Search results

The purpose of this paper is to define a methodology to analyze links between programming topics and libraries starting from GitHub data.

This paper developed an analysis over machine learning repositories on GitHub, finding communities of repositories and studying the anatomy of collaboration around a popular topic such as machine learning.

This analysis indicates the significant importance of programming languages and technologies such as Python and Jupyter Notebook. It also shows the rise of deep learning and of specific libraries such as Tensorflow from Google.

There exists no survey or analysis based on how developers influence each other for specific topics. Other researchers focused their analysis on the collaborative structure and social impact instead of topic impact. Using this methodology to analyze programming topics is important not just for machine learning but also for other topics.

Visually impaired people usually struggle with doing daily tasks due to a lack of visual cues. For image captioning assistive applications, most applications require an Internet connection for the image captioning generation function to work properly. In this study, we developed MyUEVision, an application that assists visually impaired people by generating image captions that can work with and without the Internet. This work also involves reviewing some image captioning models for this application.

The author has selected and experimented with three image captioning models for online models and two image captioning models for offline models. The user experience (UX) design was designed based on the problems faced by visually impaired users when using mobile applications. The application is developed for the Android platform, and the offline model is integrated into the application for the image captioning generation function to work offline.

After conducting experiments for selecting online and offline models, ExpansionNet V2 is chosen for the online model and VGG16 + long short-term memory (LSTM) is chosen for the offline model. The application is then developed and assessed, and the results show that the application can generate image captions with or without the Internet, providing the best result when having an Internet connection, and the image is captured in good lighting with a few objects.

MyUEVision stands out for its both online and offline functionality. This approach ensures the image captioning generator works with or without the Internet, setting it apart as a unique solution to address the needs of visually impaired individuals.

The practice of artificial intelligence (AI) is increasingly being promoted by technology developers. However, its adoption rate is still reported as low in the construction industry due to a lack of expertise and the limited reliable applications for AI technology. Hence, this paper aims to present the detailed outcome of experimentations evaluating the applicability and the performance of AI object detection algorithms for construction modular object detection.

This paper provides a thorough evaluation of two deep learning algorithms for object detection, including the faster region-based convolutional neural network (faster RCNN) and single shot multi-box detector (SSD). Two types of metrics are also presented; first, the average recall and mean average precision by image pixels; second, the recall and precision by counting. To conduct the experiments using the selected algorithms, four infrastructure and building construction sites are chosen to collect the required data, including a total of 990 images of three different but common modular objects, including modular panels, safety barricades and site fences.

The results of the comprehensive evaluation of the algorithms show that the performance of faster RCNN and SSD depends on the context that detection occurs. Indeed, surrounding objects and the backgrounds of the objects affect the level of accuracy obtained from the AI analysis and may particularly effect precision and recall. The analysis of loss lines shows that the loss lines for selected objects depend on both their geometry and the image background. The results on selected objects show that faster RCNN offers higher accuracy than SSD for detection of selected objects.

The results show that modular object detection is crucial in construction for the achievement of the required information for project quality and safety objectives. The detection process can significantly improve monitoring object installation progress in an accurate and machine-based manner avoiding human errors. The results of this paper are limited to three construction sites, but future investigations can cover more tasks or objects from different construction sites in a fully automated manner.

This paper’s originality lies in offering new AI applications in modular construction, using a large first-hand data set collected from three construction sites. Furthermore, the paper presents the scientific evaluation results of implementing recent object detection algorithms across a set of extended metrics using the original training and validation data sets to improve the generalisability of the experimentation. This paper also provides the practitioners and scholars with a workflow on AI applications in the modular context and the first-hand referencing data.

Vocabulary learning is a difficult task for children without hearing ability. Absence of enough learning centers and effective learning tools aggravate the problem. Modern technology can be utilized fruitfully to find solutions to the learning difficulties experienced by the deaf. The purpose of this paper is to present SiLearn – a novel technology based tool for teaching/learning sign vocabulary.

The proposed mobile application can act as a visual dictionary for deaf people. SiLearn is equipped with features that can automatically detect both text and physical objects and convert them to their corresponding signs. For testing the effectiveness of the proposed mobile application quantitative analyses were done. Quantitative analysis is based on testing a class of 28 students belonging to St Clare Oral School for the Deaf, Kerala, India. This group consisted of 17 boys and 11 girls. Analysis was also done through questionnaire. Questionnaires were given to teachers, parents of deaf students learning sign language and other sign language learners.

Results indicate that as SiLearn is very effective in sign vocabulary development. It can enhance vocabulary learning rate considerably.

This is the first time that artificial intelligence (AI) based techniques are used for early stage sign language learning. SiLearn can equally be used by children, parents and teachers for learning sign language.

This paper aims to Solar photovoltaic (PV) power can significantly impact the power system because of its intermittent nature. Hence, an accurate solar PV power forecasting model is required for appropriate power system planning.

In this paper, a long short-term memory (LSTM)-based double deep Q-learning (DDQL) neural network (NN) is proposed for forecasting solar PV power indirectly over the long-term horizon. The past solar irradiance, temperature and wind speed are used for forecasting the solar PV power for a place using the proposed forecasting model.

The LSTM-based DDQL NN reduces over- and underestimation and avoids gradient vanishing. Thus, the proposed model improves the forecasting accuracy of solar PV power using deep learning techniques (DLTs). In addition, the proposed model requires less training time and forecasts solar PV power with improved stability.

The proposed model is trained and validated for several places with different climatic patterns and seasons. The proposed model is also tested for a place with a temperate climatic pattern by constructing an experimental solar PV system. The training, validation and testing results have confirmed the practicality of the proposed solar PV power forecasting model using LSTM-based DDQL NN.

Utilization of sustainable materials is a global demand in the construction industry. Hence, this study aims to integrate waste management and artificial intelligence by developing an artificial neural network (ANN) model to predict the compressive strength of green concrete. The proposed model allows the use of recycled coarse aggregate (RCA), recycled fine aggregate (RFA) and fly ash (FA) as partial replacements of concrete constituents.

The model is constructed, trained and validated using python through a set of experimental data collected from the literature. The model’s architecture comprises an input layer containing seven neurons representing concrete constituents and two neurons as the output layer to represent the 7- and 28-days compressive strength. The model showed high performance through multiple metrics, including mean squared error (MSE) of 2.41 and 2.00 for training and testing data sets, respectively.

Results showed that cement replacement with 10% FA causes a slight reduction up to 9% in the compressive strength, especially at early ages. Moreover, a decrease of nearly 40% in the 28-days compressive strength was noticed when replacing fine aggregate with 25% RFA.

The research is limited to normal compressive strength of green concrete with a range of 25 to 40 MPa.

The developed model is designed in a flexible and user-friendly manner to be able to contribute to the sustainable development of the construction industry by saving time, effort and cost consumed in the experimental testing of materials.

Green concrete containing wastes can solve several environmental problems, such as waste disposal problems, depletion of natural resources and energy consumption.

This research proposes a machine learning prediction model using the Python programming language to estimate the compressive strength of a green concrete mix that includes construction and demolition waste and FA. The ANN model is used to create three guidance charts through a parametric study to obtain the compressive strength of green concrete using RCA, RFA and FA replacements.

Internet of Things (IoT) involves connecting physical objects to the internet to provide opportunities to build smart systems or applications. IoT paradigm assumes many devices connected over a conventional intent network. These devices usually have restricted resources, so moving part of the service implementation to a cloud infrastructure is a prominent solution. This study aims to proposes in this project human voice as a potential interface for one or more devices in IoT ecosystem enabling issuing commands and receiving information.

System design is the process of defining the elements of a system such as the architecture, modules and components, the different interfaces of those components and the data that goes through that system. It is meant to satisfy specific needs and requirements of a business or organization through the engineering of a coherent and well-running system.

The main aim of this proposed work is to develop a ticket booking application that performs all the operations by speech recognition. Hence, visually impaired people can make use of this application. There are several applications that help visually impaired people. This application adds extra features to those available soft wares. Using this, visually impaired people can book the tickets without the help of personal assistants. For future research, this study hopes to extend this application to perform various other operations that will help visually impaired people to do their daily activities like normal people without the help of personal assistants. For example, making a phone call, sending text messages, booking a taxi, easy navigation, etc.

System design involves the identification of classes, their relationship as well as their collaboration. In objector, classes are divided into entity classes and control classes.

Machine learning (ML) has recently gained momentum as a method for measurement in strategy research. Yet, little guidance exists regarding how to appropriately apply the method for this purpose in our discipline. We address this by offering a guide to the application of ML in strategy research, with a particular emphasis on data handling practices that should improve our ability to accurately measure our constructs of interest using ML techniques. We offer a brief overview of ML methodologies that can be used for measurement before describing key challenges that exist when applying those methods for this purpose in strategy research (i.e., sample sizes, data noise, and construct complexity). We then outline a theory-driven approach to help scholars overcome these challenges and improve data handling and the subsequent application of ML techniques in strategy research. We demonstrate the efficacy of our approach by applying it to create a linguistic measure of CEOs' motivational needs in a sample of S&P 500 firms. We conclude by describing steps scholars can take after creating ML-based measures to continue to improve the application of ML in strategy research.

This paper aims to automatically plan sequence for complex assembly products and improve assembly efficiency.

An assembly sequence planning system for workpieces (ASPW) based on deep reinforcement learning is proposed in this paper. However, there exist enormous challenges for using DRL to this problem due to the sparse reward and the lack of training environment. In this paper, a novel ASPW-DQN algorithm is proposed and a training platform is built to overcome these challenges.

The system can get a good decision-making result and a generalized model suitable for other assembly problems. The experiments conducted in Gazebo show good results and great potential of this approach.

The proposed ASPW-DQN unites the curriculum learning and parameter transfer, which can avoid the explosive growth of assembly relations and improve system efficiency. It is combined with realistic physics simulation engine Gazebo to provide required training environment. Additionally with the effect of deep neural networks, the result can be easily applied to other similar tasks.

Explainable artificial intelligence (XAI) has importance in several industrial applications. The study aims to provide a comparison of two important methods used for explainable AI algorithms.

In this study multiple criteria has been used to compare between explainable Ranked Area Integrals (xRAI) and integrated gradient (IG) methods for the explainability of AI algorithms, based on a multimethod phase-wise analysis research design.

The theoretical part includes the comparison of frameworks of two methods. In contrast, the methods have been compared across five dimensions like functional, operational, usability, safety and validation, from a practical point of view.

A comparison has been made by combining criteria from theoretical and practical points of view, which demonstrates tradeoffs in terms of choices for the user.

Our results show that the xRAI method performs better from a theoretical point of view. However, the IG method shows a good result with both model accuracy and prediction quality.