Search results for: predictive models

Authors: Yaowei Hu, Yongkai Wu, Lu Zhang, Xintao Wu

Abstract:

Causal inference from observational data is receiving wide applications in many fields. However, unidentifiable situations, where causal effects cannot be uniquely computed from observational data, pose critical barriers to applying causal inference to complicated real applications. In this paper, we develop a bounding method for estimating the average causal effect (ACE) under unidentifiable situations due to hidden confounders. We propose to parameterize the unknown exogenous random variables and structural equations of a causal model using neural networks and implicit generative models. Then, with an adversarial learning framework, we search the parameter space to explicitly traverse causal models that agree with the given observational distribution and find those that minimize or maximize the ACE to obtain its lower and upper bounds. The proposed method does not make any assumption about the data generating process and the type of the variables. Experiments using both synthetic and real-world datasets show the effectiveness of the method.

Keywords: average causal effect, hidden confounding, bound estimation, generative adversarial learning

Procedia PDF Downloads 182

Authors: Redouane Tlemsani, Redouane, Belkacem Kouninef, Abdelkader Benyettou

Abstract:

In this paper, our system is a Markovien system which we can see it like a Dynamic Bayesian Networks. One of the major interests of these systems resides in the complete training of the models (topology and parameters) starting from training data. The Bayesian Networks are representing models of dubious knowledge on complex phenomena. They are a union between the theory of probability and the graph theory in order to give effective tools to represent a joined probability distribution on a set of random variables. The representation of knowledge bases on description, by graphs, relations of causality existing between the variables defining the field of study. The theory of Dynamic Bayesian Networks is a generalization of the Bayesians networks to the dynamic processes. Our objective amounts finding the better structure which represents the relationships (dependencies) between the variables of a dynamic bayesian network. In applications in pattern recognition, one will carry out the fixing of the structure which obliges us to admit some strong assumptions (for example independence between some variables).

Keywords: Arabic on line character recognition, dynamic Bayesian network, pattern recognition, networks

Procedia PDF Downloads 608

Authors: Farid Gaci, Zoubir Nemouchi

Abstract:

A numerical study of laminar and turbulent fluid flows in 90° bend of square section was carried out. Three-dimensional meshes, based on hexahedral cells, were generated. The QUICK scheme was employed to discretize the convective term in the transport equations. The SIMPLE algorithm was adopted to treat the velocity-pressure coupling. The flow structure obtained showed interesting features such as recirculation zones and counter-rotating pairs of vortices. The performance of three different turbulence models was evaluated: the standard k- ω model, the SST k-ω model and the Reynolds Stress Model (RSM). Overall, it was found that, the multi-equation model performed better than the two equation models. In fact, the existence of four pairs of counter rotating cells, in the straight duct upstream of the bend, were predicted by the RSM closure but not by the standard eddy viscosity model nor the SST k-ω model. The analysis of the results led to a better understanding of the induced three dimensional secondary flows and the behavior of the local pressure coefficient and the friction coefficient.

Keywords: curved duct, counter-rotating cells, secondary flow, laminar, turbulent

Procedia PDF Downloads 333

Authors: Sung-Min Kim, Joon-Hong Park, Hyuk Choi

Abstract:

This study is focused on the comparative analysis and improvement to grasp the flow characteristic of the anti-splash device located under the P/V valve and new concept design models using the CFD. The P/V valve located upper deck to solve the pressure rising and vacuum condition of inner tank of the liquid cargo ships occurred oil outflow accident by transverse and longitudinal sloshing force. Anti-splash device is fitted to improve and prevent this problem in the shipbuilding industry, but the oil outflow accidents are still reported by ship owners. Thus, 4 types of new design model are presented by this study, and then comparative analysis is conducted with new models and existing model. Mostly the key criterion of this problem is flux in the outlet of the anti-splash device. Therefore, the flow and velocity are grasped by transient analysis, and then it decided optimum model and design parameters to develop model. Later, it is needed to develop an anti-splash device by flow test to get certification and verification using experiment equipments.

Keywords: anti-splash device, P/V valve, sloshing, CFD

Procedia PDF Downloads 631

Authors: Iris Xiaohong Quan

Abstract:

The global digital health market reached 175 billion U.S. dollars in 2019, and is expected to grow at about 25% CAGR to over 650 billion USD by 2025. Different terms such as digital health, e-health, mHealth, telehealth have been used in the field, which can sometimes cause confusion. The term digital health was originally introduced to refer specifically to the use of interactive media, tools, platforms, applications, and solutions that are connected to the Internet to address health concerns of providers as well as consumers. While mHealth emphasizes the use of mobile phones in healthcare, telehealth means using technology to remotely deliver clinical health services to patients. According to FDA, “the broad scope of digital health includes categories such as mobile health (mHealth), health information technology (IT), wearable devices, telehealth and telemedicine, and personalized medicine.” Some researchers believe that digital health is nothing else but the cultural transformation healthcare has been going through in the 21st century because of digital health technologies that provide data to both patients and medical professionals. As digital health is burgeoning, but research in the area is still inadequate, our paper aims to clear the definition confusion and provide an overall picture of digital health companies. We further investigate how business models are designed and differentiated in the emerging digital health sector. Both quantitative and qualitative methods are adopted in the research. For the quantitative analysis, our research data came from two databases Crunchbase and CBInsights, which are well-recognized information sources for researchers, entrepreneurs, managers, and investors. We searched a few keywords in the Crunchbase database based on companies’ self-description: digital health, e-health, and telehealth. A search of “digital health” returned 941 unique results, “e-health” returned 167 companies, while “telehealth” 427. We also searched the CBInsights database for similar information. After merging and removing duplicate ones and cleaning up the database, we came up with a list of 1464 companies as digital health companies. A qualitative method will be used to complement the quantitative analysis. We will do an in-depth case analysis of three successful unicorn digital health companies to understand how business models evolve and discuss the challenges faced in this sector. Our research returned some interesting findings. For instance, we found that 86% of the digital health startups were founded in the recent decade since 2010. 75% of the digital health companies have less than 50 employees, and almost 50% with less than 10 employees. This shows that digital health companies are relatively young and small in scale. On the business model analysis, while traditional healthcare businesses emphasize the so-called “3P”—patient, physicians, and payer, digital health companies extend to “5p” by adding patents, which is the result of technology requirements (such as the development of artificial intelligence models), and platform, which is an effective value creation approach to bring the stakeholders together. Our case analysis will detail the 5p framework and contribute to the extant knowledge on business models in the healthcare industry.

Keywords: digital health, business models, entrepreneurship opportunities, healthcare

Procedia PDF Downloads 178

Authors: Liane Marcia Freitas Silva, Fernando Augusto Silva Marins, Maria Silene Alexandre Leite

Abstract:

The adoption of a network structure, like in the supply chains, favors the increase of dependence between companies and, by consequence, their vulnerability. Environment disasters, sociopolitical and economical events, and the dynamics of supply chains elevate the uncertainty of their operation, favoring the occurrence of events that can generate break up in the operations and other undesired consequences. Thus, supply chains are exposed to various risks that can influence the profitability of companies involved, and there are several previous studies that have proposed risk classification models in order to categorize the risks and to manage them. The objective of this paper is to analyze and discuss thirty of these risk classification models by means a theoretical survey. The research method adopted for analyzing and discussion includes three phases: The identification of the types of risks proposed in each one of the thirty models, the grouping of them considering equivalent concepts associated to their definitions, and, the analysis of these risks groups, evaluating their similarities and differences. After these analyses, it was possible to conclude that, in fact, there is more than thirty risks types identified in the literature of Supply Chains, but some of them are identical despite of be used distinct terms to characterize them, because different criteria for risk classification are adopted by researchers. In short, it is observed that some types of risks are identified as risk source for supply chains, such as, demand risk, environmental risk and safety risk. On the other hand, other types of risks are identified by the consequences that they can generate for the supply chains, such as, the reputation risk, the asset depreciation risk and the competitive risk. These results are consequence of the disagreements between researchers on risk classification, mainly about what is risk event and about what is the consequence of risk occurrence. An additional study is in developing in order to clarify how the risks can be generated, and which are the characteristics of the components in a Supply Chain that leads to occurrence of risk.

Keywords: sisks classification, survey, supply chain management, theoretical discussion

Procedia PDF Downloads 625

Authors: Keyvan Ramin

Abstract:

The geometric nonlinearity of Off-Diagonal Bracing System (ODBS) could be a complementary system to covering and extending the nonlinearity of reinforced concrete material. Finite element modeling is performed for flexural frame, x-braced frame and the ODBS braced frame system at the initial phase. Then the different models are investigated along various analyses. According to the experimental results of flexural and x-braced frame, the verification is done. Analytical assessments are performed in according to three-dimensional finite element modeling. Non-linear static analysis is considered to obtain performance level and seismic behavior, and then the response modification factors calculated from each model’s pushover curve. In the next phase, the evaluation of cracks observed in the finite element models, especially for RC members of all three systems is performed. The finite element assessment is performed on engendered cracks in ODBS braced frame for various time steps. The nonlinear dynamic time history analysis accomplished in different stories models for three records of Elcentro, Naghan, and Tabas earthquake accelerograms. Dynamic analysis is performed after scaling accelerogram on each type of flexural frame, x-braced frame and ODBS braced frame one by one. The base-point on RC frame is considered to investigate proportional displacement under each record. Hysteresis curves are assessed along continuing this study. The equivalent viscous damping for ODBS system is estimated in according to references. Results in each section show the ODBS system has an acceptable seismic behavior and their conclusions have been converged when the ODBS system is utilized in reinforced concrete frame.

Keywords: FEM, seismic behaviour, pushover analysis, geometric nonlinearity, time history analysis, equivalent viscous damping, passive control, crack investigation, hysteresis curve

Procedia PDF Downloads 376

Authors: Delphine Morales, Florian Lombart, Agathe Truchot, Pauline Maire, Pascale Vigneron, Antoine Galmiche, Catherine Lok, Muriel Vayssade

Abstract:

Targeted therapy molecules are used as a first-line treatment for metastatic melanoma with B-Raf mutation. Nevertheless, these molecules can cause side effects to patients and are efficient on 50 to 60 % of them. Indeed, melanoma cell sensitivity to targeted therapy molecules is dependent on tumor microenvironment (cell-cell and cell-extracellular matrix interactions). To better unravel factors modulating cell sensitivity to B-Raf inhibitor, we have developed and compared several melanoma models: from metastatic melanoma cells cultured as monolayer (2D) to a co-culture in a 3D dermal equivalent. Cell response was studied in different melanoma cell lines such as SK-MEL-28 (mutant B-Raf (V600E), sensitive to Vemurafenib), SK-MEL-3 (mutant B-Raf (V600E), resistant to Vemurafenib) and a primary culture of dermal human fibroblasts (HDFn). Assays have initially been performed in a monolayer cell culture (2D), then a second time on a 3D dermal equivalent (dermal human fibroblasts embedded in a collagen gel). All cell lines were treated with Vemurafenib (a B-Raf inhibitor) for 48 hours at various concentrations. Cell sensitivity to treatment was assessed under various aspects: Cell proliferation (cell counting, EdU incorporation, MTS assay), MAPK signaling pathway analysis (Western-Blotting), Apoptosis (TUNEL), Cytokine release (IL-6, IL-1α, HGF, TGF-β, TNF-α) upon Vemurafenib treatment (ELISA) and histology for 3D models. In 2D configuration, the inhibitory effect of Vemurafenib on cell proliferation was confirmed on SK-MEL-28 cells (IC50=0.5 µM), and not on the SK-MEL-3 cell line. No apoptotic signal was detected in SK-MEL-28-treated cells, suggesting a cytostatic effect of the Vemurafenib rather than a cytotoxic one. The inhibition of SK-MEL-28 cell proliferation upon treatment was correlated with a strong expression decrease of phosphorylated proteins involved in the MAPK pathway (ERK, MEK, and AKT/PKB). Vemurafenib (from 5 µM to 10 µM) also slowed down HDFn proliferation, whatever cell culture configuration (monolayer or 3D dermal equivalent). SK-MEL-28 cells cultured in the dermal equivalent were still sensitive to high Vemurafenib concentrations. To better characterize all cell population impacts (melanoma cells, dermal fibroblasts) on Vemurafenib efficacy, cytokine release is being studied in 2D and 3D models. We have successfully developed and validated a relevant 3D model, mimicking cutaneous metastatic melanoma and tumor microenvironment. This 3D melanoma model will become more complex by adding a third cell population, keratinocytes, allowing us to characterize the epidermis influence on the melanoma cell sensitivity to Vemurafenib. In the long run, the establishment of more relevant 3D melanoma models with patients’ cells might be useful for personalized therapy development. The authors would like to thank the Picardie region and the European Regional Development Fund (ERDF) 2014/2020 for the funding of this work and Oise committee of "La ligue contre le cancer".

Keywords: 3D human skin model, melanoma, tissue engineering, vemurafenib efficiency

Procedia PDF Downloads 298

Authors: Bronwen Wade

Abstract:

Over the last half-century, child welfare systems have increasingly relied on quantitative risk assessment tools, such as actuarial or predictive risk tools. These tools are developed by performing statistical analysis of how attributes captured in administrative data are related to future child maltreatment. Some scholars argue that attributes in administrative data can serve as proxies for race and that quantitative risk assessment tools reify racial bias in decision-making. Others argue that these tools provide more “objective” and “scientific” guides for decision-making instead of subjective social worker judgment. This study performs a systematic review of the literature on the impact of quantitative risk assessment tools on racial disproportionality; it examines methodological biases in work on this topic, summarizes key findings, and provides suggestions for further work. A search of CINAHL, PsychInfo, Proquest Social Science Premium Collection, and the ProQuest Dissertations and Theses Collection was performed. Academic and grey literature were included. The review includes studies that use quasi-experimental methods and development, validation, or re-validation studies of quantitative risk assessment tools. PROBAST (Prediction model Risk of Bias Assessment Tool) and CHARMS (CHecklist for critical Appraisal and data extraction for systematic Reviews of prediction Modelling Studies) were used to assess the risk of bias and guide data extraction for risk development, validation, or re-validation studies. ROBINS-I (Risk of Bias in Non-Randomized Studies of Interventions) was used to assess for bias and guide data extraction for the quasi-experimental studies identified. Due to heterogeneity among papers, a meta-analysis was not feasible, and a narrative synthesis was conducted. 11 papers met the eligibility criteria, and each has an overall high risk of bias based on the PROBAST and ROBINS-I assessments. This is deeply concerning, as major policy decisions have been made based on a limited number of studies with a high risk of bias. The findings on racial disproportionality have been mixed and depend on the tool and approach used. Authors use various definitions for racial equity, fairness, or disproportionality. These concepts of statistical fairness are connected to theories about the reason for racial disproportionality in child welfare or social definitions of fairness that are usually not stated explicitly. Most findings from these studies are unreliable, given the high degree of bias. However, some of the less biased measures within studies suggest that quantitative risk assessment tools may worsen racial disproportionality, depending on how disproportionality is mathematically defined. Authors vary widely in their approach to defining and addressing racial disproportionality within studies, making it difficult to generalize findings or approaches across studies. This review demonstrates the power of authors to shape policy or discourse around racial justice based on their choice of statistical methods; it also demonstrates the need for improved rigor and transparency in studies of quantitative risk assessment tools. Finally, this review raises concerns about the impact that these tools have on child welfare systems and racial disproportionality.

Keywords: actuarial risk, child welfare, predictive risk, racial disproportionality

Procedia PDF Downloads 46

Authors: Shatha Khuzaee

Abstract:

The study develops multimodal stylistic approach to analyse a number of BBC online news articles reporting some key events from the so called ‘Arab Uprisings’. Critical stylistics (CS) and visual grammar (VG) provide insightful arguments to the ways ideology is projected through different verbal and visual modes, yet they are mode specific because they examine how each mode projects its meaning separately and do not attempt to clarify what happens intersemiotically when the two modes co-occur. Therefore, it is the task undertaken in this research to propose multimodal stylistic approach that addresses the issue of ideology construction when the two modes co-occur. Informed by functional grammar and social semiotics, the analysis attempts to integrate three linguistic models developed in critical stylistics, namely, transitivity choices, prioritizing and hypothesizing along with their visual equivalents adopted from visual grammar to investigate the way ideology is constructed, in multimodal text, when text/image participate and interrelate in the process of meaning making on the textual level of analysis. The analysis provides comprehensive theoretical and analytical elaborations on the different points of integration between CS linguistic models and VG equivalents which operate on the textual level of analysis to better account for ideology construction in news as non-literary multimodal texts. It is argued that the analysis well thought out a plan that would remark the first step towards the integration between the well-established linguistic models of critical stylistics and that of visual analysis to analyse multimodal texts on the textual level. Both approaches are compatible to produce multimodal stylistic approach because they intend to analyse text and image depending on whatever textual evidence is available. This supports the analysis maintain the rigor and replicability needed for a stylistic analysis like the one undertaken in this study.

Keywords: multimodality, stylistics, visual grammar, social semiotics, functional grammar

Procedia PDF Downloads 217

Authors: Venkat S. Somayajula

Abstract:

Ground surface temperature history prediction model plays a vital role in determining standards for international nuclear waste management. International standards for borehole based nuclear waste disposal require paleoclimate cycle predictions on scale of a million forward years for the place of waste disposal. This research focuses on developing a paleoclimate cycle prediction model using Bayesian long-short term memory (LSTM) neural architecture operated on accumulated borehole temperature history data. Bayesian models have been previously used for paleoclimate cycle prediction based on Monte-Carlo weight method, but due to limitations pertaining model coupling with certain other prediction networks, Bayesian models in past couldn’t accommodate prediction cycle’s over 1000 years. LSTM has provided frontier to couple developed models with other prediction networks with ease. Paleoclimate cycle developed using this process will be trained on existing borehole data and then will be coupled to surface temperature history prediction networks which give endpoints for backpropagation of LSTM network and optimize the cycle of prediction for larger prediction time scales. Trained LSTM will be tested on past data for validation and then propagated for forward prediction of temperatures at borehole locations. This research will be beneficial for study pertaining to nuclear waste management, anthropological cycle predictions and geophysical features

Keywords: Bayesian long-short term memory neural network, borehole temperature, ground surface temperature history, paleoclimate cycle

Procedia PDF Downloads 124

Authors: Xiaoci Li, Yonghua Huang, Hui Lin

Abstract:

Optimization of the concentration of components in mixed refrigerants leads to potential improvement of either thermodynamic cycle performance or safety performance of heat pumps and refrigerators. R32+R1234yf and R125+R290 are two promising binary mixed refrigerants for the application of heat pumps working in the cold areas. The p-ρ-T data of these mixtures are one of the fundamental and necessary properties for design and evaluation of the performance of the heat pumps. Although the property data of mixtures can be predicted by the mixing models based on the pure substances incorporated in programs such as the NIST database Refprop, direct property measurement will still be helpful to reveal the true state behaviors and verify the models. Densities of the mixtures of R32+R1234yf an d R125+R290 are measured by an Anton Paar U shape oscillating tube digital densimeter DMA-4500 in the range of temperatures from 0°C to 100 °C and pressures up to 10 MPa. The accuracy of the measurement reaches 0.00005 g/cm³. The experimental data are compared with the predictions by Refprop in the corresponding range of pressure and temperature.

Keywords: mixed refrigerant, density measurement, densimeter, thermodynamic property

Procedia PDF Downloads 290

Authors: Kefaifi Aissa, Sahraoui Tahar, Kheloufi Abdelkrim, Anas Sabiha, Hannane Farouk

Abstract:

The aim of this study is to show how the Design of Experiments Method (DOE) can be put into use as a practical approach for silica sand etching behavior modeling during its primary step of leaching. In the present work, we have studied etching effect on particle size during a primary step of leaching process on Algerian silica sand with florid acid (HF) at 20% and 30 % during 4 and 8 hours. Therefore, a new purity of the sand is noted depending on the time of leaching. This study was expanded by a numerical approach using a method of experiment design, which shows the influence of each parameter and the interaction between them in the process and approved the obtained experimental results. This model is a predictive approach using hide software. Based on the measured parameters experimentally in the interior of the model, the use of DOE method can make it possible to predict the outside parameters of the model in question and can give us the optimize response without making the experimental measurement.

Keywords: acid leaching, design of experiments method(DOE), purity silica, silica etching

Procedia PDF Downloads 278

Authors: Queen Suraajini Rajendran, Sai Hung Cheung

Abstract:

Statistical down scaling models are required to connect the global climate model outputs and the local weather variables for climate change impact prediction. For reliable climate change impact studies, the uncertainty associated with the model including natural variability, uncertainty in the climate model(s), down scaling model, model inadequacy and in the predicted results should be quantified appropriately. In this work, a new approach is developed by the authors for statistical classification, statistical down scaling and uncertainty assessment and is applied to Singapore rainfall. It is a robust Bayesian uncertainty analysis methodology and tools based on coupling dependent modeling error with classification and statistical down scaling models in a way that the dependency among modeling errors will impact the results of both classification and statistical down scaling model calibration and uncertainty analysis for future prediction. Singapore data are considered here and the uncertainty and prediction results are obtained. From the results obtained, directions of research for improvement are briefly presented.

Keywords: statistical downscaling, global climate model, climate change, uncertainty

Procedia PDF Downloads 363

Authors: Saeedullah J. Mandokhail, Shamsher Sadiq, Meer H. Khan

Abstract:

This paper presents the comparison of excess pore water pressure ratio (ru) predicted by using accumulated stress based pore pressure model and plasticity model. One dimensional effective stress site response analyses were performed on a 30 m deep sand column (consists of a liquefiable layer in between non-liquefiable layers) using accumulated stress based pore pressure model in Deepsoil and PDMY2 (PressureDependentMultiYield02) model in Opensees. Three Input motions with different peak ground acceleration (PGA) levels of 0.357 g, 0.124 g, and 0.11 g were used in this study. The developed excess pore pressure ratio predicted by the above two models were compared and analyzed along the depth. The time history of the ru at mid of the liquefiable layer and non-liquefiable layer were also compared. The comparisons show that the two models predict mostly similar ru values. The predicted ru is also consistent with the PGA level of the input motions.

Keywords: effective stress, excess pore pressure ratio, pore pressure model, site response analysis

Procedia PDF Downloads 221

Authors: Rosa Tsegaye Aga, Xuan Jiang, Pavel Vazquez Faci, Siqing Liu, Simon Rayner, Endalkachew Alemu, Markos Abebe

Abstract:

Tuberculosis (TB) is a major cause of disease globally. In most cases, TB is treatable and curable, but only with the proper treatment. There is a time when drug-resistant TB occurs when bacteria become resistant to the drugs that are used to treat TB. Current strategies to identify drug-resistant TB bacteria are laboratory-based, and it takes a longer time to identify the drug-resistant bacteria and treat the patient accordingly. But machine learning (ML) and data science approaches can offer new approaches to the problem. In this study, we propose to develop an ML-based model to predict the antibiotic resistance phenotypes of TB isolates in minutes and give the right treatment to the patient immediately. The study has been using the whole genome sequence (WGS) of TB isolates as training data that have been extracted from the NCBI repository and contain different countries’ samples to build the ML models. The reason that different countries’ samples have been included is to generalize the large group of TB isolates from different regions in the world. This supports the model to train different behaviors of the TB bacteria and makes the model robust. The model training has been considering three pieces of information that have been extracted from the WGS data to train the model. These are all variants that have been found within the candidate genes (F1), predetermined resistance-associated variants (F2), and only resistance-associated gene information for the particular drug. Two major datasets have been constructed using these three information. F1 and F2 information have been considered as two independent datasets, and the third information is used as a class to label the two datasets. Five machine learning algorithms have been considered to train the model. These are Support Vector Machine (SVM), Random forest (RF), Logistic regression (LR), Gradient Boosting, and Ada boost algorithms. The models have been trained on the datasets F1, F2, and F1F2 that is the F1 and the F2 dataset merged. Additionally, an ensemble approach has been used to train the model. The ensemble approach has been considered to run F1 and F2 datasets on gradient boosting algorithm and use the output as one dataset that is called F1F2 ensemble dataset and train a model using this dataset on the five algorithms. As the experiment shows, the ensemble approach model that has been trained on the Gradient Boosting algorithm outperformed the rest of the models. In conclusion, this study suggests the ensemble approach, that is, the RF + Gradient boosting model, to predict the antibiotic resistance phenotypes of TB isolates by outperforming the rest of the models.

Keywords: machine learning, MTB, WGS, drug resistant TB

Procedia PDF Downloads 42

Authors: Zhifeng Kong

Abstract:

Over-parameterized neural networks have attracted a great deal of attention in recent deep learning theory research, as they challenge the classic perspective of over-fitting when the model has excessive parameters and have gained empirical success in various settings. While a number of theoretical works have been presented to demystify properties of such models, the convergence properties of such models are still far from being thoroughly understood. In this work, we study the convergence properties of training two-hidden-layer partially over-parameterized fully connected networks with the Rectified Linear Unit activation via gradient descent. To our knowledge, this is the first theoretical work to understand convergence properties of deep over-parameterized networks without the equally-wide-hidden-layer assumption and other unrealistic assumptions. We provide a probabilistic lower bound of the widths of hidden layers and proved linear convergence rate of gradient descent. We also conducted experiments on synthetic and real-world datasets to validate our theory.

Keywords: over-parameterization, rectified linear units ReLU, convergence, gradient descent, neural networks

Procedia PDF Downloads 138

Authors: Merve Aygün Esastürk, Deren Ataç Yılmaz, Görkem Oğur, Emre Özgen Kuzu, Sadık Ödemiş

Abstract:

Hydrocracker plants are one of the most complicated and most profitable units in the refinery process. It takes long chain paraffinic hydrocarbons as feed and turns them into smaller and more valuable products, mainly kerosene and diesel under high pressure with the excess amount of hydrogen. Controlling the product qualities well directly contributes to the unit profit. Control of a plant is mainly based on PID and MPC controllers. Controlling the reaction section is important in terms of reaction severity. However, controlling the fractionation section is more crucial since the end products are separated in fractionation section. In this paper, the importance of well-configured base layer control mechanism, composed of PID controllers, is highlighted. For this purpose, two different base layer control scheme is applied in a hydrocracker fractionator column performances of schemes, which is a direct contribution to better product quality, are compared.

Keywords: controller, distillation, configuration selection, hydrocracker, model predictive controller, proportional-integral-derivative controller

Procedia PDF Downloads 436

Authors: Zeinab Yazdanfar, Dilan Robert, Daniel Lester, S. Setunge

Abstract:

Scour has been identified as the most common threat to bridge stability worldwide. Traditionally, scour around bridge piers is calculated using the empirical approaches that have considerable limitations and are difficult to generalize. The multi-physic nature of scouring which involves turbulent flow, soil mechanics and solid-fluid interactions cannot be captured by simple empirical equations developed based on limited laboratory data. These limitations can be overcome by direct numerical modeling of coupled hydro-mechanical scour process that provides a robust prediction of bridge scour and valuable insights into the scour process. Several numerical models have been proposed in the literature for bridge scour estimation including Eulerian flow models and coupled Euler-Lagrange models incorporating an empirical sediment transport description. However, the contact forces between particles and the flow-particle interaction haven’t been taken into consideration. Incorporating collisional and frictional forces between soil particles as well as the effect of flow-driven forces on particles will facilitate accurate modeling of the complex nature of scour. In this study, a coupled Computational Fluid Dynamics and Discrete Element Model (CFD-DEM) has been developed to simulate the scour process that directly models the hydro-mechanical interactions between the sediment particles and the flowing water. This approach obviates the need for an empirical description as the fundamental fluid-particle, and particle-particle interactions are fully resolved. The sediment bed is simulated as a dense pack of particles and the frictional and collisional forces between particles are calculated, whilst the turbulent fluid flow is modeled using a Reynolds Averaged Navier Stocks (RANS) approach. The CFD-DEM model is validated against experimental data in order to assess the reliability of the CFD-DEM model. The modeling results reveal the criticality of particle impact on the assessment of scour depth which, to the authors’ best knowledge, hasn’t been considered in previous studies. The results of this study open new perspectives to the scour depth and time assessment which is the key to manage the failure risk of bridge infrastructures.

Keywords: bridge scour, discrete element method, CFD-DEM model, multi-phase model

Procedia PDF Downloads 127

Authors: Yuan Long, Fabrizio Ceschin, David Harrison

Abstract:

Environmental sustainability is under the threat of excessive single-use plastic packaging waste, and current waste management fails to address this issue. Therefore, it has led to a reidentification of the alternative, which can curb the packaging waste without reducing social needs. Reusable packaging represents a circular approach to close the loop of consumption in which packaging can stay longer in the system to satisfy social needs. However, the implementation of reusable packaging is fragmented and lacks systematic approaches. The product-service system (PSS) is widely regarded as a sustainable business model innovation for embracing circular consumption. As a result, applying PSS to reusable packaging solutions will be promising to address the packaging waste issue. This paper aims at filling the knowledge gap relating to apply PSS to reusable packaging solutions and provide a strategic design tool that could support packaging professionals to design reusable packaging solutions. The methodology of this paper is case studies and workshops to provide a design tool. The respondents are packaging professionals who are packaging consultants, NGO professionals, and entrepreneurs. 57 cases collected show that 15 archetypal models operate in the market. Subsequently, a polarity diagram is developed to embrace those 15 archetypal models, and a total number of 24 experts were invited for the workshop to evaluate the design tool. This research finally provides a strategic design tool to support packaging professionals to design reusable packaging solutions. The application of the tool is to support the understanding of the reusable packaging solutions, analyzing the markets, identifying new opportunities, and generate new business models. The implication of this research is to provide insights for academics and businesses in terms of tackling single-use packaging waste and build a foundation for further development of the reusable packaging solution tool.

Keywords: environmental sustainability, product-service system, reusable packaging, design tool

Procedia PDF Downloads 143

Authors: Grmanesh Gidey Kahsay

Abstract:

In modern technical engineering applications, quality is defined in two ways. The first one is that quality is the parameter that measures a product or service’s characteristics to meet and satisfy the pre-stated or fundamental needs (reliability, durability, serviceability). The second one is the quality of a product or service free of any defect or deficiencies. The American Society for Quality (ASQ) describes quality as a pursuit of optimal solutions to confirm successes and fulfillment to be accountable for the product or service's requirements and expectations. This article focuses on quality engineering tools in modern industrial applications. Quality engineering is a field of engineering that deals with the principles, techniques, models, and applications of the product or service to guarantee quality. Including the entire activities to analyze the product’s design and development, quality engineering emphasizes how to make sure that products and services are designed and developed to meet consumers’ requirements. This episode acquaints with quality tools such as quality systems, auditing, product design, and process control. The finding presents thoughts that aim to improve quality engineering proficiency and effectiveness by introducing essential quality techniques and tools in some selected industries.

Keywords: essential quality tools, quality systems and models, quality management systems, and quality assurance

Procedia PDF Downloads 146

Authors: S. Khamtree, T. Ratanawilai, C. Nuntadusit

Abstract:

Rubberwood is a crucial commercial timber in Southern Thailand. All processes in a rubberwood production depend on the knowledge and expertise of the technicians, especially the drying process. This research aims to develop an empirical model for drying kinetics in rubberwood. During the experiment, the temperature of the hot air and the average air flow velocity were kept at 80-100 °C and 1.75 m/s, respectively. The moisture content in the samples was determined less than 12% in the achievement of drying basis. The drying kinetic was simulated using an empirical solver. The experimental results illustrated that the moisture content was reduced whereas the drying temperature and time were increased. The coefficient of the moisture ratio between the empirical and the experimental model was tested with three statistical parameters, R-square (R²), Root Mean Square Error (RMSE) and Chi-square (χ²) to predict the accuracy of the parameters. The experimental moisture ratio had a good fit with the empirical model. Additionally, the results indicated that the drying of rubberwood using the Henderson and Pabis model revealed the suitable level of agreement. The result presented an excellent estimation (R² = 0.9963) for the moisture movement compared to the other models. Therefore, the empirical results were valid and can be implemented in the future experiments.

Keywords: empirical models, rubberwood, moisture ratio, hot air drying

Procedia PDF Downloads 259

Authors: A. Hol

Abstract:

21^st century brought waves of business and industry eTransformations. The impact of change is also being seen in education. To identify the extent of this, scenario analysis methodology was utilised with the aim to assess business transformations across industry sectors ranging from craftsmanship, medicine, finance and manufacture to innovations and adoptions of new technologies and business models. Firstly, scenarios were drafted based on the current eTransformation models and its dimensions. Following this, eTransformation framework was utilised with the aim to derive the key eTransformation parameters, the essential characteristics that have enabled eTransformations across the sectors. Following this, identified key parameters were mapped to the transforming domain-education. The mapping assisted in deriving a cognitive eTransformation framework for education sector. The framework highlights the importance of context and the notion that education today needs not only to deliver content to students but it also needs to be able to meet the dynamically changing demands of specific student and industry groups. Furthermore, it pinpoints that for such processes to be supported, specific technology is required, so that instant, on demand and periodic feedback as well as flexible, dynamically expanding study content can be sought and received via multiple education mediums.

Keywords: education sector, business transformation, eTransformation model, cognitive model, cognitive systems, eTransformation

Procedia PDF Downloads 132

Authors: Oladapo Tolulope Ibitoye

Abstract:

Neural networks have become prominent and widely engaged in algorithmic-based machine learning networks. They are perfect in solving day-to-day issues to a certain extent. Neural networks are computing systems with several interconnected nodes. One of the numerous areas of application of neural networks is object detection. This is a prominent area due to the coronavirus disease pandemic and the post-pandemic phases. Wearing a face mask in public slows the spread of the virus, according to experts’ submission. This calls for the development of a reliable and effective model for detecting face masks on people's faces during compliance checks. The existing neural network models for facemask detection are characterized by their black-box nature and large dataset requirement. The highlighted challenges have compromised the performance of the existing models. The proposed model utilized Faster R-CNN Model on Inception V3 backbone to reduce system complexity and dataset requirement. The model was trained and validated with very few datasets and evaluation results shows an overall accuracy of 96% regardless of skin tone.

Keywords: convolutional neural network, face detection, face mask, masked faces

Procedia PDF Downloads 59

Authors: Susan Diamond

Abstract:

Machine Learning workloads have traditionally been run in high-performance computing (HPC) environments, where users log in to dedicated machines and utilize the attached GPUs to run training jobs on huge datasets. Training of large neural network models is very resource intensive, and even after exploiting parallelism and accelerators such as GPUs, a single training job can still take days. Consequently, the cost of hardware is a barrier to entry. Even when upfront cost is not a concern, the lead time to set up such an HPC environment takes months from acquiring hardware to set up the hardware with the right set of firmware, software installed and configured. Furthermore, scalability is hard to achieve in a rigid traditional lab environment. Therefore, it is slow to react to the dynamic change in the artificial intelligent industry. Watson Deep Learning as a service, a cloud-based deep learning platform that mitigates the long lead time and high upfront investment in hardware. It enables robust and scalable sharing of resources among the teams in an organization. It is designed for on-demand cloud environments. Providing a similar user experience in a multi-tenant cloud environment comes with its own unique challenges regarding fault tolerance, performance, and security. Watson Deep Learning as a service tackles these challenges and present a deep learning stack for the cloud environments in a secure, scalable and fault-tolerant manner. It supports a wide range of deep-learning frameworks such as Tensorflow, PyTorch, Caffe, Torch, Theano, and MXNet etc. These frameworks reduce the effort and skillset required to design, train, and use deep learning models. Deep Learning as a service is used at IBM by AI researchers in areas including machine translation, computer vision, and healthcare. 

Keywords: deep learning, machine learning, cognitive computing, model training

Procedia PDF Downloads 203

Authors: Sedat Yayla, Mehmet Oruc, Shakhwan Yaseen

Abstract:

Cavitation phenomena might rigorously impair machine parts such as pumps, propellers and impellers or devices as the pressure in the fluid declines under the liquid's saturation pressure. To evaluate the influence of cavitation, in this research two-dimensional computational fluid dynamics (CFD) venturi models with variety of inlet pressure values, throat lengths and vapor fluid contents were applied. In this research three different vapor contents (0%, 5% 10%), four inlet pressures (2, 4, 6, 8 and 10 atm) and two venturi models were employed at different throat lengths ( 5, 10, 15 and 20 mm) for discovering the impact of each parameter on the cavitation number. It is uncovered that there is a positive correlation between pressure inlet and vapor fluid content and cavitation number. Furthermore, it is unveiled that velocity remains almost constant at the inlet pressures of 6, 8,10atm, nevertheless increasing the length of throat results in the substantial escalation in the velocity of the throat at inlet pressures of 2 and 4 atm. Furthermore, velocity and cavitation number were negatively correlated. The results of the cavitation number varied between 0.092 and 0.495 depending upon the velocity values of the throat.

Keywords: cavitation number, computational fluid dynamics, mixture of fluid, two-phase flow, velocity of throat

Procedia PDF Downloads 390

Authors: V. Keim, J. Spachtholz, J. Hammer

Abstract:

The importance of fibre reinforced plastics continually increases due to the excellent mechanical properties, low material and manufacturing costs combined with significant weight reduction. Today, components are usually designed and calculated numerically by using finite element methods (FEM) to avoid expensive laboratory tests. These programs are based on material models including material specific deformation characteristics. In this research project, material models for short glass fibre reinforced plastics are presented to simulate the visco-elasto-plastic deformation behaviour. Prior to modelling specimens of the material EMS Grivory HTV-5H1, consisting of a Polyphthalamide matrix reinforced by 50wt.-% of short glass fibres, are characterized experimentally in terms of the highly time dependent deformation behaviour of the matrix material. To minimize the experimental effort, the cyclic deformation behaviour under tensile and compressive loading (R = −1) is characterized by isothermal complex low cycle fatigue (CLCF) tests. Combining cycles under two strain amplitudes and strain rates within three orders of magnitude and relaxation intervals into one experiment the visco-elastic deformation is characterized. To identify visco-plastic deformation monotonous tensile tests either displacement controlled or strain controlled (CERT) are compared. All relevant modelling parameters for this complex superposition of simultaneously varying mechanical loadings are quantified by these experiments. Subsequently, two different material models are compared with respect to their accuracy describing the visco-elasto-plastic deformation behaviour. First, based on Chaboche an extended 12 parameter model (EVP-KV2) is used to model cyclic visco-elasto-plasticity at two time scales. The parameters of the model including a total separation of elastic and plastic deformation are obtained by computational optimization using an evolutionary algorithm based on a fitness function called genetic algorithm. Second, the 12 parameter visco-elasto-plastic material model by Launay is used. In detail, the model contains a different type of a flow function based on the definition of the visco-plastic deformation as a part of the overall deformation. The accuracy of the models is verified by corresponding experimental LCF testing.

Keywords: complex low cycle fatigue, material modelling, short glass fibre reinforced polyphthalamides, visco-elasto-plastic deformation

Procedia PDF Downloads 211

Authors: Jose Ramon Calvo-Ferrer

Abstract:

Video games have been extensively employed in educational contexts to teach contents and skills, upon the premise that they engage students and provide instant feedback, which makes them adequate tools in the field of education and training. Term frequency, along with metacognition and implicit corrective feedback, has often been identified as powerful variables in the learning of vocabulary in a foreign language. This study analyses the learning of L2 mobile operating system terminology by a group of students and uses the data collected by the video game The Conference Interpreter to identify the predictive strength of term frequency (times a term is shown), positive metacognition (times a right answer is provided), and negative metacognition (times a term is shown as wrong) regarding L2 vocabulary learning and perceived learning outcomes. The regression analysis shows that the factor ‘positive metacognition’ is a positive predictor of both dependent variables, whereas the other factors seem to have no statistical effect on any of them.

Keywords: digital game-based learning, feedback, metacognition, frequency, video games

Procedia PDF Downloads 149

Authors: Wanhyun Cho, Soonja Kang, Sanggoon Kim, Soonyoung Park

Abstract:

We present probabilistic multinomial Dirichlet classification model for multidimensional data and Gaussian process priors. Here, we have considered an efficient computational method that can be used to obtain the approximate posteriors for latent variables and parameters needed to define the multiclass Gaussian process classification model. We first investigated the process of inducing a posterior distribution for various parameters and latent function by using the variational Bayesian approximations and important sampling method, and next we derived a predictive distribution of latent function needed to classify new samples. The proposed model is applied to classify the synthetic multivariate dataset in order to verify the performance of our model. Experiment result shows that our model is more accurate than the other approximation methods.

Keywords: multinomial dirichlet classification model, Gaussian process priors, variational Bayesian approximation, importance sampling, approximate posterior distribution, marginal likelihood evidence

Procedia PDF Downloads 435

Authors: John Adebayo Oloyhede

Abstract:

This paper examines effect of financial liberalization on the stability of the demand for money function and its implication for exchange rate behaviour of three African countries. As the demand for money function is regarded as one of the two main building blocks of most exchange rate determination models, the other being purchasing power parity, its stability is required for the monetary models of exchange rate determination to hold. To what extent has the liberalisation policy of these countries, for instance liberalised interest rate, affected the demand for money function and what has been the consequence on the validity and relevance of floating exchange rate models? The study adopts the Autoregressive Instrumental Package (AIV) of multiple regression technique and followed the Almon Polynomial procedure with zero-end constraint. Data for the period 1986 to 2011 were drawn from three developing countries of Africa, namely: Gambia, Ghana and Nigeria, which did not only start the liberalization and floating system almost at the same period but share similar and diverse economic and financial structures. Its findings show that the demand for money was a stable function of income and interest rate at home and abroad. Other factors such as exchange rate and foreign interest rate exerted some significant effect on domestic money demand. The short-run and long-run elasticity with respect to income, interest rates, expected inflation rate and exchange rate expectation are not greater than zero. This evidence conforms to some extent to the expected behaviour of the domestic money function and underscores its ability to serve as good building block or assumption of the monetary model of exchange rate determination. This will, therefore, assist appropriate monetary authorities in the design and implementation of further financial liberalization policy packages in developing countries.

Keywords: financial liberalisation, exchange rates, demand for money, developing economies

Procedia PDF Downloads 366