of nanoparticles greatly facilitates the transport of active agents (drugs) across biological membranes and allows them to pass through the smallest capillaries in the body that are 5-6 μm in diameter, and can minimize possible undesired side effects. AgNPs are non-toxic, inert, stable, and has a high binding capacity and thus can be considered as biomaterials. AgNPs were synthesized from the nutrient broth supernatant after the culture of environment-friendly bacteria Bacillus subtilis. The AgNPs were found to show the surface plasmon resonance (SPR) band at 425 nm. The Cp capped Ag nanoparticles formation was complete within 30 minutes, which was confirmed from absorbance spectroscopy. Physico-chemical nature of the AgNPs-Cp system was confirmed by Dynamic Light Scattering (DLS), Transmission Electron Microscopy (TEM) etc. The AgNPs-Cp system size was found to be in the range of 30-40 nm. To monitor the kinetics of drug release from the surface of nanoparticles, the release of Cp was carried out by careful dialysis keeping AgNPs-Cp system inside the dialysis bag at pH 7.4 over time. The drug release was almost complete after 30 hrs. During the drug delivery process, to understand the AgNPs-Cp system in a better way, the sincere theoretical investigation is been performed employing Density Functional Theory. Electronic charge transfer, electron density, binding energy as well as thermodynamic properties like enthalpy, entropy, Gibbs free energy etc. has been predicted. The electronic and thermodynamic properties, governed by the AgNPs-Cp interactions, indicate that the formation of AgNPs-Cp system is exothermic i.e. thermodynamically favorable process. The binding energy and charge transfer analysis implies the optimum stability of the AgNPs-Cp system. Thus, the synthesized Cp-Ag nanoparticles can be effectively used for biological purposes due to its environmentally benign routes of synthesis procedures, which is clean, biocompatible, non-toxic, safe, cost-effective, sustainable and eco-friendly. The Cp-AgNPs as biomaterials can be successfully used for drug delivery procedures due to slow release of drug from nanoparticles over a considerable period of time. The kinetics of the drug release show that this drug-nanoparticle assembly can be effectively used as potential tools for therapeutic applications. The ease of synthetic procedure, lack of possible chemical toxicity and their biological activity along with excellent application as drug delivery agent will open up vista of using nanoparticles as effective and successful drug delivery agent to be used in modern days.

Keywords: silver nanoparticles, ciprofloxacin, density functional theory, drug delivery

Procedia PDF Downloads 384

Authors: Milica Karadzic, Lidija Jevric, Sanja Podunavac-Kuzmanovic, Strahinja Kovacevic, Anamarija Mandic, Katarina Penov Gasi, Marija Sakac, Aleksandar Okljesa, Andrea Nikolic

Abstract:

The present paper deals with chromatographic lipophilicity prediction of newly synthesized steroid derivatives. The prediction was achieved using in silico generated molecular descriptors and quantitative structure-retention relationship (QSRR) methodology with the artificial neural networks (ANN) approach. Chromatographic lipophilicity of the investigated compounds was expressed as retention factor value logk. For QSRR modeling, a feedforward back-propagation ANN with gradient descent learning algorithm was applied. Using the novel sum of ranking differences (SRD) method generated ANN models were ranked. The aim was to distinguish the most consistent QSRR model that can be found, and similarity or dissimilarity between the models that could be noticed. In this study, SRD was performed with average values of retention factor value logk as reference values. An excellent correlation between experimentally observed retention factor value logk and values predicted by the ANN was obtained with a correlation coefficient higher than 0.9890. Statistical results show that the established ANN models can be applied for required purpose. This article is based upon work from COST Action (TD1305), supported by COST (European Cooperation in Science and Technology).

Keywords: artificial neural networks, liquid chromatography, molecular descriptors, steroids, sum of ranking differences

Procedia PDF Downloads 319

Authors: Farid Khosravikia, Patricia Clayton

Abstract:

The main objective of this work is to evaluate the advantages and disadvantages of various machine learning techniques in two key steps of seismic hazard and risk assessment of different types of structures. The first step is the development of ground-motion models, which are used for forecasting ground-motion intensity measures (IM) given source characteristics, source-to-site distance, and local site condition for future events. IMs such as peak ground acceleration and velocity (PGA and PGV, respectively) as well as 5% damped elastic pseudospectral accelerations at different periods (PSA), are indicators of the strength of shaking at the ground surface. Typically, linear regression-based models, with pre-defined equations and coefficients, are used in ground motion prediction. However, due to the restrictions of the linear regression methods, such models may not capture more complex nonlinear behaviors that exist in the data. Thus, this study comparatively investigates potential benefits from employing other machine learning techniques as statistical method in ground motion prediction such as Artificial Neural Network, Random Forest, and Support Vector Machine. The results indicate the algorithms satisfy some physically sound characteristics such as magnitude scaling distance dependency without requiring pre-defined equations or coefficients. Moreover, it is shown that, when sufficient data is available, all the alternative algorithms tend to provide more accurate estimates compared to the conventional linear regression-based method, and particularly, Random Forest outperforms the other algorithms. However, the conventional method is a better tool when limited data is available. Second, it is investigated how machine learning techniques could be beneficial for developing probabilistic seismic demand models (PSDMs), which provide the relationship between the structural demand responses (e.g., component deformations, accelerations, internal forces, etc.) and the ground motion IMs. In the risk framework, such models are used to develop fragility curves estimating exceeding probability of damage for pre-defined limit states, and therefore, control the reliability of the predictions in the risk assessment. In this study, machine learning algorithms like artificial neural network, random forest, and support vector machine are adopted and trained on the demand parameters to derive PSDMs for them. It is observed that such models can provide more accurate estimates of prediction in relatively shorter about of time compared to conventional methods. Moreover, they can be used for sensitivity analysis of fragility curves with respect to many modeling parameters without necessarily requiring more intense numerical response-history analysis.

Keywords: artificial neural network, machine learning, random forest, seismic risk analysis, seismic hazard analysis, support vector machine

Procedia PDF Downloads 106

Authors: Mostafa Aminzadeh, Min Deng

Abstract:

Modeling loss data is an important part of actuarial science. Actuaries use models to predict future losses and manage financial risk, which can be beneficial for marketing purposes. In the insurance industry, small claims happen frequently while large claims are rare. Traditional distributions such as Normal, Exponential, and inverse-Gaussian are not suitable for describing insurance data, which often show skewness and fat tails. Several authors have studied classical and Bayesian inference for parameters of composite distributions, such as Exponential-Pareto, Weibull-Pareto, and Inverse Gamma-Pareto. These models separate small to moderate losses from large losses using a threshold parameter. This research introduces a computational approach using a nonlinear regression model for loss data that relies on multiple predictors. Simulation studies were conducted to assess the accuracy of the proposed estimation method. The simulations confirmed that the proposed method provides precise estimates for regression parameters. It's important to note that this approach can be applied to datasets if goodness-of-fit tests confirm that the composite distribution under study fits the data well. To demonstrate the computations, a real data set from the insurance industry is analyzed. A Mathematica code uses the Fisher information algorithm as an iteration method to obtain the maximum likelihood estimation (MLE) of regression parameters.

Keywords: maximum likelihood estimation, fisher scoring method, non-linear regression models, composite distributions

Procedia PDF Downloads 34

Authors: Widi Astuti, Rizki Agus Hermawan, Hariono Mukti, Nurul Retno Sugiyono

Abstract:

The removal of lead ion (Pb²⁺) from aqueous solution by activated carbon with phosphoric acid activation employing mangrove propagule as precursor was investigated in a batch adsorption system. Batch studies were carried out to address various experimental parameters including pH and contact time. The Langmuir and Freundlich models were able to describe the adsorption equilibrium, while the pseudo first order and pseudo second order models were used to describe kinetic process of Pb²⁺ adsorption. The results show that the adsorption data are seen in accordance with Langmuir isotherm model and pseudo-second order kinetic model.

Keywords: activated carbon, adsorption, equilibrium, kinetic, lead, mangrove propagule

Procedia PDF Downloads 167

Authors: Funmilayo L. Amao, Amos O. Amao

Abstract:

It has been observed that majority of the people are living in poor housing quality or totally homeless in urban center despite all governmental policies to provide housing to the public. On the supply side, various government policies in the past have been formulated towards overcoming the huge shortage through several Housing Reform Programmes. Despite these past efforts, housing continues to be a mirage to ordinary Nigerian. Currently, there are various mass housing delivery programmes such as the affordable housing scheme that utilize the Public Private Partnership effort and several Private Finance Initiative models could only provide for about 3% of the required stock. This suggests the need for a holistic solution in approaching the problem. The aim of this research is to find out the problems hindering the delivery of housing in Nigeria and its effects on housing affordability. The specific objectives are to identify the causes of housing delivery problems, to examine different housing policies over years and to suggest a way out for sustainable housing delivery. This paper also reviews the past and current housing delivery programmes in Nigeria and analyses the demand and supply side issues. It identifies the various housing delivery mechanisms in current practice. The objective of this paper, therefore, is to give you an insight into the delivery option for the sustainability of housing in Nigeria, given the existing delivery structures and the framework specified in the New National Housing Policy. The secondary data were obtained from books, journals and seminar papers. The conclusion is that we cannot copy models from other nations, but should rather evolve workable models based on our socio-cultural background to address the huge housing shortage in Nigeria. Recommendations are made in this regard.

Keywords: housing, sustainability, housing delivery, housing policy, housing affordability

Procedia PDF Downloads 296

Authors: P. Pötters, A. Marquet, B. Leyendecker

Abstract:

The objective of this paper is to investigate different implementation approaches for the implementation of Lean production in companies. Furthermore, a structured overview of those different approaches is to be made. Therefore, the present work is intended to answer the following research question: What differences and similarities exist between the various systematic approaches and phase models for the implementation of Lean Production? To present various approaches for the implementation of Lean Production discussed in the literature, a qualitative content analysis was conducted. Within the framework of a qualitative survey, a selection of texts dealing with lean production and its introduction was examined. The analysis presents different implementation approaches from the literature, covering the descriptive aspect of the study. The study also provides insights into similarities and differences among the implementation approaches, which are drawn from the analysis of latent text contents and author interpretations. In this study, the focus is on identifying differences and similarities among systemic approaches for implementing Lean Production. The research question takes into account the main object of consideration, objectives pursued, starting point, procedure, and endpoint of the implementation approach. The study defines the concept of Lean Production and presents various approaches described in literature that companies can use to implement Lean Production successfully. The study distinguishes between five systemic implementation approaches and seven phase models to help companies choose the most suitable approach for their implementation project. The findings of this study can contribute to enhancing transparency regarding the existing approaches for implementing Lean Production. This can enable companies to compare and contrast the available implementation approaches and choose the most suitable one for their specific project.

Keywords: implementation, lean production, phase models, systematic approaches

Procedia PDF Downloads 104

Authors: Malek Ali

Abstract:

Composites materials of Poly (vinyl alcohol) (PVA)/Chitosan (CS) have been synthesized and characterized successfully. HNTs have been added to composites to enhance the mechanical and degradation properties by hydrogen bonding interactions, compatibility, and chemical crosslink between HNTs and PVA. PVA/CS/HNTs composites prepared with different concentration ratio. SEM micrographs of composites surface showed that more agglomeration with more chitosan ratio. Mechanical and degradation properties were characterized and the result indicates that Mechanical and degradation of 80%PVA/5%Chitosan/15%HNTs higher than the others PVA/CS/HNTs composites.

Keywords: PVA/chitosan, composites, PVA/CS/HNTs, HNTs

Procedia PDF Downloads 292

Authors: Dianelys Vega, Carlos Magluta, Ney Roitman

Abstract:

The simulation of loads induced by walking people in civil engineering structures is still challenging It has been the focus of considerable research worldwide in the recent decades due to increasing number of reported vibration problems in pedestrian structures. One of the most important key in the designing of slender structures is the Human-Structure Interaction (HSI). How moving people interact with structures and the effect it has on their dynamic responses is still not well understood. To rely on calibrated pedestrian models that accurately estimate the structural response becomes extremely important. However, because of the complexity of the pedestrian mechanisms, there are still some gaps in knowledge and more reliable models need to be investigated. On this topic several authors have proposed biodynamic models to represent the pedestrian, whether these models provide a consistent approximation to physical reality still needs to be studied. Therefore, this work comes to contribute to a better understanding of this phenomenon bringing an experimental validation of a pedestrian walking model and a Human-Structure Interaction model. In this study, a bi-dimensional bipedal walking model was used to represent the pedestrians along with an interaction model which was applied to a prototype footbridge. Numerical models were implemented in MATLAB. In parallel, experimental tests were conducted in the Structures Laboratory of COPPE (LabEst), at Federal University of Rio de Janeiro. Different test subjects were asked to walk at different walking speeds over instrumented force platforms to measure the walking force and an accelerometer was placed at the waist of each subject to measure the acceleration of the center of mass at the same time. By fitting the step force and the center of mass acceleration through successive numerical simulations, the model parameters are estimated. In addition, experimental data of a walking pedestrian on a flexible structure was used to validate the interaction model presented, through the comparison of the measured and simulated structural response at mid span. It was found that the pedestrian model was able to adequately reproduce the ground reaction force and the center of mass acceleration for normal and slow walking speeds, being less efficient for faster speeds. Numerical simulations showed that biomechanical parameters such as leg stiffness and damping affect the ground reaction force, and the higher the walking speed the greater the leg length of the model. Besides, the interaction model was also capable to estimate with good approximation the structural response, that remained in the same order of magnitude as the measured response. Some differences in frequency spectra were observed, which are presumed to be due to the perfectly periodic loading representation, neglecting intra-subject variabilities. In conclusion, this work showed that the bipedal walking model could be used to represent walking pedestrians since it was efficient to reproduce the center of mass movement and ground reaction forces produced by humans. Furthermore, although more experimental validations are required, the interaction model also seems to be a useful framework to estimate the dynamic response of structures under loads induced by walking pedestrians.

Keywords: biodynamic models, bipedal walking models, human induced loads, human structure interaction

Procedia PDF Downloads 132

Authors: Wang Ye, Wei Wei

Abstract:

Residential block construction of big cities in China began in the 1950s, and four models had far-reaching influence on modern residential block in its development process, including unit compound and residential district in 1950s to 1980s, and gated community and open community in 1990s to now. Based on analysis of the four models’ fabric, the article takes residential blocks in Hangzhou west area as an example and carries on the studies from urban structure level and block special level, mainly including urban road network, land use, community function, road organization, public space and building fabric. At last, the article puts forward semi-open sub-community strategy to improve the current fabric.

Keywords: Hangzhou west area, residential block model, residential block fabric, semi-open sub-community strategy

Procedia PDF Downloads 417

Authors: Azita Ramezani, Ghazal Mashhadiagha, Bahareh Sanabakhsh

Abstract:

This study researches the combination of Random. Forest classifiers with large language models (LLMs) and natural language processing (NLP) to improve diagnostic accuracy in chest X-ray analysis using the Indiana University dataset. Utilizing advanced NLP techniques, the research preprocesses textual data from radiological reports to extract key features, which are then merged with image-derived data. This improved dataset is analyzed with Random Forest classifiers to predict specific clinical results, focusing on the identification of health issues and the estimation of case urgency. The findings reveal that the combination of NLP, LLMs, and machine learning not only increases diagnostic precision but also reliability, especially in quickly identifying critical conditions. Achieving an accuracy of 99.35%, the model shows significant advancements over conventional diagnostic techniques. The results emphasize the large potential of machine learning in medical imaging, suggesting that these technologies could greatly enhance clinician judgment and patient outcomes by offering quicker and more precise diagnostic approximations.

Keywords: natural language processing (NLP), large language models (LLMs), random forest classifier, chest x-ray analysis, medical imaging, diagnostic accuracy, indiana university dataset, machine learning in healthcare, predictive modeling, clinical decision support systems

Procedia PDF Downloads 44

Authors: Judson P. LaGrone

Abstract:

Simulation-based education in curricula was once a luxurious component of nursing programs but now serves as a vital element of an individual’s learning experience. A debriefing occurs after the simulation scenario or clinical experience is completed to allow the instructor(s) or trained professional(s) to act as a debriefer to guide a reflection with a purpose of acknowledging, assessing, and synthesizing the thought process, decision-making process, and actions/behaviors performed during the scenario or clinical experience. Debriefing is a vital component of the simulation process and educational experience to allow the learner(s) to progressively build upon past experiences and current scenarios within a safe and welcoming environment with a guided dialog to enhance future practice. The aim of this integrative review was to assess current practices of debriefing models in simulation-based education for health care professionals and students. The following databases were utilized for the search: CINAHL Plus, Cochrane Database of Systemic Reviews, EBSCO (ERIC), PsycINFO (Ovid), and Google Scholar. The advanced search option was useful to narrow down the search of articles (full text, Boolean operators, English language, peer-reviewed, published in the past five years). Key terms included debrief, debriefing, debriefing model, debriefing intervention, psychological debriefing, simulation, simulation-based education, simulation pedagogy, health care professional, nursing student, and learning process. Included studies focus on debriefing after clinical scenarios of nursing students, medical students, and interprofessional teams conducted between 2015 and 2020. Common themes were identified after the analysis of articles matching the search criteria. Several debriefing models are addressed in the literature with similarities of effectiveness for participants in clinical simulation-based pedagogy. Themes identified included (a) importance of debriefing in simulation-based pedagogy, (b) environment for which debriefing takes place is an important consideration, (c) individuals who should conduct the debrief, (d) length of debrief, and (e) methodology of the debrief. Debriefing models supported by theoretical frameworks and facilitated by trained staff are vital for a successful debriefing experience. Models differed from self-debriefing, facilitator-led debriefing, video-assisted debriefing, rapid cycle deliberate practice, and reflective debriefing. A reoccurring finding was centered around the emphasis of continued research for systematic tool development and analysis of the validity and effectiveness of current debriefing practices. There is a lack of consistency of debriefing models among nursing curriculum with an increasing rate of ill-prepared faculty to facilitate the debriefing phase of the simulation.

Keywords: debriefing model, debriefing intervention, health care professional, simulation-based education

Procedia PDF Downloads 142

Authors: Cecile Meier, Drago Diaz Aleman, Itahisa Perez Conesa, Jose Luis Saorin Perez, Jorge De La Torre Cantero

Abstract:

In this work, two ways of creating small-sized metal sculptures are proposed: the first by means of microcasting and the second by electroforming from models printed in 3D using an FDM (Fused Deposition Modeling‎) printer or using a DLP (Digital Light Processing) printer. It is viable to replace the wax in the processes of the artistic foundry with 3D printed objects. In this technique, the digital models are manufactured with resin using a low-cost 3D FDM printer in polylactic acid (PLA). This material is used, because its properties make it a viable substitute to wax, within the processes of artistic casting with the technique of lost wax through Ceramic Shell casting. This technique consists of covering a sculpture of wax or in this case PLA with several layers of thermoresistant material. This material is heated to melt the PLA, obtaining an empty mold that is later filled with the molten metal. It is verified that the PLA models reduce the cost and time compared with the hand modeling of the wax. In addition, one can manufacture parts with 3D printing that are not possible to create with manual techniques. However, the sculptures created with this technique have a size limit. The problem is that when printed pieces with PLA are very small, they lose detail, and the laminar texture hides the shape of the piece. DLP type printer allows obtaining more detailed and smaller pieces than the FDM. Such small models are quite difficult and complex to melt using the lost wax technique of Ceramic Shell casting. But, as an alternative, there are microcasting and electroforming, which are specialized in creating small metal pieces such as jewelry ones. The microcasting is a variant of the lost wax that consists of introducing the model in a cylinder in which the refractory material is also poured. The molds are heated in an oven to melt the model and cook them. Finally, the metal is poured into the still hot cylinders that rotate in a machine at high speed to properly distribute all the metal. Because microcasting requires expensive material and machinery to melt a piece of metal, electroforming is an alternative for this process. The electroforming uses models in different materials; for this study, micro-sculptures printed in 3D are used. These are subjected to an electroforming bath that covers the pieces with a very thin layer of metal. This work will investigate the recommended size to use 3D printers, both with PLA and resin and first tests are being done to validate use the electroforming process of microsculptures, which are printed in resin using a DLP printer.

Keywords: sculptures, DLP 3D printer, microcasting, electroforming, fused deposition modeling

Procedia PDF Downloads 135

Authors: Kasim Görenekli, Ali Gülbağ

Abstract:

This study presents a comprehensive analysis of water consumption patterns in Kocaeli province, Turkey, utilizing various machine learning approaches. We analyzed data from 5,000 water subscribers across residential, commercial, and official categories over an 80-month period from January 2016 to August 2022, resulting in a total of 400,000 records. The dataset encompasses water consumption records, weather information, weekends and holidays, previous months' consumption, and the influence of the COVID-19 pandemic.We implemented and compared several machine learning models, including Linear Regression, Random Forest, Support Vector Regression (SVR), XGBoost, Artificial Neural Networks (ANN), Long Short-Term Memory (LSTM), and Gated Recurrent Units (GRU). Particle Swarm Optimization (PSO) was applied to optimize hyperparameters for all models.Our results demonstrate varying performance across subscriber types and models. For official subscribers, Random Forest achieved the highest R² of 0.699 with PSO optimization. For commercial subscribers, Linear Regression performed best with an R² of 0.730 with PSO. Residential water usage proved more challenging to predict, with XGBoost achieving the highest R² of 0.572 with PSO.The study identified key factors influencing water consumption, with previous months' consumption, meter diameter, and weather conditions being among the most significant predictors. The impact of the COVID-19 pandemic on consumption patterns was also observed, particularly in residential usage.This research provides valuable insights for effective water resource management in Kocaeli and similar regions, considering Turkey's high water loss rate and below-average per capita water supply. The comparative analysis of different machine learning approaches offers a comprehensive framework for selecting appropriate models for water consumption prediction in urban settings.

Keywords: mMachine learning, water consumption prediction, particle swarm optimization, COVID-19, water resource management

Procedia PDF Downloads 15

Authors: Shalini Arora, Sri Sivakumar

Abstract:

The deep removal of high molecular weight sulphur compounds (e.g., 4,6, dimethyl dibenzothiophene) is challenging due to their steric hindrance. Hydrogenation desulfurization (HYD) pathway is the main pathway to remove these sulfur compounds, and it is mainly governed by the number of type 2 sites. The formation of type 2 sites can be enhanced by modulating the pore structure and the interaction between the active metal and support. To this end, we report the enhanced HDS catalytic activity of ultrasmall NiMo supported on amorphous alumina (A-Al₂O₃) catalysts by one pot colloidal synthesis method followed by calcination and sulfidation. The amorphous alumina (A-Al₂O₃) was chosen as the support due to its lower surface energy, better physicochemical properties, and enhanced acidic sites (due to the dominance of tetra and penta coordinated [Al] sites) than crystalline alumina phase. At 20% metal oxide composition, NiMo supported on A-Al₂O₃ catalyst showed 1.4 and 1.2 times more reaction rate constant and turn over frequency (TOF) respectively than the conventional catalyst (wet impregnated NiMo catalysts) for HDS reaction of dibenzothiophene reactant molecule. A-Al₂O₃ supported catalysts represented enhanced type 2 sites formation (because this catalystpossesses higher sulfidation degree (80%) and NiMoS sites (19.3 x 10¹⁷ sites/mg) with desired optimum stacking degree (2.5) than wet impregnated catalyst at same metal oxide composition 20%) along with higher active metal dispersion, Mo edge site fraction. The experimental observations were also supported by DFT simulations. Lower heat of adsorption (< 4.2 ev for MoS2 interaction and < 3.15 ev for Ni doped MoS2 interaction) values for A-Al₂O₃ confirmed the presence of weaker metal-support interaction in A-Al₂O₃ in contrast to crystalline ℽ-Al₂O3. The weak metal-support interaction for prepared catalysts clearly suggests the higher formation of type 2 sites which leads to higher catalytic activity for HDS reaction.

Keywords: amorphous alumina, colloidal, desulfurization, metal-support interaction

Procedia PDF Downloads 267

Authors: Masoumeh Ghorbania, Falk Liebnerb, Hendrikus W.G. van Herwijnenc, Johannes Konnertha

Abstract:

Phenol formaldehyde (PF) resins are widely used as wood adhesives for variety of industrial products such as plywood, laminated veneer lumber and others. Lignin as a main constituent of wood has become well-known as a potential substitute for phenol in PF adhesives because of their structural similarity. During the last decades numerous research approaches have been carried out to substitute phenol with pulping-derived lignin, whereby the lower reactivity of resins synthesized with shares of lignin seem to be one of the major challenges. This work reports about a systematic screening of different types of lignin (plant origin and pulping process) for their suitability to replace phenol in phenolic resins. Lignin from different plant sources (softwood, hardwood and grass) were used, as these should differ significantly in their reactivity towards formaldehyde of their reactive phenolic core units. Additionally a possible influence of the pulping process was addressed by using the different types of lignin from soda, kraft, and organosolv process and various lignosulfonates (sodium, ammonium, calcium, magnesium). To determine the influence of lignin on the adhesive performance beside others the rate of viscosity development, bond strength development of varying hot pressing time and other thermal properties were investigated. To evaluate the performance of the cured end product, a few selected properties were studied at the example of solid wood-adhesive bond joints, compact panels and plywood. As main results it was found that lignin significantly accelerates the viscosity development in adhesive synthesis. Bonding strength development during curing of adhesives decelerated for all lignin types, while this trend was least for pine kraft lignin and spruce sodium lignosulfonate. However, the overall performance of the products prepared with the latter adhesives was able to fulfill main standard requirements, even after exposing the products to harsh environmental conditions. Thus, a potential application can be considered for processes where reactivity is less critical but adhesive cost and product performance is essential.

Keywords: phenol formaldehyde resin, lignin phenol formaldehyde resin, ABES, DSC

Procedia PDF Downloads 237

Authors: C. Baidada, M. H. Abidi, A. Jakimi, E. H. El Kinani

Abstract:

Reverse engineering has become a viable method to measure an existing system and reconstruct the necessary model from tis original. The reverse engineering of behavioral models consists in extracting high-level models that help understand the behavior of existing software systems. In this paper, we propose an approach for the reverse engineering of sequence diagrams from the analysis of execution traces produced dynamically by an object-oriented application using petri nets. Our methods show that this approach can produce state diagrams in reasonable time and suggest that these diagrams are helpful in understanding the behavior of the underlying application. Finally we will discuss approachs and tools that are needed in the process of reverse engineering UML behavior. This work is a substantial step towards providing high-quality methodology for effectiveand efficient reverse engineering of sequence diagram.

Keywords: reverse engineering, UML behavior, sequence diagram, execution traces, petri nets

Procedia PDF Downloads 445

Authors: Adel Edwar Waheeb Louka

Abstract:

Purpose AI-driven solutions are at the forefront of many pathology and medical imaging methods. Using algorithms designed to better the experience of medical professionals within their respective fields, the efficiency and accuracy of diagnosis can improve. In particular, X-rays are a fast and relatively inexpensive test that can diagnose diseases. In recent years, X-rays have not been widely used to detect and diagnose COVID-19. The under use of Xrays is mainly due to the low diagnostic accuracy and confounding with pneumonia, another respiratory disease. However, research in this field has expressed a possibility that artificial neural networks can successfully diagnose COVID-19 with high accuracy. Models and Data The dataset used is the COVID-19 Radiography Database. This dataset includes images and masks of chest X-rays under the labels of COVID-19, normal, and pneumonia. The classification model developed uses an autoencoder and a pre-trained convolutional neural network (DenseNet201) to provide transfer learning to the model. The model then uses a deep neural network to finalize the feature extraction and predict the diagnosis for the input image. This model was trained on 4035 images and validated on 807 separate images from the ones used for training. The images used to train the classification model include an important feature: the pictures are cropped beforehand to eliminate distractions when training the model. The image segmentation model uses an improved U-Net architecture. This model is used to extract the lung mask from the chest X-ray image. The model is trained on 8577 images and validated on a validation split of 20%. These models are calculated using the external dataset for validation. The models’ accuracy, precision, recall, f1-score, IOU, and loss are calculated. Results The classification model achieved an accuracy of 97.65% and a loss of 0.1234 when differentiating COVID19-infected, pneumonia-infected, and normal lung X-rays. The segmentation model achieved an accuracy of 97.31% and an IOU of 0.928. Conclusion The models proposed can detect COVID-19, pneumonia, and normal lungs with high accuracy and derive the lung mask from a chest X-ray with similarly high accuracy. The hope is for these models to elevate the experience of medical professionals and provide insight into the future of the methods used.

Keywords: artificial intelligence, convolutional neural networks, deeplearning, image processing, machine learningSarapin, intraarticular, chronic knee pain, osteoarthritisFNS, trauma, hip, neck femur fracture, minimally invasive surgery

Procedia PDF Downloads 73

Authors: Harriet Koshie Lamptey, Richard Boateng

Abstract:

Mobile learning (m-learning) is a novel approach to knowledge acquisition and dissemination and is gaining global attention. Steady progress in wireless technologies and the portability of communication devices continue to broaden the scope and use of mobiles. With the convergence of Web functionality onto mobile platforms and the affordability and availability of mobile technology, m-learning has the potential of being the next prevalent channel of education in both formal and informal settings. There is substantive literature on developed countries but the state in developing countries (DCs) however appears vague. This paper is a synthesis of extant literature on mobile learning in DCs. The research interest is based on the fact that in DCs, mobile communication and internet connectivity are popular. However, its use in education is under explored. There are some reviews on the state, conceptualizations, trends and teacher education, but to the authors’ knowledge, no study has focused on mobile learning adoption and integration issues. This study examines issues and gaps associated with its adoption and integration in DCs higher education institutions. A qualitative build-up of literature was conducted using articles pooled from electronic databases (Google Scholar and ERIC). To enable criteria for inclusion and incorporate diverse study perspectives, search terms used were m-learning, DCs, higher education institutions, challenges, benefits, impact, gaps and issues. The synthesis revealed that though mobile technology has diffused globally, its pedagogical pursuit in DCs remains quite low. The absence of a mobile Web and the difficulty of resource conversion into mobile format due to lack of funding and technical competence is a stumbling block. Again, the lack of established design and implementation rules to guide the development of m-learning platforms in DCs is a hindrance. The absence of access restrictions on devices poses security threats to institutional systems. Negative perceptions that devices are taking over faculty roles lead to resistance in some situations. Resistance to change can be a hindrance to the acceptance and success of new systems. Lack of interest for m-learning is also attributed to lower technological literacy levels of the underprivileged masses. Scholarly works on m-learning in DCs is yet to mature. Most technological innovations are handed down from developed countries, and this constantly creates a lag for DCs. Lack of theoretical grounding was also identified which reduces the objectivity of study reports. The socio-cultural terrain of DCs results in societies with different views and needs that have been identified as a hindrance to research. Institutional commitment decisions, adequate funding for the necessary infrastructural development as well as multiple stakeholder participation is important for project success. Evidence suggests that while adoption decisions are readily made, successful integration of the concept for its full benefits to be realized is often neglected. Recommendations to findings were made to provide possible remedies to identified issues.

Keywords: developing countries, higher education institutions, mobile learning, literature review

Procedia PDF Downloads 225

Authors: Florian Mach, Eric Hund, Malte Stonis

Abstract:

The dismantling of disused industrial facilities such as nuclear power plants or refineries is an enormous challenge for the planning and control of the logistic processes. Existing control models do not meet the requirements for a proper dismantling of industrial plants. Therefore, the paper presents an approach for the control of dismantling and post-processing processes (e.g. decontamination) in plant decommissioning. In contrast to existing approaches, the dismantling sequence and depth are selected depending on the capacity utilization of required post-processing processes by also considering individual characteristics of respective dismantling tasks (e.g. decontamination success rate, uncertainties regarding the process times). The results can be used in the dismantling of industrial plants (e.g. nuclear power plants) to reduce dismantling time and costs by avoiding bottlenecks such as capacity constraints.

Keywords: dismantling management, logistics planning and control models, nuclear power plant dismantling, reverse logistics

Procedia PDF Downloads 304

Authors: I. A. Simsek, S. N. Dogan, A. S. Kipcak, E. Morodor Derun, N. Tugrul

Abstract:

In this study, the drying characteristics of shrimp are studied by using the traditional drying method of oven. Drying temperatures are selected between 60-80°C. Obtained experimental drying results are applied to eleven mathematical models of Alibas, Aghbashlo et al., Henderson and Pabis, Jena and Das, Lewis, Logaritmic, Midilli and Kucuk, Page, Parabolic, Wang and Singh and Weibull. The best model was selected as parabolic based on the highest coefficient of determination (R²) (0.999990 at 80°C) and the lowest χ² (0.000002 at 80°C), and the lowest root mean square error (RMSE) (0.000976 at 80°C) values are compared to other models. The effective moisture diffusivity (Deff) values were calculated using the Fick’s second law’s cylindrical coordinate approximation and are found between 6.61×10⁻⁸ and 6.66×10⁻⁷ m²/s. The activation energy (Ea) was calculated using modified form of Arrhenius equation and is found as 18.315 kW/kg.

Keywords: activation energy, drying, effective moisture diffusivity, modelling, oven, shrimp

Procedia PDF Downloads 188