Inteligencia Artificial

Learning Terrain Traversability for a Mobile Robot based on Information Fusion

Sat, 19 Oct 2024 00:00:00 +0200

In this work, we propose an approach to determine terrain traversability for a car-like robot. Our approach has two main modules: a neural network classifier that makes use of sensors' readings to assign traversability levels to control inputs of the robot, and a second neural network that, based on the outputs of the first network, mimics the control selection performed by a human driver. The approach incorporates sensor fusion from a variety of sources to enhance the traversability estimation, and it is trained employing a semi-supervised learning scheme with examples resulting from the interaction of the car with the environment. This semi-supervised scheme avoids exhausting manual labeling and is built on the premise that there is a correlation between the terrain traversability and the required and observed behaviors of the vehicle. The method is validated with data obtained from a physical electric car.

Real time distracted driver detection using Xception architecture and Raspberry Pi

Sat, 19 Oct 2024 00:00:00 +0200

Researchers are concentrating on developing technologies to identify and caution drivers against driving while distracted because it is a major cause of traffic accidents. According to the National Highway Traffic Safety Administrator's report, distracted driving is to blame for roughly one in every five car accidents.Our goal is to create an accurate and dependable method for identifying distracted drivers and alerting them to their lack of focus. We take inspiration from the success of convolutional neural networks in computer vision to do this. Our strategy entails putting in place a CNN-based system that can recognize when a driver is distracted as well as pinpoint the precise cause of their preoccupation. Real-time detection, however, necessitates three apparently mutually exclusive requirements for an optimal network: a small number of parameters, high accuracy, and fast speed.

ALHK: Integrating 3D Holograms and Gesture Interaction for Elementary Education

Abeer Hakeem, Hind Bitar, Ayman Alfahid — Sat, 19 Oct 2024 00:00:00 +0200

The integration of technology in elementary education offers innovative ways to enhance learning. One such advancement is the use of three-dimensional holograms (3DH), which provide immersive displays that merge seamlessly with the learner’s environment, creating a dynamic and engaging atmosphere. Educators have found that 3D visual tools significantly improve student comprehension, with 94.4% agreeing in a preliminary study. However, using interactive 3D holography alone has limitations, such as the inability for students to physically touch or manipulate holographic objects. To address this, Active Learning with Holo-Kid (ALHK) is introduced as a desktop application for elementary school students (grades 1 to 6). ALHK combines Leap Motion technology’s precision with interactive 3D holography to overcome these limitations. The combination allows students to interact with virtual objects in a more immersive and realistic manner. Holograms provide visual representation, while Leap Motion enables precise gesture recognition and hand tracking, resulting in a seamless and intuitive user experience. Initial evaluations demonstrate improved student engagement and comprehension. Future iterations aim to enhance scalability by incorporating features like custom object upload, multi-user interaction, and broader age applicability. ALHK shows promise as a tool for creating an immersive and intuitive learning environment using 3D holograms and interactive technology in elementary education.

Semantic Alignment in Disciplinary Tutoring System: Leveraging Sentence Transformer Technology

Rosana Abdoune, Lydia Lazib, Farida Dahmani-Bouarab, Nada Mimouni — Sat, 19 Oct 2024 00:00:00 +0200

In this work, we present a disciplinary e-tutoring system that integrates ONTO-TDM, an ontology designed for teaching domain modeling, with advanced transformer technology. Our primary objective is to enhance semantic similarity tasks within the system by fine-tuning a Sentence Transformer model. By carefully adjusting training parameters with a curated dataset of question-answer pairs focused on algorithms and data structures, we achieved a notable improvement in system performance. The Sentence Transformer model, combined with domain ontology, achieved an accuracy of 91%, a precision of 93%, a recall of 89%, and an F1-score of 90%, significantly surpassing the results of existing works. This methodology highlights the potential to deliver personalized support and guidance in tutoring scenarios. It effectively addresses the evolving needs of modern education by offering tailored answers and reducing the necessity for constant learner-tutor interaction, thereby improving the efficiency of educational support systems.

Breast Cancer Classification Using Gradient Boosting Algorithms Focusing on Reducing the False Negative and SHAP for Explainability

João Manoel Herrera Pinheiro, Marcelo Becker — Wed, 04 Dec 2024 00:00:00 +0100

Cancer is one of the diseases that kill the most women in the world, with breast cancer being responsible for the highest number of cancer cases and consequently deaths. However, it can be prevented by early detection and, consequently, early treatment. Any development for detection or perdition this kind of cancer is important for a better healthy life. Many studies focus on a model with high accuracy in cancer prediction, but sometimes accuracy alone may not always be a reliable metric. This study implies an investigative approach to studying the performance of different machine learning algorithms based on boosting to predict breast cancer focusing on the recall metric. Boosting machine learning algorithms has been proven to be an effective tool for detecting medical diseases. The dataset of the University of California, Irvine (UCI) repository has been utilized to train and test the model classifier that contains their attributes. The main objective of this study is to use state-of-the-art boosting algorithms such as AdaBoost, XGBoost, CatBoost and LightGBM to predict and diagnose breast cancer and to find the most effective metric regarding recall, ROC-AUC, and confusion matrix. Furthermore, previous studies have applied Optuna to individual algorithms like XGBoost or LightGBM, but no prior research has collectively examined all four boosting algorithms within a unified Optuna framework, a library for hyperparameter optimization, and the SHAP method to improve the interpretability of our model, which can be used as a support to identify and predict breast cancer. We were able to improve AUC or recall for all the
models and reduce the False Negative for AdaBoost and LigthGBM the final AUC were more than 99.41% for all models.

An Efficient Deep Learning Technique for Brain Abnormality Detection Using MRI

Shilpa Mahajan, Anuradha Dhull, Aryan Dahiya — Wed, 04 Dec 2024 00:00:00 +0100

This research proposes an effective and reliable deep learning method for detecting brain abnormalities via magnetic resonance imaging (MRI). The technique consists of two primary stages: first, a binary classifier that divides pictures into "Brain" and "Non-Brain" categories; second, multi-class classifiers that explicitly recognise categories such pituitary adenomas, gliomas, and meningiomas. The labelled and preprocessed data were taken from a collection of 7,753 pictures provided by Qhills Technologies Pvt. Ltd. Additional data from the Brain Tumour MRI collection was also incorporated to improve the model's generalisation skills. VGG-16 outperforms the other machine learning models, with an accuracy rate of 96.4%, when compared to ANN, CNN, VGG-16, and AlexNet. A thorough model evaluation and hyperparameter tweaking process was conducted using the accuracy, precision, recall F1-score. The findings of this study point to the potential of deep learning techniques in identifying brain disorders fast and precisely, opening the door to more precise diagnosis in clinical settings.

Artificial Intelligence (AI) in Dental Age Estimation Studies: A Scientometric Analysis

Muhammad Zaid Zainuddin, Azliyana Azizan , Mohd Yusmiaidil Putera Mohd Yusof — Mon, 13 Jan 2025 00:00:00 +0100

Dental age estimation (DAE) is important in age-related studies ranging from forensics, clinical dentistry and bioanthropology. DAE heavily relies on image analysis and morphometrics and has underwent academic scrutiny to improve its level of reliability and accuracy. The recent rise of artificial intelligence (AI) in data analysis allows accurate analysis without the influence of human error. As AI has penetrated DAE research, there is a lack of scientometric analysis regarding AI-driven DAE studies. This scientometric study presents an analysis of AI-driven DAE research based on data from the Scopus and Web of Science literature databases. This study examines various parameters, such as publication trends, prolific countries and research institutions, active journals and highly cited publications as well as highly used keywords pertaining to AI-driven DAE studies. Notably, though the niche area is fairly recent, there has been a substantial increase in the number of publications in AI-driven DAE research in the past few years. Countries such as China, Malaysia and South Korea are currently at the forefront of publications on the application of AI in DAE studies. This study also finds that a variety of journals ranging from dentistry, law, forensics and computer science are publishing studies on AI-driven DAE. Prominent keywords such as “age estimation”, “artificial intelligence” and age-group related keywords were amongst the dominant keywords used. This scientometric analysis provides an overview of studies pertaining to AI-driven DAE, which serves to help researchers stay informed regarding the latest research trend and may help identify possible research gaps.

Two competitive hybridization approaches based on combining of Giza Pyramids Construction with Particle Swarm Optimization for solving global optimization problems

Sasan Harifi, Faraz Davachi, Narges Mohammadi, Soheil FaridMohammadzadegan — Mon, 13 Jan 2025 00:00:00 +0100

Optimization problems are complex problems that are very difficult to solve. Although these types of problems are solved in the real world using exact methods, these methods are very time-consuming and costly. By using soft computing methods, the time and cost of problem-solving can be reduced to some extent. Engineering problems are among the complex real-world problems that can be solved through soft computing methods. One of these methods is the use of metaheuristic algorithms to optimize the solution of these types of problems. The Particle Swarm Optimization (PSO) algorithm is a common and state-of-the-art metaheuristic algorithm used to solve engineering optimization problems. This algorithm is known as swarm-based optimization techniques and has a very powerful mathematical basis. Another recently published algorithm is the Giza Pyramids Construction (GPC) algorithm. The GPC algorithm models the technological advancements of construction in ancient times. Both algorithms have many advantages through which optimization problems can be solved effectively. To increase the power of metaheuristic algorithms and solve optimization problems more effectively through them, the idea of competitive hybridization algorithms has been proposed. In this paper, two competitive hybrid approaches of combining PSO and GPC algorithms are presented. These two competitive hybridization approaches have been first applied to 45 benchmark functions and have been evaluated and analyzed statistically. Then they have been applied to six classic engineering problems. Algorithms presented in each step have been compared with Genetic Algorithm (GA) and PSO algorithms and their combined approaches. The results of experiments and statistical analysis show that the solution to engineering problems has been done more effectively by using the two proposed combinations.