Samti Ahmed1, 2, Semeh ben salem1, 2, Sami Naouali1, 2, 1Sciences and Technologies for Defense (STD), Military Academy of Fandouk Jedid, Tunisia, 2Military Research Center (MRC), L’Aouina Military Base, Tunisia
The explosion of social network activity in recent years has led to massive volumes of user data, including status updates, posts, blog articles, forum entries, recommendations, login requests, and suggestions. This has given rise to new topics, including social media analytics and social network analysis. Analyzing online data to uncover terrorist trends is an essential task. It not only aids in comprehending terrorist events, including the actors, communities, methods, and operational tactics involved but also assists in predicting future attacks. However, this process remains challenging and error-prone, as terrorist events often deviate from conventional attack patterns. This paper introduces a scoring model, the Keyword Feature Score (KFS), for collecting data from social networks. The KFS model aims to assist investigators in conducting focused and specific analyses. Researchers can employ the KFS model to score and categorize suspicious comments related to homeland security within the Online Social Network (OSN) dataset, thereby further strengthening the model’s robustness.
Text scoring, natural language processing, Text classification, open source intelligence, social Network, homeland Security, text mining.
Hanna Abi Akl, Data ScienceTech Institute, 4 Rue de la Collégiale, 75005, Paris, France
We introduce SHADOW, a fine-tuned language model trained on an intermediate task using associative deductive reasoning, and mea- sure its performance on a knowledge base construction task using Wiki- data triple completion. We evaluate SHADOW on the LM-KBC 2024 challenge and show that it outperforms the baseline solution by 20% with a F1 score of 68.72%.
Knowledge Graphs, Large Language Models, Ontologies.
Lishuo Tao1, Samuel Silverberg2, 1La Jolla Country Day School, 9490 Genesse Ave, San Diego, CA 92037, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
This paper explores the development and evaluation of a mental health tracking app designed to monitor mood patterns and provide personalized support [1]. The app integrates AI technology to offer real-time guidance and recommendations based on user inputs, while a calendar feature visualizes mood trends over time [2]. We conducted two experiments to assess the accuracy of the AI responses and the effectiveness of the calendar in capturing mood patterns, finding both to be effective, although improvements in empathy and user engagement are needed. By comparing our approach with other mental health apps, we demonstrate the apps unique strengths in offering a tailored, interactive experience. Limitations such as reliance on user participation and AI empathy were identified, but proposed enhancements suggest potential for improved functionality. Our app presents a promising tool for mental health management, blending technology and self-reflection to foster better emotional well-being [3].
Mental Health Tracking, AI Integration, Mood Patterns, Mobile Health (mHealth), Personalized Support.
Ebtesam Hussain Almansour, Computer Science Applied collage, Najran University, Najran 66462, Saudi Arabia
The growing acceptance of social media networks as a platform to share opinions on several feature semerged opinion mining or sentiment analysis (SA) as an active investigation part. In recent times, SA has attracted significant attention owing to its various applications in different features of our lives. SA is one of the Natural Language Processing (NLP) that purposes to analyze and process data that is transcribed in human languages. Even though the Arabic language is the most extensively spoken language utilized for content sharing through social media, the SA on Arabic content is restricted owing to numerous challenges with the language’s morphologic structures, the dialects variabilities, and the absence of the proper corpora. In recent times, deep learning (DL) and machine learning (ML) have demonstrated extraordinary achievements in the field of SA for Arabic tweet classification in social media platforms. In this manuscript, we design and develop an Integrated Deep Learning with Natural Language Processing Models for Sentiment Analysis and Classification (IDLNLPM-SAC) technique. The IDLNLPM-SAC model presents a sentiment analysis and classification using Arabic tweets. The presented IDLNLPM-SAC model follows different levels of data preprocessing to transform the raw Arabic tweet data into a compatible format. For the process of word embedding, the latent semantic analysis (LSA) technique can be deployed. Besides, the hybrid of parallel temporal convolutional network–gated recurrent unit (PTCN-GRU) classifier can be implemented for the classification process. Eventually, the parameter choice of the PTCN-GRU algorithm can be implemented by the design of the improved marine predator algorithm (IMPA). The simulation evaluation of the IDLNLPM-SAC technique takes place using the Arabic tweets database. The experimental results pointed out the heightened solution of the IDLNLPM-SAC technique compared to recent approaches.
Sentiment Analysis; Deep Learning; Arabic Tweet; Latent Semantic Analysis; Marine Predator Algorithm.
Jiadong Gu1, Moddwyn Andaya2, 1Bellevue high school, 9235 NE 25th street, Clyde Hill, WA 98004, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
My research focuses on a gamified approach to teaching financial education, targeting students aged 10-18 [1]. I outline a method involving a simulated retail environment where players manage a retail store, enabling them to understand economic concepts through interactive gameplay. In my method analysis, I discuss three key algorithms: ordering items, tracking sales, and evaluating business performance. Each algorithm incorporates real-time data and complex calculations to simulate realistic retail operations, such as inventory management and sales probabilities based on customer foot traffic, location premium, and time of day. I aim to assess student preferences for this gamified learning model compared to traditional platforms like textbooks or Khan Academy, using a system of surveys to gather demographic information and feedback that provided excellent and satisfactory results [2]. The issue of financial literacy is urgent, highlighting statistics that reveal a significant knowledge gap in our youth.
Financial education, Gamified education, Retail simulation.
Michael Jin , Andrew Park,1Lexington High School, 251 Waltham Street, Lexington, MA 02421,2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
This project addresses the challenge of simulating rocket landings across different planetary environments by using Unity ML-Agents to train AI models [1]. The reusability of rockets, critical for space exploration, requires precise control and adaptability to varying gravitational conditions. We proposed a solution combining AI-driven controls with interactive user input to create a flexible and realistic rocket landing simulator. The methodology incorporated machine learning to train models for complex control tasks, applying reinforcement learning to adjust for Earth, Mars, and Moon environments. Our experiments focused on testing the model’s adaptation to these environments and assessing how rocket parameters like mass and thrust affected performance [2]. The most significant finding was that the AI performed well on Earth and the Moon but required further tuning on Mars due to faster descent speeds [3]. Our approach provides an engaging and educational platform for studying reusable rocket technology, making it a valuable tool for both academic and practical applications.
Machine Learning, Rockets, Landing, Reinforcement Learning.
Mesay Moges Menebo1,1Associate professor, University of Southeastern Norway, Campus
Background Influenza presents a significant public health challenge globally, with recurrent seasonal outbreaks straining healthcare systems, particularly during peak seasons. Internet search data has emerged as a valuable source for real-time forecasting of influenza trends, offering potential improvements over traditional surveillance systems. This study aimed to assess the effectiveness of using Google Trends search query data, alongside influenza-like illness (ILI) incidence, to forecast influenza trends in Norway using various machine learning models.
Eileen Weiyun Ho1, Armando Contreras2, 1Lexington High School, 251 Waltham St, Lexington, MA 02421, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
Indicator organisms such as Escherichia coli (E. coli) are vital for monitoring microbiological water quality [16]. However, current testing methods are reactive, which may cause delays in reporting E. coli levels after contamination. This can make timely interventions difficult, especially in locations lacking in testing infrastructure. Our proposal involves the creation of a machine learning-based algorithm and application that predicts and displays microbiological water quality and any potential infractions. Our research examined correlations between E. coli levels, date, and temperature. We found that E. coli levels peaked in July, modeled by an exponential trendline; temperature showed a strong correlation, likely due to its influence in the other variables. We also validated our app s predictions of E. coli levels using data from the Massachusetts Department of Public Health (MDPH) data. Our application had an average prediction difference of 2 units across 50 locations. These findings suggest reliable, real-time water safety information. Through machine learning, our application aims to provide proactive insights into water quality to enhance public health and safety.
Pathogenic Bacteria Prediction, Water Quality Prediction, Machine Learning, Environmental Health and Safety.
Juran Liu1, Moddwyn Andaya2,1Sage hill school, 65 Longchamp, Irvine, CA 92602,, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
During the long-term process of doing this tennis project, I had created and added many different features into the game, including different types of game modes, competitive tournament map, and music and sounds. One of the most significant features is the AI opponent with three different levels of difficulties for all the three game modes [1]. I designed these three game modes because of the challenges that I faced when I was trying to teach and communicate with children with autism, and that I think it might be a good idea for them to enjoy and relax themselves [2]. My idea of this project appeared in my mind after I was done with my volunteering event, I really wanted to help them out because of the situations that they are having in their lives. From the experiment that I did, I did find some inaccuracy of the swings in the game because sometimes when I swing my arm in front of the camera, the pose estimation didn’t really capture it. However, the data that I collected in this experiment tells me that I just need a little improvement on the calculations of the motion capturing system. People should start to try playing my game, especially the kids with autism, because the game could be played through a projector which allows the kids to experience the realism of the game of tennis and it builds up important social skills for them and provides a bunch of benefits to their lives [3]. This project is a game that could slowly help them to build up a better understanding for them about the game of tennis, which could bring benefits to their physical health while having fun playing the tennis mini games.
Autism-Friendly Gaming, Adaptive AI Opponents, Physical Engagement, Social Skill Development.
Jiahao Lai,Hangzhou Jianxiake Technology Co., Ltd.Hangzhou, China
High-dimensional data often presents challenges in visualization and interpretation due to its complex structures and intricate relationships. Traditional dimensionality reduction techniques, such as PCA and t-SNE, often struggle to preserve the topological features of such data, leading to the loss of critical structural information. To address this, we propose a novel dimensionality reduction technique rooted in topological data analysis, which aims to maintain the intrinsic topological structure while mapping the data into a lower-dimensional space. Our approach extracts persistent homology groups and critical points as topological features, ensuring their invariance in the reduced representation, and optimizes the bottleneck distance using a mapper-based skeleton. We demonstrate the effectiveness of our method on complex real-world datasets, showcasing its ability to uncover meaningful structures that are often overlooked by traditional methods.
Topological Data Analysis, Dimensionality Reduction, Persistent Homology, Mapper Algorithm.
Sean Li1, Andrew Park2, 1Sage Hill School, 20402 Newport Coast Dr, Irvine, CA 92603, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
Visual impairments are an unfortunate trait that can be obtained during birth, unexpected incidents, or natural causes like age. These reasons shouldn’t be the determining factor of whether the individual can remain physically healthy or not. Our goal is to provide an assistant to improve fitness journeys of those who need guidance on form and varieties of exercises. We intend to solve these problems with a dedicated fitness device that has voice guidance and a built-in camera to help lead users into building good exercise habits to prevent the risk of injury and help them remain in their best possible shape [2]. Some key technologies would be the Raspberry Pi, USB speaker, a mini microphone, and the use of AI [3]. Some challenges that we faced would be how to make the AI comprehend the words clearly and ignore outside noises. We fixed this issue with the use of OpenCV and MediaPipe. This idea is ultimately something that people should use because it provides the tools to a successful physical transformation or just daily routines to keep someone in shape, giving the visually impaired a key to a more convenient life [4].
Computer Vision, Accessibility, Fitness, Voice Command.
Yutong Zhang1, Ang Li2, 1Sage Hill School, Newport Coast, CA 92657 ,2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
In the realm of home safety, the heightened risk of injury among unsupervised children, particularly from window-related falls, represents a significant challenge [1]. This study introduces an innovative solution to mitigate such risks through a novel integration of technology and artificial intelligence. We propose a comprehensive system that harnesses the power of Deep Learning, specifically utilizing the YOLOv8n algorithm, in conjunction with a Raspberry Pi platform for real-time hazard detection [2]. To address the critical aspect of data transfer, our system employs Firebase for efficient and timely communication between components. Acknowledging the limitations posed by initial model inaccuracies, our approach involved augmenting our dataset to ensure diversity and mitigate the risk of overfitting, thereby enhancing the models predictive accuracy. This paper details our experimentation with various configurations, including an attempt to utilize YOLOv8x, which was ultimately revised to YOLOv8n due to computational constraints of the Raspberry Pi [3]. The robustness of our system was rigorously tested across diverse scenarios involving windows and doors, establishing a comprehensive dataset that underscores the systems effectiveness. By integrating real-time detection with an intuitive user interface, our system offers a proactive tool for parents to enhance home safety for their children. This contribution not only addresses a pressing societal issue but also advances the application of Deep Learning and IoT technologies in the domain of domestic safety.
YoloV8n, Object detection and recognition, Raspberry PI, OpenCV, Picamera2.
Khaoula Marhane, Fatima Taif, Mohamed Azzouazi, Faculty of science Ben M’sik University, Casablanca, Morroco
Universities Risk is the probability of an unwanted activity occurring causing loss or harm to the organization. Managing risk includes identification, assessment, impact evaluation, acceptance and mitigation. In today’s well dis-tributed and diversified organizations, risks exist in virtually every part or function of the organization. Increasingly it is being recognized that one central risk management team may not be in the best of position to identify, assess and manage all the risks across the organization. The integration of blockchain technology into collaborative risk management has gained significant attention in recent years, reflecting an increasing recognition of its potential to enhance transparency, security, and efficiency. The literature reveals a progressive exploration of blockchain applications across various sectors, particularly in public governance and supply chain management. sThis paper proposes the idea of Collaborative Risk Management where in the stakeholders across the organization, collaborate seamlessly using blockchain technology to determine and manage risks on an ongoing manner. in addition, we present a proof of concept implementation using Hyperledger Fabric.
Risk management, Blockchain, Hyperledger Fabric, Stakeholder.
Copyright © MLTEC 2024