Search results for: fused deep representations

Authors: G. K. Awari, Vishwajeet V. Ambade, S. W. Rajurkar

Abstract:

The field of additive manufacturing is growing, and discoveries are being made. 3D printing machines are also being developed to accommodate a wider range of 3D printing materials, including plastics, metals (metal AM powders), composites, filaments, and other materials. There are numerous printing materials available for industrial additive manufacturing. Such materials have their unique characteristics, advantages, and disadvantages. In order to avoid errors in additive manufacturing, key elements such as 3D printing material type, texture, cost, printing technique and procedure, and so on must be examined. It can be complex to select the best material for a particular job. Polylactic acid (PLA) is made from sugar cane or cornstarch, both of which are renewable resources. "Black plastic" is another name for it. Because it is safe to use and print, it is frequently used in primary and secondary schools. This is also how FDM screen printing is done. PLA is simple to print because of its low warping impact. It's also possible to print it on a cold surface. When opposed to ABS, it allows for sharper edges and features to be printed. This material comes in a wide range of colours. Polylactic acid (PLA) is the most common material used in fused deposition modelling (FDM). PLA can be used to print a wide range of components, including medical implants, household items, and mechanical parts. The mechanical behaviour of the printed item is affected by variations in infill patterns that are subjected to compressive tests in the current investigation to examine their behaviour under compressive stresses.

Keywords: fused deposition modelling, polylactic acid, infill density, infill pattern, compressive strength

Procedia PDF Downloads 60

Authors: Hongmei Wang, Ziyun Xiang, Ying liu, Li Yu, Dongsheng Yue

Abstract:

Intro: The COVID-19 (Corona Virus Disease 2019) has greatly changed the global economic, political and financial ecology. The mutation of the coronavirus in the UK in December 2020 has brought new panic to the world. Deep learning was performed on Chest Computed tomography (CT) of COVID-19 and Influenza and describes their characteristics. The predominant features of COVID-19 pneumonia was ground-glass opacification, followed by consolidation. Lesion density: most lesions appear as ground-glass shadows, and some lesions coexist with solid lesions. Lesion distribution: the focus is mainly on the dorsal side of the periphery of the lung, with the lower lobe of the lungs as the focus, and it is often close to the pleura. Other features it has are grid-like shadows in ground glass lesions, thickening signs of diseased vessels, air bronchi signs and halo signs. The severe disease involves whole bilateral lungs, showing white lung signs, air bronchograms can be seen, and there can be a small amount of pleural effusion in the bilateral chest cavity. At the same time, this year's flu season could be near its peak after surging throughout the United States for months. Chest CT for Influenza infection is characterized by focal ground glass shadows in the lungs, with or without patchy consolidation, and bronchiole air bronchograms are visible in the concentration. There are patchy ground-glass shadows, consolidation, air bronchus signs, mosaic lung perfusion, etc. The lesions are mostly fused, which is prominent near the hilar and two lungs. Grid-like shadows and small patchy ground-glass shadows are visible. Deep neural networks have great potential in image analysis and diagnosis that traditional machine learning algorithms do not. Method: Aiming at the two major infectious diseases COVID-19 and influenza, which are currently circulating in the world, the chest CT of patients with two infectious diseases is classified and diagnosed using deep learning algorithms. The residual network is proposed to solve the problem of network degradation when there are too many hidden layers in a deep neural network (DNN). The proposed deep residual system (ResNet) is a milestone in the history of the Convolutional neural network (CNN) images, which solves the problem of difficult training of deep CNN models. Many visual tasks can get excellent results through fine-tuning ResNet. The pre-trained convolutional neural network ResNet is introduced as a feature extractor, eliminating the need to design complex models and time-consuming training. Fastai is based on Pytorch, packaging best practices for in-depth learning strategies, and finding the best way to handle diagnoses issues. Based on the one-cycle approach of the Fastai algorithm, the classification diagnosis of lung CT for two infectious diseases is realized, and a higher recognition rate is obtained. Results: A deep learning model was developed to efficiently identify the differences between COVID-19 and influenza using chest CT.

Keywords: COVID-19, Fastai, influenza, transfer network

Procedia PDF Downloads 130

Authors: Alejandra Sierra, Maria de los Angeles Covarrubias, Guillermo Julian Gonzalez, Noe Alfaro

Abstract:

The objective of this study was to identify the cultural dimensions of the terms unplanned pregnancy and unwanted pregnancy built by adolescent women, through the focus of the social representations. Two associative methods were used: free listings and the paired comparison. 72 female students between the ages of 15 and 19 were interviewed, from the downtown area of Leon Guanajuato, Mexico. Words related to inducer terms were classified into five thematic categories: facilitators, consequences, reactions, expectations, and lexicon. The results showed that the social representations of unplanned pregnancy highlighted elements related to economic difficulties and negative emotional aspects, while unwanted pregnancy was associated with negative emotional aspects such as anger, anxiety, and sadness. The meanings each person attributes to terms related to pregnancy are culturally constructed and differ between populations; therefore, more attention should be paid to understanding the cultural meanings and attitudes of people in fertility decision-making, including also the views of adolescent men and other types of population, stratified by age groups and social conditions.

Keywords: adolescent, qualitative research, unplanned pregnancy, unwanted pregnancy

Procedia PDF Downloads 197

Authors: M. Yosofi, S. Ezeddini, A. Ollivier, V. Lavaste, C. Mayousse

Abstract:

Additive manufacturing processes have changed significantly in a wide range of industries and their application progressed from rapid prototyping to production of end-use products. However, their environmental impact is still a rather open question. In order to support the growth of this technology in the industrial sector, environmental aspects should be considered and predictive models may help monitor and reduce the environmental footprint of the processes. This work presents predictive models based on a previously developed methodology for the environmental impact evaluation combined with a technical and economical assessment. Here we applied the methodology to the Fused Deposition Modeling process. First, we present the predictive models relative to different types of machines. Then, we present a decision-making tool designed to identify the optimum manufacturing strategy regarding technical, economic, and environmental criteria.

Keywords: additive manufacturing, decision-makings, environmental impact, predictive models

Procedia PDF Downloads 113

Authors: Yingisani Chabalala, Elhadi Adam, Khalid Adem Ali

Abstract:

Mapping smallholder fruit plantations using optical data is challenging due to morphological landscape heterogeneity and crop types having overlapped spectral signatures. Furthermore, cloud covers limit the use of optical sensing, especially in subtropical climates where they are persistent. This research assessed the effectiveness of Sentinel-1 (S1) and Sentinel-2 (S2) data for mapping fruit trees and co-existing land-use types by using support vector machine (SVM) and random forest (RF) classifiers independently. These classifiers were also applied to fused data from the two sensors. Feature ranks were extracted using the RF mean decrease accuracy (MDA) and forward variable selection (FVS) to identify optimal spectral windows to classify fruit trees. Based on RF MDA and FVS, the SVM classifier resulted in relatively high classification accuracy with overall accuracy (OA) = 0.91.6% and kappa coefficient = 0.91% when applied to the fused satellite data. Application of SVM to S1, S2, S2 selected variables and S1S2 fusion independently produced OA = 27.64, Kappa coefficient = 0.13%; OA= 87%, Kappa coefficient = 86.89%; OA = 69.33, Kappa coefficient = 69. %; OA = 87.01%, Kappa coefficient = 87%, respectively. Results also indicated that the optimal spectral bands for fruit tree mapping are green (B3) and SWIR_2 (B10) for S2, whereas for S1, the vertical-horizontal (VH) polarization band. Including the textural metrics from the VV channel improved crop discrimination and co-existing land use cover types. The fusion approach proved robust and well-suited for accurate smallholder fruit plantation mapping.

Keywords: smallholder agriculture, fruit trees, data fusion, precision agriculture

Procedia PDF Downloads 37

Authors: Wentian Shi, Daming Shi, Maysam Orouskhani, Feng Tian

Abstract:

Whereas currently the most prevalent deep learning methods require a large amount of data for training, few-shot learning tries to learn a model from limited data without extensive retraining. In this paper, we present a loss function based on triplet loss for solving few-shot problem using metric based learning. Instead of setting the margin distance in triplet loss as a constant number empirically, we propose an adaptive margin distance strategy to obtain the appropriate margin distance automatically. We implement the strategy in the deep siamese network for deep metric embedding, by utilizing an optimization approach by penalizing the worst case and rewarding the best. Our experiments on image recognition and co-segmentation model demonstrate that using our proposed triplet loss with adaptive margin distance can significantly improve the performance.

Keywords: few-shot learning, triplet network, adaptive margin, deep learning

Procedia PDF Downloads 155

Authors: Ariangelo Hauer Dias Filho, Gustavo Antoniácomi de Carvalho, Benjamim de Melo Carvalho

Abstract:

The work´s purpose was to evaluate the possibility of manufacturing casting tools utilizing Fused Filament Fabrication, a 3D printing technique, without any post-processing on the printed part. Taguchi Orthogonal array was used to evaluate the influence of extrusion temperature, bed temperature, layer height, and infill on the dimensional accuracy of a 3D-Printed Polymer Model. A Zeiss T-SCAN CS 3D Scanner was used for dimensional evaluation of the printed parts within the limit of ±0,2 mm. The mold capabilities were tested with the printed model to check how it would interact with the green sand. With little adjustments in the 3D model, it was possible to produce rapid tools without the need for post-processing for iron casting. The results are important for reducing time and cost in the development of such tools.

Keywords: additive manufacturing, Taguchi method, rapid tooling, fused filament fabrication, casting mold

Procedia PDF Downloads 129

Authors: M. Mortas, H. Asar, P. Dursun Cebi

Abstract:

The visionary works of architectural representation in the 21st century's present situation, are practiced through the methodologies which try to expose the intellectual and theoretical essences of futurologist positions that are revealed with this era's interactions. Expansions of conceptual and contextual inputs related to one architectural design representation, depend on its deepness of critical attitudes, its interactions with the concepts such as experience, meaning, affection, psychology, perception and aura, as well as its communication with spatial, cultural and environmental factors. The purpose of this research study is to be able to offer methodological application areas for the design dimensions of experiential practices into architectural design studios, by focusing on the architectural representative narrations of 'transformation,' 'metamorphosis,' 'morphogenesis,' 'in-betweenness', 'superposition' and 'intertwine’ in which they affect and are affected by the today’s spatiotemporal hybridizations of architecture. The narrative representations and the visual theory paradigms of the designers are chosen under the main title of 'transformation' for the investigation of these visionary and critical representations' dismantlings and decodings. Case studies of this research area are chosen from Neil Spiller, Bryan Cantley, Perry Kulper and Dan Slavinsky’s transformative, morphogenetic representations. The theoretical dismantlings and decodings which are obtained from these artists’ contemporary architectural representations are tried to utilize and practice in the structural design studios as alternative methodologies when to approach architectural design processes, for enriching, differentiating, diversifying and 'transforming' the applications of so far used design process precedents. The research aims to indicate architectural students about how they can reproduce, rethink and reimagine their own representative lexicons and so languages of their architectural imaginations, regarding the newly perceived tectonics of prosthetic, biotechnology, synchronicity, nanotechnology or machinery into various experiential design workshops. The methodology of this work can be thought as revealing the technical and theoretical tools, lexicons and meanings of contemporary-visionary architectural representations of our decade, with the essential contents and components of hermeneutics, etymology, existentialism, post-humanism, phenomenology and avant-gardism disciplines to re-give meanings the architectural visual theorists’ transformative representations of our decade. The value of this study may be to emerge the superposed and overlapped atmospheres of futurologist architectural representations for the students who need to rethink on the transcultural, deterritorialized and post-humanist critical theories to create and use the representative visual lexicons of themselves for their architectural soft machines and beings by criticizing the now, to be imaginative for the future of architecture.

Keywords: architectural design studio, visionary lexicon, narrative architecture, transformative representation

Procedia PDF Downloads 130

Authors: Arpit Rai

Abstract:

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

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

Procedia PDF Downloads 77

Authors: Alison Wilde, Stephen Millett

Abstract:

The cultural value of disabled people has largely been overlooked within forms of media and cultural analysis until the 1980s, when disabled people and disability studies highlighted the cultural misrecognition of disabled people and called for improved forms of cultural recognition and representation. Despite an increase in cultural analysis of representations of disabled people, much has been assumed about how images are read, and little work has been done on the value attributed to disabled people by media audiences and the viewing interests and encounters of film audiences. In particular, there has been little work on film reception, or on the way that young people interpret images of disability. We set out to understand some of the ways that young people read disability imagery, by showing small groups of young people different types of film featuring impairments, chosen from three different eras in film. These were Freaks, Rear Window (remake), and Finding Nemo. The discussions after these films allowed them to explore their own experiences of disability alongside the evolution of cultural representations; in so doing they discussed significant themes of cultural value and reflected on their own identities, e.g. in/dependency, autonomy, and competency and the ways these intersected with self-identity, and attitudes to disabled people.

Keywords: film, audience, identity, disability

Procedia PDF Downloads 410

Authors: G. Purushotham, Joel Hemanth

Abstract:

An investigation has been carried out to fabricate and evaluate the strength and soundness of chilled composites consisting of nickel matrix and fused silica particles (size 40–150 μm) in the matrix. The dispersoid added ranged from 3 to 12 wt. % in steps of 3%. The resulting composites cast in moulds containing metallic and non-metallic chill blocks (MS, SiC, and Cu) were tested for their microstructure and mechanical properties. The main objective of the present research is to obtain fine grain Ni/SiO2 chilled sound composite having very good mechanical properties. Results of the investigation reveal the following: (1) Strength of the composite developed is highly dependent on the location of the casting from where the test specimens are taken and also on the dispersoid content of the composite. (2) Chill thickness and chill material, however, does significantly affect the strength and soundness of the composite. (3) Soundness of the composite developed is highly dependent on the chilling rate as well as the dispersoid content. An introduction of chilling and increase in the dispersoid content of the material both result in an increase in the ultimate tensile strength (UTS) of the material. The temperature gradient developed during solidification and volumetric heat capacity (VHC) of the chill used is the important parameters controlling the soundness of the composite. (4) Thermal properties of the end chills are used to determine the magnitude of the temperature gradient developed along the length of the casting solidifying under the influence of chills.

Keywords: metal matrix composite, mechanical properties, corrosion behavior, nickel alloy, fused silica, chills

Procedia PDF Downloads 389

Authors: Athanasios Raftopoulos

Abstract:

In this paper, an attempt is made to explain the evolution or development of human’s representational arsenal from its humble beginnings to its modern abstract symbols. Representations are physical entities that represent something else. To represent a thing (in a general sense of “thing”) means to use in the mind or in an external medium a sign that stands for it. The sign can be used as a proxy of the represented thing when the thing is absent. Representations come in many varieties, from signs that perceptually resemble their representative to abstract symbols that are related to their representata through conventions. Relying the distinction among indices, icons, and symbols, it is explained how symbolic representations gradually emerged from indices and icons. To understand the development or evolution of our representational arsenal, the development of the cognitive capacities that enabled the gradual emergence of representations of increasing complexity and expressive capability should be examined. The examination of these factors should rely on a careful assessment of the available empirical neuroscientific and paleo-anthropological evidence. These pieces of evidence should be synthesized to produce arguments whose conclusions provide clues concerning the developmental process of our representational capabilities. The analysis of the empirical findings in this paper shows that Homo Erectus was able to use both icons and symbols. Icons were used as external representations, while symbols were used in language. The first step in the emergence of representations is that a sensory-motor purely causal schema involved in indices is decoupled from its normal causal sensory-motor functions and serves as a representation of the object that initially called it into play. Sensory-motor schemes are tied to specific contexts of the organism-environment interactions and are activated only within these contexts. For a representation of an object to be possible, this scheme must be de-contextualized so that the same object can be represented in different contexts; a decoupled schema loses its direct ties to reality and becomes mental content. The analysis suggests that symbols emerged due to selection pressures of the social environment. The need to establish and maintain social relationships in ever-enlarging groups that would benefit the group was a sufficient environmental pressure to lead to the appearance of the symbolic capacity. Symbols could serve this need because they can express abstract relationships, such as marriage or monogamy. Icons, by being firmly attached to what can be observed, could not go beyond surface properties to express abstract relations. The cognitive capacities that are required for having iconic and then symbolic representations were present in Homo Erectus, which had a language that started without syntactic rules but was structured so as to mirror the structure of the world. This language became increasingly complex, and grammatical rules started to appear to allow for the construction of more complex expressions required to keep up with the increasing complexity of social niches. This created evolutionary pressures that eventually led to increasing cranial size and restructuring of the brain that allowed more complex representational systems to emerge.

Keywords: mental representations, iconic representations, symbols, human evolution

Procedia PDF Downloads 40

Authors: Reza Samsami

Abstract:

The aim of the study was to degradation of polycyclic aromatic hydrocarbons (PAHs) by proxy-acid method. The amounts of PAHs were determined in a silty-clay soil sample of an aged oil refinery field in Abadan, Iran. Proxy-acid treatment method was investigated. The results have shown that the proxy-acid system is an effective method for degradation of PAHs. The results also demonstrated that the number of fused aromatic rings have not significant effects on PAH removal by proxy-acid method. The results also demonstrated that the number of fused aromatic rings have not significant effects on PAH removal by proxy-acid method.

Keywords: proxy-acid treatment, silty-clay soil, PAHs, degradation

Procedia PDF Downloads 256

Authors: Romisaa Ali

Abstract:

The Policy Gradient approach is one of the deep reinforcement learning families that combines deep neural networks (DNN) with reinforcement learning RL to discover the optimum of the control problem through experience gained from the interaction between the robot and its surroundings. In contrast to earlier policy gradient algorithms, which were unable to handle these two types of error because of over-or under-estimation introduced by the deep neural network model, this article will discuss the state-of-the-art SOTA policy gradient technique, trust region policy optimization (TRPO), by applying this method in various environments compared to another policy gradient method, the Proximal Policy Optimization (PPO), to explain their robust optimization, using this SOTA to gather experience data during various training phases after observing the impact of hyper-parameters on neural network performance.

Keywords: deep neural networks, deep reinforcement learning, proximal policy optimization, state-of-the-art, trust region policy optimization

Procedia PDF Downloads 156

Authors: B. Khatri, K. Lappe, M. Habedank, T. Müller, C. Megnin, T. Hanemann

Abstract:

We present a process flow for the development of ceramic-filled polymer composite filaments compatible with the fused deposition modeling (FDM) 3D printing process. Thermoplastic-ceramic composites were developed using acrylonitrile butadiene styrene (ABS) and 10- and 20 vol.% barium titanate (BaTiO3) powder (corresponding to 39.47- and 58.23 wt.% respectively) and characterized for their flow properties. To make them compatible with the existing FDM process, the composites were extruded into filaments. These composite filaments were subsequently structured into tensile stress specimens using a commercially available FDM 3D printer and characterized for their mechanical properties. Rheometric characterization of the material composites revealed non-Newtonian behavior with the viscosity logarithmically decreasing over increasing shear rates, as well as higher viscosities for samples with higher BaTiO3 filler content for a given shear rate (with the ABS+20vol.% BaTiO3 composite being over 50% more viscous compared to pure ABS at a shear rate of 1x〖10〗^3 s^(-1)). Mechanical characterization of the tensile stress specimens exhibited increasingly brittle behavior as well as a linearly decreasing ultimate tensile strength of the material composites with increasing volumetric ratio of BaTiO3 (from σ_max=32.4MPa for pure ABS to σ_max=21.3MPa for ABS+20vol.% BaTiO3). Further studies being undertaken include the development of composites with higher filler concentrations, sintering of the printed composites to yield pure dielectric structures and the determination of the dielectric characteristics of the composites.

Keywords: ceramic composites, fused deposition modeling, material characterization, rapid prototyping

Procedia PDF Downloads 324

Authors: Nishank Raisinghani

Abstract:

Efficiently predicting drug response remains a challenge in the realm of drug discovery. To address this issue, we propose four model architectures that combine graphical representation with varying positions of multiheaded self-attention mechanisms. By leveraging two types of multi-omics data, transcriptomics and genomics, we create a comprehensive representation of target cells and enable drug response prediction in precision medicine. A majority of our architectures utilize multiple transformer models, one with a graph attention mechanism and the other with a multiheaded self-attention mechanism, to generate latent representations of both drug and omics data, respectively. Our model architectures apply an attention mechanism to both drug and multiomics data, with the goal of procuring more comprehensive latent representations. The latent representations are then concatenated and input into a fully connected network to predict the IC-50 score, a measure of cell drug response. We experiment with all four of these architectures and extract results from all of them. Our study greatly contributes to the future of drug discovery and precision medicine by looking to optimize the time and accuracy of drug response prediction.

Keywords: drug discovery, transformers, graph neural networks, multiomics

Procedia PDF Downloads 130

Authors: Ghorab Sara, Meziani Lila, Rubin Harvey Stuart

Abstract:

The rapid growth of deep learning has led to intricate and resource-intensive deep neural networks widely used in computer vision tasks. However, their complexity results in high computational demands and memory usage, hindering real-time application. To address this, research focuses on model compression techniques. The paper provides an overview of recent advancements in compressing neural networks and categorizes the various methods into four main approaches: network pruning, quantization, network decomposition, and knowledge distillation. This paper aims to provide a comprehensive outline of both the advantages and limitations of each method.

Keywords: model compression, deep neural network, pruning, knowledge distillation, quantization, low-rank decomposition

Procedia PDF Downloads 32

Authors: Efrain Rodriguez, Sergio Pertuz, Cristhian Riano

Abstract:

Tool-path generation is an essential step in the FFF (Fused Filament Fabrication)-based Additive Manufacturing (AM) process planning. In the manufacture of a mechanical part by using additive processes, high resource consumption and prolonged production times are inherent drawbacks of these processes mainly due to non-optimized tool-path generation. In this work, we propose a heuristic-search intelligent algorithm-based approach for optimized tool-path generation for FFF-based AM. The main benefit of this approach is a significant reduction of travels without material deposition when the AM machine performs moves without any extrusion. The optimization method used reduces the number of travels without extrusion in comparison with commercial software as Slic3r or Cura Engine, which means a reduction of production time.

Keywords: additive manufacturing, tool-path optimization, fused filament fabrication, process planning

Procedia PDF Downloads 432

Authors: Elisaveta Doncheva, Jelena Djokikj, Ognen Tuteski, Bojana Hadjieva

Abstract:

In the past decade, additive manufacturing technology or 3D printing has been promoted as an efficient method for fabricating hybrid composite materials and structures with superior mechanical properties and complex shape and geometry. Fused deposition modeling (FDM) process is commonly used additive manufacturing technique for production of polymer products. Therefore, many studies and experiments are focused on investigating the possibilities for improving the obtained results on product properties as a key factor for expanding the spectrum of their application. This article provides an extensive review on recent research advances in FDM and reports on studies that cover the effects of process parameters, material, and design of the product properties. The paper conclusions provide a clear up-to date information for optimum efficiency and enhancement of the mechanical properties of 3D printed samples and recommends further research work and investigations.

Keywords: additive manufacturing, critical parameters, filament, print orientation, 3D printing

Procedia PDF Downloads 179

Authors: Lokes Parvatha Kumaran S., Sakthi Jay Mahenthar C., Sathyaprakash P., Jayakumar V., Shobanadevi A.

Abstract:

With the evolution of technology, the need to solve complex computational problems like machine learning and deep learning has shot up. But even the most powerful classical supercomputers find it difficult to execute these tasks. With the recent development of quantum computing, researchers and tech-giants strive for new quantum circuits for machine learning tasks, as present works on Quantum Machine Learning (QML) ensure less memory consumption and reduced model parameters. But it is strenuous to simulate classical deep learning models on existing quantum computing platforms due to the inflexibility of deep quantum circuits. As a consequence, it is essential to design viable quantum algorithms for QML for noisy intermediate-scale quantum (NISQ) devices. The proposed work aims to explore Variational Quantum Circuits (VQC) for Deep Reinforcement Learning by remodeling the experience replay and target network into a representation of VQC. In addition, to reduce the number of model parameters, quantum information encoding schemes are used to achieve better results than the classical neural networks. VQCs are employed to approximate the deep Q-value function for decision-making and policy-selection reinforcement learning with experience replay and the target network.

Keywords: quantum computing, quantum machine learning, variational quantum circuit, deep reinforcement learning, quantum information encoding scheme

Procedia PDF Downloads 109

Authors: Yifan Fan, Xudong Luo, Pingping Lin

Abstract:

Nowadays, dialogue systems increasingly become the way for humans to access many computer systems. So, humans can interact with computers in natural language. A dialogue system consists of three parts: understanding what humans say in natural language, managing dialogue, and generating responses in natural language. In this paper, we survey deep learning based methods for dialogue management, response generation and dialogue evaluation. Specifically, these methods are based on neural network, long short-term memory network, deep reinforcement learning, pre-training and generative adversarial network. We compare these methods and point out the further research directions.

Keywords: dialogue management, response generation, deep learning, evaluation

Procedia PDF Downloads 153

Authors: Atef Ahmad Bleibel

Abstract:

Most codes of practice, like ACI 318-14, require the use of strut-and-tie modeling to analyze and design reinforced concrete deep beams. Though, investigations that conducted on deep beams do not include ring deep beams of influential parameters. This work presents an analytical parametric study using strut-and-tie modeling stated by ACI 318-14 to predict load capacity of 20 reinforced concrete ring deep beam specimens with different parameters. The parameters that were under consideration in the current work are ring diameter (Dc), number of supports (NS), width of ring beam (bw), concrete compressive strength (f'c) and width of bearing plate (Bp). It is found that the load capacity decreases by about 14-36% when ring diameter increases by about 25-75%. It is also found that load capacity increases by about 62-189% when number of supports increases by about 33-100%, while the load capacity increases by about 25-75% when the beam ring width increases by about 25-75%. Finally, it is found that load capacity increases by about 24-76% when compressive strength increases by about 24-76%, while the load capacity increases by about 5-16% when Bp increases by about 25-75%.

Keywords: load parameters, reinforced concrete, ring deep beam, strut and tie

Procedia PDF Downloads 91

Authors: Letizia Maria Drammis, Maria Antonella Brandimonte

Abstract:

The idea that people use space representations to think about time and number received support from several lines of research. However, how these representations develop in children and then shape space-time and space-number mappings is still a debated issue. In the present study, 40 children (20 pre-schoolers and 20 elementary-school children) performed 4 main tasks, which required the use of more concrete (non-symbolic) or more abstract (symbolic) space-time and space-number associations. In the non-symbolic conditions, children were required to order pictures of everyday-life events occurring in a specific temporal order (Temporal sequences) and of quantities varying in numerosity (Numerical sequences). In the symbolic conditions, they were asked to perform the typical time-to-position and number-to-position tasks by mapping time-related words and numbers onto lines. Results showed that children performed reliably better in the non-symbolic Time conditions than the symbolic Time conditions, independently of age, whereas only pre-schoolers performed worse in the Number-to-position task (symbolic) as compared to the Numerical sequence (non-symbolic) task. In addition, only older children mapped time-related words onto space following the typical left-right orientation, pre-schoolers’ performance being somewhat mixed. In contrast, mapping numbers onto space showed a clear left-right orientation, independently of age. Overall, these results indicate a cross-domain difference in the way younger and older children process time and number, with time-related tasks being more difficult than number-related tasks only when space-time tasks require symbolic representations.

Keywords: space-time associations, space-number associations, orientation, children

Procedia PDF Downloads 324

Authors: Bih-O Lee, Hsiu-Wan Hsieh, Hsiu-Chen Liu, Mer Yu Pan

Abstract:

Background: Illness perceptions are developed when people face health-threatening situations. Previous research suggests that understanding discrepancies between illness perceptions of patients and caregivers may need to improve quality of health care. Objective: This study examined the differences between illness perceptions of injured patients and those of their caregivers. Methods: Comparative study design was used. The study setting was the surgical wards of a teaching hospital in Taiwan. Participants were 127 pairs of injured patients and their caregivers. The participants completed socio-demographic data and completed the Chinese Illness Perception Questionnaire Revised-Trauma, which comprises eight subscales. Clinical data of the injured patients was obtained from medical records. Results: This study found that injured patients were more pessimistic than their caregivers about the injury. There were significant differences between patients and caregivers insofar as patients perceived more physical symptoms, scored higher in terms of reasons for their injury, had more negative emotions and experienced more consequences than caregivers. Elderly caregivers and caregivers for patients who were over 65, severely injured and admitted to an ICU perceived more negative perceptions about the injury. Conclusions: This study indicated that patients and caregivers had negative illness representations several months after injury although the intensity of their perceptions was different. The interventions should highlight the need to assist patients and caregivers after injury.

Keywords: illness representations, injury, caregivers, comparative study

Procedia PDF Downloads 365

Authors: Sushish Baral, Pratibha Joshi, Yaman Maharjan

Abstract:

Glaucoma is a chronic eye condition that causes vision loss that is irreversible. Early detection and treatment are critical to prevent vision loss because it can be asymptomatic. For the identification of glaucoma, multiple deep learning algorithms are used. Transformer-based architectures, which use the self-attention mechanism to encode long-range dependencies and acquire extremely expressive representations, have recently become popular. Convolutional architectures, on the other hand, lack knowledge of long-range dependencies in the image due to their intrinsic inductive biases. The aforementioned statements inspire this thesis to look at transformer-based solutions and investigate the viability of adopting transformer-based network designs for glaucoma detection. Using retinal fundus images of the optic nerve head to develop a viable algorithm to assess the severity of glaucoma necessitates a large number of well-curated images. Initially, data is generated by augmenting ocular pictures. After that, the ocular images are pre-processed to make them ready for further processing. The system is trained using pre-processed images, and it classifies the input images as normal or glaucoma based on the features retrieved during training. The Vision Transformer (ViT) architecture is well suited to this situation, as it allows the self-attention mechanism to utilise structural modeling. Extensive experiments are run on the common dataset, and the results are thoroughly validated and visualized.

Keywords: glaucoma, vision transformer, convolutional architectures, retinal fundus images, self-attention, deep learning

Procedia PDF Downloads 181

Authors: C. Jameke

Abstract:

During the cognitivist era in cognitive psychology, a theory of internal rules and symbolic representations was posited as an account of human cognition. This type of cognitive architecture had its heyday during the 1970s and 80s, but it has now been largely abandoned in favour of subsymbolic architectures (e.g. connectionism), non-representational frameworks (e.g. dynamical systems theory), and statistical approaches such as Bayesian theory. In this presentation I describe this changing landscape of research, and comment on the increasing influence of neuroscience on cognitive psychology. I then briefly review a few recent developments in connectionism, and neurocomputation relevant to cognitive psychology, and critically discuss the assumption made by some researchers in these frameworks that higher-level aspects of human cognition are simply emergent properties of massively large distributed neural networks

Keywords: connectionism, emergentism, postocgnitivist, representations, subsymbolic archiitecture

Procedia PDF Downloads 555

Authors: Natalia Fijol, Aji P. Mathew

Abstract:

We present the processing and characterisation of three-dimensional (3D) monolith filters based on polylactic acid (PLA) reinforced with various nature-derived nanospecies such as hydroxyapatite, modified cellulose fibers and chitin fibers. The nanospecies of choice were dispersed in PLA through Thermally Induced Phase Separation (TIPS) method. The biocomposites were developed via solvent-assisted blending and the obtained pellets were further single-screw extruded into 3D-printing filaments and processed into various geometries using Fused Deposition Modelling (FDM) technique. The printed prototypes included cubic, cylindrical and hour-glass shapes with diverse patterns of printing infill as well as varying pore structure including uniform and multiple level gradual pore structure. The pores and channel structure as well as overall shape of the prototypes were designed in attempt to optimize the flux and maximize the adsorption-active time. FDM is a cost and energy-efficient method, which does not require expensive tools and elaborated post-processing maintenance. Therefore, FDM offers the possibility to produce customized, highly functional water purification filters with tuned porous structures suitable for removal of wide range of common water pollutants. Moreover, as 3D printing becomes more and more available worldwide, it allows producing portable filters at the place and time where they are most needed. The study demonstrates preparation route for the PLA-based, fully biobased composite and their processing via FDM technique into water purification filters, addressing water treatment challenges on an industrial scale.

Keywords: fused deposition modelling, water treatment, biomaterials, 3D printing, nanocellulose, nanochitin, polylactic acid

Procedia PDF Downloads 105

Authors: Maylen Villamanan Alba

Abstract:

Latin America is marked by its history: conquest, colonialism, and slavery left deep footprints in most Latin American countries. Also, the past century has been affected by wars, military dictatorships, and political violence, which profoundly influenced Latin American popular culture. Consequently, reminiscences of historical crimes are frequently present in daily life, media, public opinion, and arts. This legacy is remembered in novels, paintings, songs, and films. In fact, Latin American cinema has a trend which refers to the verisimilitude with reality in fiction films. These films about historical violence are narrated as fictional characters, but their stories are based on real historical contexts. Therefore, cultural criminology has considered films as a significant field to understand social representations of violence related to historical crimes. The aim of the present contribution is to analyze the legacy of past and historical memory in social representations of violence in Latin American cinema as a critical approach to historical crimes. This qualitative research is based on content analysis. The sample is seven multi-award winning films of the International Festival of New Latin American Cinema of Havana. The films selected are Kamchatka, Argentina (2002); Carandiru, Brazil (2003); Enlightened by fire, Argentina (2005); Post-mortem, Chile (2010); No, Chile (2012) Wakolda; Argentina (2013) and The Clan, Argentina (2015). Cultural criminology highlights that cinema shapes meanings of social practices such as historical crimes. Critical criminology offers a critical theory framework to interpret Latin American cinema. This analysis reveals historical conditions deeply associated with power relationships, policy, and inequality issues. As indicated by this theory, violence is characterized as a structural process based on social asymmetries. These social asymmetries are crossed by social scopes, including institutional and personal dimensions. Thus, institutions of the states are depicted through personal stories of characters involved with human conflicts. Intimacy and social background are linked in the personages who simultaneously perform roles such as soldiers, policemen, professionals or inmates and they are at the same time depict as human beings with family, gender, racial, ideological or generational issues. Social representations of violence related to past legacy are a portrait of historical crimes perpetrated against Latin American citizens. Thereby, they have contributed to political positions, social behaviors, and public opinion. The legacy of these historical crimes suggests a path that should never be taken again. It means past legacy is a reminder, a warning, and a historic lesson for Latin American people. Social representations of violence are permeated by historical memory as denunciation under a critical approach.

Keywords: Latin American cinema, historical memory, social representation, violence

Procedia PDF Downloads 136

Authors: Yin Yuanling

Abstract:

A Deep Feedback Recurrent Neural Network (DFRNN) and Bidirectional Long Short-Term Memory (BiLSTM) are designed to address the problem of low accuracy of traditional relationship extraction models. This method combines a deep feedback-based recurrent neural network (DFRNN) with a bi-directional long short-term memory (BiLSTM) approach. The method combines DFRNN, which extracts local features of text based on deep feedback recurrent mechanism, BiLSTM, which better extracts global features of text, and Self-Attention, which extracts semantic information. Experiments show that the method achieves an F1 value of 76.69% on the CEC dataset, which is 0.0652 better than the BiLSTM+Self-ATT model, thus optimizing the performance of the deep learning method in the event relationship extraction task.

Keywords: event relations, deep learning, DFRNN models, bi-directional long and short-term memory networks

Procedia PDF Downloads 123

Authors: Delowar Hossain, Genci Capi

Abstract:

This paper concerns with the problem of deep learning parameters tuning using a genetic algorithm (GA) in order to improve the performance of deep learning (DL) method. We present a GA based DL method for robot object recognition and grasping. GA is used to optimize the DL parameters in learning procedure in term of the fitness function that is good enough. After finishing the evolution process, we receive the optimal number of DL parameters. To evaluate the performance of our method, we consider the object recognition and robot grasping tasks. Experimental results show that our method is efficient for robot object recognition and grasping.

Keywords: deep learning, genetic algorithm, object recognition, robot grasping

Procedia PDF Downloads 338