September 21-22, 2024, Copenhagen, Denmark
Kamronbek Yusupov1, Md Rezanur Islam1, Insu Oh2, Mahdi Sahlabadi2, and Kangbin Yim2, 1Software Convergence, Soonchunhyang University, Asan-si, South Korea, 2Department of Information Security Engineering, Soonchunhyang University, Asan-si, South Korea
This research focuses on evaluating the security of an intrusion detection system in a CAN bus-based vehicle control network. A series of studies were conducted to evaluate the performance of models proposed by previous researchers, testing their effectiveness in real-world scenarios as opposed to those on which they were trained. The article demonstrates that models trained and tested on the same dataset can only sometimes be considered adequate. An approach that included models trained only on CAN ID, Payload, or full data was chosen. The research results show that such methods are ineffective enough in real-world attack scenarios because they cannot distinguish between new scenarios not presented during training. The results of testing the models in various attack scenarios are presented, and their limitations are identified. In addition, a new method is proposed explicitly for attack scenarios that may occur in the real-world use of an in-vehicle CAN communication system.
Intrusion Detection System, Controller Area Network, In-Vehicle Network, LSTM.
Omar Saad Almousa, Jordan University of Science and Technology, Jordan
Passwords are ubiquitous and this will continue for long. Strong passwords are a necessity to protect sensitive information. However, users not only tend to pick weak passwords, but also reuse them over several authentication systems. The existence of weak passwords in a system not only jeopardize that system, but also other systems with overlapping users because of password reuse phenomena. Investigating users’ behaviour in password creation leads to finding ways to avoid weak passwords. One aspect of that is to study the very passwords. In this study we analyse 662 passwords created by fresh students in our faculty. The students picked their passwords to authenticate themselves to a platform for programming practice and assignment solving. Our analysis relied on basic structural parameters such as password length, constructing characters, and entropy. To that end, we coined two definitions for weak and strong passwords. One is alphabet-based, and the other is entropy based. Accordingly, we found that majority of students do not tend to create strong passwords. We believe that this is due to the lack of enforcement of a strong password policy.
Passwords, Analysis, Weak password, Strong password.
Alex Chen1, Christopher Wadley2, 1Mark’s School, 25 Marlboro rd, Southborough, 2California State Polytechnic University, USA
ConnACT is a mobile application designed to connect individuals seeking sports partners [11]. It addresses the challenge of finding compatible players by considering factors such as sports preferences, skill levels, location, and any relevant medical conditions. The app uses its algorithm to match users with suitable games and partners, fostering a more inclusive and active sports community. It also offers features for creating and joining games, including options for Special Olympics athletes [12]. The development of ConnACT involved overcoming challenges related to platform limitations, geolocation integration, and user profile customization. The effectiveness of the apps game sorting and rating systems was evaluated through multiple measures of experimentation, demonstrating both functions properly, and revealing insights into how the app works [13]. ConnACT distinguishes itself from existing solutions by prioritizing inclusivity and comprehensive game-finding. This app has the potential to significantly enhance the accessibility and enjoyment of sports for individuals of all abilities, promoting physical activity, social interaction, and overall well-being.
Sports, Community, Games, Special Olympics.
Niloufar Ataeepour, University Of Quebec In Montreal, Canada
There have been great successes in recent language teaching methodologies over the past decades which still raise ambivalent questions and criticism. Language experts have been focusing to appear with different language teaching manners concerning the second language learners. While almost everyone agrees with the need of developping the oral competency, the teaching procedures used are not necessarily the same. The trend is normally to use one or other existing approaches, because choosing one unique teaching method would not be an easy task (Moore, 2000). To adress the issue and rather than not opting for particu- lar pedagogical technique, this PhD research seeks to demonstrate the necessity of combing the salient aspects of the two recent language teaching methodologies of plurilinguism ( Lau &Vari Viegen, 2020; May 2014) and the one of monolingualism (neurolingualism); ( Germain and Netton, 2012); to come up with a new enriched persepective better meeting the French second language learners need and providing them with more social authentic opportunities to advance in their oral communicative mastery.The study aims to contribute to the literature regarding sociolinguistic, critical sociology and the sociolinguistic mobility leading to language learners perception of their own learning process. Through an ethnographic study of the international students participating in French classes in Quebec / Canada, this PhD proposal will be focused on the sociolinguistic practices of these new Francophones in the context of their daily life; outside of the formal language classes; the way they will invest socially in practicing the language, challenges they might be faced with and the strategies they would apply to overcome the eventual language barriers, given their plurilinguistic repertoires.
Oral competency, FLS (French as a second language), plurilingual and monolingual (neurolingual) approaches.
Yingqi Wang, Marisabel Chang, Emma Willard School, California State Polytechnic University, USA
This research explores the development and evaluation of PlantAR, a system designed to enhance indoor spaces using augmented reality (AR) and artificial intelligence (AI) technologies [1][2]. The problem of reduced indoor air quality and psychological well-being due to a lack of greenery is addressed by providing users with personalized plant recommendations and AR visualizations [3]. The systems key components include an AI engine for real-time data analysis, a database for managing user and plant information, and a user interface that integrates AR functionality. Experiments conducted to assess AR accuracy and user engagement revealed that while the system performs well under optimal conditions, further improvements are needed to handle challenging environments and sustain user interest. Overall, PlantAR demonstrates significant potential as a tool for improving indoor environments, with opportunities for future enhancements.
Augmented Reality (AR), Artificial Intelligence (AI), Plant Recommendations, Indoor Environment, Smart Home Technology.
Hanes Oliveira, Algemetric INC, USA
Given the large size of genome files and the current importance and widespread use of DNA sequencing in medicine, sciences, and even everyday life, we introduce a transform targeted to these files. Our transform, Slot Machine (SMT), improves the statistical distribution of the nucleotide sequences stored in FASTA and Multi-FASTA files while greatly decreasing their size. This decrease in size is comparable to some of the most generic compression algorithms such as gzip. Moreover, Slot Machine allows random access, which allows a user to retrieve only the desired parts of the file. In contrast, gzip requires the entire file to be decompressed before any parts can be accessed. As an added bonus, the output of Slot Machine can be further compressed by pairing with existing compressors. Through our experiments, we achieved compression-ratios of up to 5.21, or a compression rate of 80.84%. Our results improved upon the results of NAF in every experiment and are comparable to MFCompress.
Compression, compressed pattern matching, information retrieval, dna, fasta, multi-fasta, bioinformatics.
HoiNi Yeung1, Ang Li2, 1Ulink College of Shanghai, China, 2California State Polytechnic University, USA
Sailing requires specific practice conditions, such as access to open water and expensive equipment, limiting its accessibility for many. To overcome these barriers, I developed a virtual reality (VR) sailing simulator that allows users to practice sailing techniques anywhere, regardless of weather or location [5]. The simulator is built using Unity for realistic game environments, Visual Studio for robust coding, and VR technology for immersive interaction [6]. Key challenges included accurately modeling the physics of wind dynamics and sail adjustments and ensuring realistic responses to user inputs. These challenges were addressed through detailed data analysis and iterative refinement. The simulator was tested across various scenarios to replicate real-world conditions, demonstrating its effectiveness in improving users sailing skills. This VR sailing simulator offers a cost-effective, accessible training solution, enabling both beginners and experienced sailors to practice and enhance their skills in a safe, controlled environment, making it an essential tool for sailing education and practice.
Unity, Simulation, Sailing Training, Virtual Reality.
Yifei Zhang1, Jonathan Sahagun2, 1Troy High School, 2California State Polytechnic University, USA
This paper addresses the challenge of automating household tasks, focusing on enhancing robot arm capabilities for tasks such as cloth handling and dynamic object manipulation [1]. Our proposed solution involves a robot arm equipped with advanced computer vision, angle calculation modules, and serial communication [2]. We tested the systems performance in object classification and handling dynamic environments. Challenges included inaccuracies in object recognition and adaptability issues. We addressed these by improving the vision model with more diverse training data and enhancing the arm’s mechanical and computational capabilities [3]. The experiments demonstrated the system’s potential for real-time, effective task automation [4]. Our improvements lead to a more adaptable and precise solution, making it a valuable tool for household and environmental applications. The project ultimately offers a significant advancement in automating mundane tasks, providing a practical and efficient solution for everyday use.
Robot Arm Automation, Dynamic Object Manipulation, Computer Vision Integration, Household Task Automation.
Batuhan Erdogan, Bogazici University, Istanbul, Turkey
This study examines the limitations of OpenAIs ChatGPT models (GPT-3.5 and GPT-4) in interpreting and utilizing indexicals. While GPT-4 shows some performance improvements over GPT-3.5, both models frequently misinterpret indexicals in prompts and occasionally err in producing them in specific contexts. The models abilities vary with the type of contextual environment simulated by the user, demonstrating better competence in discrete environments and conversational implicatures. ChatGPT generally avoids context-dependent language in its responses. Through word frequency analysis of four demonstrative indexicals across essays written by humans and the two GPT models, we found GPT-4 significantly more likely to produce such indexicals than GPT-3.5. Inspired by Heideggers concept of Being-in-the-World, we propose a new training method using narratives with multiple first-person perspectives within a fictional world to enhance the models handling of pronominal indexicals.
Artificial Intelligence, Pragmatics, Semantics, Linguistics, Indexicals, LLMs, Artificial Neural Networks, Philosophy of Artificial Intelligence.
Mehdi Mekni, Kaitlin I. Singer, and Candace Williams, University of New Haven West Haven, CT 06516
The rapid evolution of higher education, particularly in technology and innovation, has prompted Connecticut to leverage its education ecosystem to maintain a competitive workforce. The Connecticut Higher Education Tech Talent Accelerator (TTA) aims to meet emerging credentials needs through innovative IndustryRecognized Credentials (IRCs) and employer partnerships. Currently, no available comprehensive methodologies guide the successful integration of IRCs in curricula. To address this, our project aims to integrate Unity Technologies’ credentials in our Bachelor of Science in Computer Science with Game Design and Development concentration (BSCS-G2D) at The University of New Haven (UNewHaven)to enhance Connecticut’s game development workforce. The project’s goals include proposing a comprehensive methodology to integrate IRCs, identifying challenges, evaluating industry collaboration, and formulating a robust workforce development strategy.
Industry Recognized Credentials, Knowledge Skills Abilities, Game Design and Development.
Haoyun Yang1, Yu Cao2, 1Lutheran High School of Orange County, 2222 North Santiago Boulevard, Orange, CA 92867, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
Our application aims to assist students with class materials by providing a platform with multiple-choice and freeresponse questions for them to practice [13]. We built this application with Flutter and Firebase which provides compatibility through different mobile platforms and secure data storage [14]. To provide an effective and userfriendly interface, we allow everyone to create their own quizzes based on their focus area and explore quizzes others created. We also developed a hint system where the user can access the hint if they struggle to find the correct solution. Moreover, the user can access their recent quiz attempts to further review the concepts and materials they missed. We validated the effectiveness of the application by conducting an experiment on the user s quiz score over time given continuous practice. We also conducted another experiment on the impact of hints on users understanding of materials which concluded a mixed results in their accuracy. Throughout our tests, we proved the effectiveness of the application and its success in fulfilling its purpose.
Education, Academic, Mobile Application, Practice.
Wenya Zhang Institution of Education, University College London, London, UK
The National Assessment Program for Literacy and Numeracy (NAPLAN) Reform in Australia, launched in 2008, has emerged as the countrys most significant and contentious reform. However, due to its highstakes nature and standardization, testing presents various challenges. These challenges include the combination of accountability with the my school website, overlooking higher-order cognitive abilities, exacerbating students anxiety and stress, and creating inequity for Language Background Other Than English (LBOTE) students. This re-port assesses the achievements and obstacles of the NAPLAN reform, pro-posing recommendations such as transitioning to online testing, enhancing content and platforms, increasing public assessment literacy, and investing more in LBOTE education. These suggestions aim to strike a balance be-tween standardized testing and authentic educational pursuits, adapting to the evolving needs of students to create a fair, inclusive educational environment that addresses the demands of the 21st century.
NAPLAN, High Stakes Assessment, Accountability, Education Policy .
Marcus Schoenharl1, Ang Li2, 1BBIS Berlin Brandenburg International School, Schopfheimer Allee 10, 14532 Kleinmachnow, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
Physics education often struggles with student engagement and limited access to hands-on experimentation, which are essential for deep understanding. Physics Playground addresses these challenges by integrating Artificial Intelligence (AI) and Augmented Reality (AR) into an interactive, personalized learning platform [1][2]. The platform leverages AI to deliver customized quizzes and instant feedback, while AR provides immersive virtual experiments, allowing students to explore physics concepts in a dynamic environment. Key challenges, such as integrating AI and AR seamlessly and ensuring user-friendliness, were addressed through iterative design, extensive user testing, and modular development. Experimental results demonstrated that Physics Playground significantly enhances learning outcomes and user satisfaction compared to traditional teaching methods, with marked improvements in engagement, comprehension, and retention [3][4]. By offering an innovative blend of AI-driven personalization and AR-based interactivity, Physics Playground provides an effective, scalable solution for diverse educational settings, making it a valuable tool for advancing physics education globally.
Interactive Physics Education, Augmented Reality Learning, AI-Powered Quizzes, Virtual Physics Experiments.
Yuan Cheng1, Tyler Boulom2, 1Fairmont Preparatory Academy, 2200 West Sequoia Ave, Anaheim, CA 92801, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
Traditional drumming education often involves high costs, noise constraints, and limited access to skilled instructors, making it difficult for many to learn. "Virtual Drummer VR" addresses these challenges by integrating Virtual Reality (VR) and Artificial Intelligence (AI) to create an immersive, interactive drumming experience [1]. The program combines VR environments with AI-driven performance analysis to provide real-time feedback, personalized learning paths, and a comprehensive educational framework [2]. Key challenges included optimizing the software for different hardware and reducing latency, which were overcome using advanced algorithms and scalable cloud computing. Experimental results demonstrated that this approach significantly enhances user engagement, skill acquisition, and retention. Participants showed marked improvement in drumming proficiency and reported high satisfaction levels. "Virtual Drummer VR" offers an innovative, accessible, and scalable solution for drumming education, making it a valuable tool for learners of all skill levels, from beginners to advanced drummers, by providing a modern, effective approach to music learning.
Virtual Reality Drumming, AI-Driven Music Education, Real-Time Performance Feedback, Immersive Learning Environments.
Sükran Sungur and Gülbin ÖzkanDepartment of Mathematics and Science Education, Uludag University, Bursa, Turkiye
The purpose of this study is to determine the opinions of pre-service science teachers about web-based teaching and distance education. A case study was carried out with undergraduate science teacher students (n=15) studying at a state university in Istanbul. The study was carried out through Material Design in Science Teaching lesson. Students took this course for 12 weeks and at the end of this course student opinions about web-based teaching and their opinions about distance education by moving from the experiences of students during the pandemic were received. Examining all the data reveals that while pre-service science teachers have many favorable opinions of web-based learning, they have few favorable opinions about distance education. Since students work with web 2.0 tools, they stated the advantages and disadvantages of using these tools in science education. Students suggestions regarding web-based and distance education will contribute to future studies about web-based and distance education.
Web-based teaching, distance education, science teaching, pre-service science teaching .
Simon M¨unker, Kai Kugler, and Achim Rettinger
Filtering and annotating textual data are routine tasks in many areas, like social media or news analytics. Automating these tasks allows to scale the analyses wrt. speed and breadth of content covered and decreases the manual effort required. Due to technical advancements in Natural Language Processing, specifically the success of large foundation models, a new tool for automating such annotation processes by using a text-to-text interface given written guidelines without providing training samples has become available. In this work, we assess these advancements in-the-wild by empirically testing them in an annotation task on German Twitter data about social and political European crises. We compare the prompt-based results with our human annotation and preceding classification approaches, including Naive Bayes and a BERT-based fine-tuning/domain adaptation pipeline. Our results show that the prompt-based approach – despite being limited by local computation resources during the model selection – is comparable with the fine-tuned BERT but without any annotated training data. Our findings emphasize the ongoing paradigm shift in the NLP landscape, i.e., the unification of downstream tasks and elimination of the need for pre-labeled training data.
foundation models, automating text annotation, zero-shot classification, social and political EU crises.
Xiaotian Zhu1, Tongchen He2, 1Westtown School, 975 Westtown Rd, West Chester, PA 19382, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
This app aims to expand teenager’s awareness with common health problems and provide a platform for users to seek guidance from configured AI [4]. FlutterFlow is the service we use to build the app where users access the information pages and the survey feature. On the “Survey question” page, users are able to fill out the questions. Then all results would then be sent to GPT using the python code which would guide AI on how to generate the answer [5]. Then the answer would be generated and sent back to FlutterFlow for users to view on the “Result” page. And we use Firebase to store information of the “Topic” and “survey questions” [6]. My ultimate goal for this app is to be a secure platform where patients will get immediate diagnosis from professionals when there is an emergency, or they face trouble seeking help. And there will be a chat feature for patients to send out more detailed information including pictures.
Teenager Health, Generative AI, Mobile App, Medical Suggestions.
Lezhi Wu1, Tann Nguyen2, 1Basis International School Nanjing, No. 18, Lingshan North Road, Qixia District, Nanjing,Jiangsu, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
This program aims to solve the problem for those people who find it hard to comprehend and get the necessary information they need when reading difficult and long articles online [1]. The program is a Google extension available to everyone and can summarize the articles users find on all the websites into bullet points, providing a specific explanation when needed [2]. The program contains three important systems, including prompt engineering, flask server, and Google extension. When the user clicks the summarize button, the text on the website will be sent to Chat GPT, and it will summarize the article based on a specific prompt we set [3]. The server allows the program to respond to users’ requests, and Google extensions provide access to end users. We conducted an experiment that tested the accuracy of the summary on 20 different websites and got an average accuracy of 8.35 out of 10. We found out that the program works well for long articles that usually are harder to read, accomplishing our goal to help people read long and difficult articles and get information efficiently.
Article Summarization, Google Extension, Prompt Engineering, User Comprehension.
Maayeesha Farzana and Md. Shafiul Alam Forhad, Department of Computer Science & Engineering, Chittagong University of Engineering & Technology,Chittagong 4349
Due to the increase in expressing thoughts on communication platforms online, partic-ularly within the Bengali community, instances of hate speech are becoming increasingly common.While text analysis methods have been extensively studied, audio analysis remains relatively un-explored due to the scarcity of Bengali audio speech resources. This paper endeavors to fill this gap by constructing an audio hate speech dataset sourced from publicly available platforms like YouTube, aiming at a binary class classification of hate speech. Our approach involves analyzing crucial audio features for hate speech detection. We assessed Support Vector Machine (SVM),K-Nearest Neighbors (KNN), Logistic Regression (LR), Random Forest (RF), and Convolutional Neural Network (CNN) models for classifying audio features. Our research underscores the impor-tance of addressing hate speech detection across various modalities and emphasizes the necessity of robust datasets to facilitate model training and evaluation.
Spoken Language Processing, Hate Speech, Deep Learning, Machine Learning, Ben-gali Natural Language Processing, Audio Analysis, Computational Linguistics.
Denis B´echet1 and Annie Foret2, 1Nantes University, France, 2Univ. Rennes and IRISA, France
We consider the family of Categorial Dependency Grammars (CDG), as computational grammars for language processing. CDG are a class of categorial grammars defining dependency structures. They can be viewed as a formal system, where types are attached to words, combining the classical categorial grammars’ elimination rules with valency pairing rules that are able to define non-projective (discontinuous) dependencies. Whereas the problem of closure under iteration is open for the original version of CDG, we define “CDG extended with barriers”, an extended version of the original CDG, that solves this formal issue. We provide a rule system and we show that the extended version defines an Abstract Family of Languages (AFL), while preserving advantages of the original CDG, in terms of expressivity, parsing and efficiency.
Logical approach to natural language, Type calculus, Categorial Gram- mar, Dependency Grammar, Abstract Family of Languages.
Xuxin Wang1, Fangrui Guo2, 1Harvard-Westlake School, 3700 Coldwater Canyon Ave, Los Angeles, CA 91604, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
This research addresses the challenge of providing an engaging and effective Latin language learning experience by leveraging advanced AI technologies [1]. The proposed solution involves developing an educational app that integrates the GPT-4o Mini model for personalized learning, combining language instruction with cultural insights [2]. Key components include a Flask-based backend, AI-driven content generation, and interactive games designed to enhance user engagement [3]. Challenges such as balancing AI performance with response time were tackled by selecting GPT-4o Mini for its optimal mix of speed and accuracy. Experiments confirmed the models effectiveness across various benchmarks, demonstrating its superior performance and cost-efficiency. The app offers a comprehensive and immersive learning experience, making it a valuable tool for modern learners interested in classical studies [4]. This approach not only improves traditional methods but also sets a new standard for language education by integrating advanced AI and interactive elements.
Literature Education, Nature Language Processing, Large Language Model, Machine Learning.
Ziwei Mi1, Ang Li2, 1Sandy Spring Friends School, 16923 Norwood Rd, Sandy Spring, MD 20860, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
The "History Around You" app is designed to make historical knowledge accessible and engaging by integrating AI-generated content and user contributions [1]. The project addresses the issue of declining interest in history by providing an interactive platform where users can explore and contribute to historical narratives. The app leverages ChatGPT to generate brief historical articles based on user inputs, while Google Maps integration allows for an intuitive exploration of historical sites [2]. Experiments conducted revealed that while ChatGPT is a powerful tool for content generation, it occasionally produces inaccuracies, necessitating additional validation mechanisms. Additionally, user-generated content significantly boosted engagement, though it introduced variability in content quality [3]. To mitigate these challenges, the app could benefit from improved moderation systems and offline capabilities. Overall, the "History Around You" app shows great potential as an educational tool, making history more accessible and interactive for users worldwide.
Historical Knowledge, AI-generated Content, Educational Technology, Mobile Application, Google Maps Integrati.
Morteza Sadeghi and Abdolreza Torabi, Department of Engineering Science, University of Tehran, Tehran, IRAN
The Multilevel Fast Multipole Algorithm (MLFMA) has known applications in scientific modeling in the fields of telecommunications, physics, mechanics, and chemistry. Accelerating calculation of far-field using GPUs and GPU clusters for large-scale problems has been studied for more than a decade. The acceleration of the Near Field Computation (P2P operator) however was less of a concern because it does not face the challenges of distributed processing which does far field. This article proposes a modification of the P2P algorithm and uses performance models to determine its optimality criteria. By modeling the speedup, we found that making threads independence by creating redundancy in the data makes the algorithm for lower dense problems nearly 13 times faster than non-redundant mode.
Multilevel Fast Multi-Pole Algorithm, Graphics Processors, Performance Evaluation .
Yan Lee1, Jonathan Sahagun2, 1Arnold O. Beckman High School, 3588 Bryan Ave, Irvine, CA 92602, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
This paper addresses the critical need for personal air quality monitoring, focusing on CO2 exposures impact on health [1]. Current solutions often lack portability and user-specific recommendations [2]. We propose CAIR, an electronic necklace with a mobile app, equipped with a CO2 sensor and microcontroller, to provide real-time air quality data and personalized health advice. Key technologies include advanced sensors, microcontrollers, and secure mobile communication. Challenges such as sensor accuracy and data privacy were mitigated through calibration and encryption. Applied in various scenarios, CAIR demonstrated reliable performance in monitoring air quality and providing actionable insights [3]. Our results indicate that CAIR effectively empowers users to manage their exposure to CO2, making it a valuable tool for enhancing personal health and safety in diverse environments. This innovative approach promotes greater awareness and proactive management of air quality, ensuring better health outcomes.
Carbon dioxide, Personal health, Air quality, Wearable.
Camus Hu1, Yu Sun2, 1Fairmont Preparatory Academy, 2200 W Sequoia Ave, Anaheim, CA 92801, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768,USA
Wristwell presents a novel approach to environmental monitoring, offering real-time insights into air, water, and noise pollution levels [1]. This paper discusses the development, implementation, and potential of Wristwell in addressing critical environmental challenges [2]. We explore the systems architecture, sensor integration, data processing algorithms, and user interface design. Additionally, we examine the challenges and limitations faced by Wristwell, including sensor accuracy, network connectivity, and scalability issues. We propose strategies to mitigate these challenges, such as sensor calibration, offline capabilities, and user experience enhancements. Through these efforts, Wristwell aims to provide accurate, reliable, and actionable pollution data to support informed decision- making and promote environmental sustainability.
Environmental Monitoring Systems, Real-Time Pollution Insights, Sensor Integration and Data Processing.
Paul Knowles1, Bart Gajderowicz2, 1The Human Colossus Foundation, Geneva, Switzerland, 2Industrial Engineering Department, University of Toronto, Toronto, Ontario, Canada
This paper introduces a foundational Informatics domain model, a comprehensive data-centric design paradigm that addresses the challenges and limitations of current computational system designs and models. The Informatics model shifts the focus from node-centric to data-centric categorization, leveraging a multimodal framework of objects, events, concepts, and actions. Developed through rigorous research, the model draws on diverse disciplines, relying primarily on precise definitions from the Oxford Dictionary. It provides a foundational reference for system designers and data architects, enabling secure role-based access solutions and promoting semantic interoperability. The accompanying core data ontology enhances knowledge representation and fosters consistent understanding across distributed data ecosystems. The paper discusses the design of the ontology and presents the ontology in OWL 2. The paper also discusses the applications and benefits of the Informatics model, highlights its scalability, and outlines future directions for research. Readers may explore the Informatics Domain Model presentation deck for a more visual representation of the model. This paper provides a frame of reference and catalyst for adopting the Informatics model as a transformative approach in various domains and industries.
Informatics model, distributed data ecosystems, cryptographic data security, semantic interoperability, ontology design.
Arlene Chang1, Jonathan Sahagun2, 1Northwood High School, 4515 Portola Pkwy, Irvine, CA 92620, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
Sign language, the main language used by Deaf individuals, has been historically suppressed in favor of speech and oralism. Most hearing people also do not know sign language, thus creating a linguistic and cultural gap between Deaf and hearing communities today. This application proposes an American Sign Language (ASL) recognition model, consisting of a two-way translation assistant. One function uses machine learning to detect ASL handshapes to translate signs into English while the second function uses motion capture techniques to translate written English into ASL through a generated animated character. We additionally use ChatGPT’s large language model for auto- completion, creating a prediction service to infer the given message. A large amount of training data was needed, given the variety of backgrounds, lighting, size, color, length, and width of human hands. The results demonstrate a ___% accuracy score. Although several other methods exist such as communication service groups for the Deaf or sensory augmentation technologies, our application uses an AI model to bridge verbal communication gaps, allowing Deaf individuals to overcome language barriers without needing to accommodate a dominantly hearing society. Both the hearing and Deaf individual can use their natural language in real-time for more emotional, personal, and effective communication.
American Sign Language, Machine Learning, Translation, Flutte.
Haoyu Li1, Marisabel Chang2, 1Yorba Linda high school, 19900 Bastanchury Rd, Yorba Linda, CA 92886, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
The EcoFishCast system is an innovative tool designed to predict Dissolved Inorganic Carbon (DIC) levels in oceanographic environments using machine learning models [1]. By integrating a mobile application with a robust backend server, the system allows users to input environmental data and receive accurate predictions. Experiments conducted as part of the project identified Gradient Boosting and Random Forest as the most reliable models, particularly when combined with data scaling techniques, which significantly improved prediction accuracy [2][3]. While the system performs well, future enhancements are planned to address limitations related to training data diversity and computational efficiency, ensuring EcoFishCast remains a powerful and reliable resource for oceanographic analysis.
Dissolved Inorganic Carbon (DIC), Marine Science, Mobile Application, Machine Learning, Oceanographic Analysis.
Paul Knowles1, Bart Gajderowicz2, 1The Human Colossus Foundation, Geneva, Switzerland, 2Industrial Engineering Department, University of Toronto, Toronto, Ontario, Canada
This paper introduces a foundational Informatics domain model, a comprehensive data-centric design paradigm that addresses the challenges and limitations of current computational system designs and models. The Informatics model shifts the focus from node-centric to data-centric categorization, leveraging a multimodal framework of objects, events, concepts, and actions. Developed through rigorous research, the model draws on diverse disciplines, relying primarily on precise definitions from the Oxford Dictionary. It provides a foundational reference for system designers and data architects, enabling secure role-based access solutions and promoting semantic interoperability. The accompanying core data ontology enhances knowledge representation and fosters consistent understanding across distributed data ecosystems. The paper discusses the design of the ontology and presents the ontology in OWL 2. The paper also discusses the applications and benefits of the Informatics model, highlights its scalability, and outlines future directions for research. Readers may explore the Informatics Domain Model presentation deck for a more visual representation of the model. This paper provides a frame of reference and catalyst for adopting the Informatics model as a transformative approach in various domains and industries.
Informatics model, distributed data ecosystems, cryptographic data security, semantic interoperability, ontology design.
Jayden Szeto1, Rodrigo Onate2, 1American High School, 36300 Fremont Blvd, Fremont, CA 94536, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
The problem addressed by CareerCompass is the inefficiency and stress of the job search process, particularly for high school students [1]. CareerCompass integrates advanced mapping technologies and artificial intelligence to create a visual, easy-to-interact-with platform that displays job opportunities on a map, making it easier to find jobs based on specific geographical preferences [2]. Key technologies include Google Maps for better geolocation, Selenium for web scraping, and OpenAI GPT for job assessment and personalized lesson generation [3]. An experiment testing the accuracy of the job assessment system showed a 80% accuracy rate, highlighting its reliability and areas for potential improvement. By providing a user-friendly solution to job searching, the app reduces the time and stress associated with traditional job searching, empowering high school students to make better career decisions and bridging the gap between their current skills and job market requirements. This approach not only streamlines the job search process but also supports users in achieving their career goals through personalized learning paths, making CareerCompass a transformative tool for younger job seekers.
CareerCompass, Job Search Efficiency, AI-Powered Career Guidance, Geolocation Technology.
Anjun Chen1, Garret Washburn2, 1Sage Hill School, 20402 Newport Coast Dr., Newport Beach, CA92657, 2Computer Science Department, California State Polytechnic University, Pomona, CA91768
This research is intended to display and encapsulate a methodology for solving the problem of the desperate lackof information and resources for learning about and practicing opening moves in Chess. The methodology, GrandMaster Openings, comes in the shape of a mobile application available on both the Apple and Google Playstores and can be downloaded right now [1]. The GrandMaster Openings app seeks to give the user insightful information on opening moves in Chess, and even comes equipped with an AI chat feature with an AI GrandMastertrained on the same data that is on display for users to see within the app [2]. The most prominent technologies that were utilized to develop this application are ChatGPT’s trainable chat model features and the Flutter mobileapplication development framework, as well as a few extensive chess match datasets. During development, thecreation and implementation of a back-end server was necessary, as the transmission of quite extensive data andthehousing of a big AI model became cumbersome to keep on a user’s device, and proved to be quite challenging duetothe new implications of a back-end server [3]. Within this essay are experiments that were conducted specificallytargeted at the back-end server in order to discover any current or potential issues with it. Ultimately, after properdevelopment and thorough review, the GrandMaster Openings mobile application is a great resource for thoselooking to learn more about Chess openings, get immediate feed from a trained AI model GrandMaster, and reviewexcellently collected real life game data.
Chess, Opening, AI, LLM.
Ewelina Szczekocka, Orange Polska S.A., Poland
The article presents a state-of-the-art review on AI Risk Management, a first result in a research company project. It highlights crucial questions on practical implementations and depicts major challenges for organizations, finally proposing further directions in solving these challenges (ongoing work).
AI Risk Management, organizational challenge, standardisation.
Abderrazak Belabes, King Abdulaziz University, Jeddah, Saudi Arabia
Despite significant progress, AI is largely based on a statistical process that functions as a closed entropic system. This path can turn into a speculative bubble that can burst at any time. The challenge will be to enrich the computational approach, centred on numbers, with approaches open to forms and meanings, to link things that are difficult to associate by calculation. Hence the interest in the theme of Grothendieck topos, which offer unsuspected possibilities for enriching AI systems, beyond the accomplishment of tasks and the resolution of problems. The advantage of reading by meaning is that it offers the possibility of enriching both the theme of the topos and that of AI, through an interactive process, with a permanent back-and-forth.
Philosophy of Science, AI, Toposes, Computation, Geometric Forms, Reading by Meaning.
Zelin Jason Hu1, Garret Washburn2, 1Westminster School, 995 Hopmeadow St. Simsbury, Connecticut, 06070, 2Computer Science Department, California State Polytechnic University, Pomona, CA91768
The need for immediate medical or psychological attention is something that humanity has always needed as an essential service [1]. However, it is apparent nowadays that most individuals would say that the medical attention they receive is hardly immediate and doesn’t provide a diagnosis or advice in a quick fashion. To solve this issue, we propose the Physiomed mobile application available on both the Apple and Google Play Stores. The Physiomed mobile app provides immediate professional AI generated diagnosis and advice to individuals who have encountered physical or emotional symptoms in their day-to-day life [2]. The Physiomed application utilizes ChatGPT AI prompt generation on a hosted back-end server that receives the symptoms from different users and can provide accurate diagnosis based on real life medical data. The biggest challenge during development was creation of the prompt to receive insightful and accurate diagnosis, as well as the communication between the back-end server and the different user’s mobile devices.
AI-powered diagnosis, Mobile health app, ChatGPT integration, Real-time medical advice.
Zexi Chen1, Bobby Nguyen2, 1Arcadia High School, 180 Campus Dr, Arcadia, CA 91006, 2Computer Science Department, California State Polytechnic University, Pomona, CA91768
Parkinson 's disease, a progressive neurological disorder, affects millions globally, presenting challenges in symptom management and medication adherence. Levio is a mobile application developed to address these challenges comprehensively. Levio integrates several key features: a symptom tracker for logging and monitoring symptoms, a medication reminder system, voice and speech therapy exercises, and a movement and exercise coach. It also provides an online forum where users can ask and answer questions. The methodology involved using Flutter for the app development and Firebase for data storage. Key challenges included ensuring user engagement with symptom tracking, customizing speech therapy exercises, and providing accurate exercise guidance. These were addressed by implementing user-friendly interfaces, leveraging machine learning for personalized therapy, and incorporating AI-based motion detection. During testing, Levio demonstrated high reliability in document registration and machine learning accuracy, with mean and median success rates indicating robust performance. The app’s holistic approach provides a practical and integrated solution for managing Parkinson’s disease. Levio’s potential impact lies in its ability to consolidate multiple management aspects into a single platform, offering a significant improvement over existing fragmented tools and resources. .
Parkinson, Computer Vision, Machine Learning, AI.
Alexandr Silonosov and Lawrence Henesey, Blekinge Institute of Technology, Karlskrona, Sweden
The research presented in the paper evaluates practices of Attribute-Based Encryption as a key encapsulation mechanism, leading to a proposed end-to-end encryption architecture for a cloud-based ship tracking system. Though extensively used for efficiently gathering and sharing maritime data, these systems draw information from Automated Identification Systems, ports, and vessels, which can lead to cyber security vulnerabilities. This paper presents a study addressing the current state of knowledge, methodologies, and challenges associated with supporting crypto graphic agility for End-to-End Encryption (E2EE) for AIS data. To enhance cryptographic agility performance, a new metric has been introduced for cryptographic library analysis that improves the methodology by comparing Attribute-Based Encryption (ABE) with state of the art CRYSTALS Kyber key encapsulation mechanism (KEM) that belongs to Post-Quantum Cryptography (PQC). Acomprehensive series of experiments are undertaken to simulate large-scale cryptographic migra tion within the proposed system, showcasing the practical applicability of the proposed approach in measuring cryptographic agility performance.
AIS ship tracking data, Key encapsulation mechanism, end-to-end encryption, cryp tographic agility, CRYSTALS-Kyber,Post-Quantum Cryptography.
Alexander Xu1, Jonathan Sahagun2, 1Mt. SAC Early College Academy, 2226 E Rio Verde Dr, West Covina, CA 91791, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
This paper addresses the challenge of designing an adaptive egg-cracking device suitable for various egg sizes and shell strengths, especially for users with mobility issues [1]. The proposed solution integrates a mechanical egg cracker controlled via a Bluetooth-connected mobile application, utilizing adjustable cracking settings and machine learning algorithms for optimization [2]. The system is comprised of three main components: the hardware cracker, Bluetooth communication, and the mobile app interface [3]. Key challenges included variability in egg characteristics, ensuring reliable Bluetooth connectivity, and developing a suitable dataset for machine learning. Through a series of experiments, we evaluated the device’s performance and connectivity, revealing areas for improvement and showcasing its potential for versatility. The project’s findings suggest that this egg-cracking solution offers a more accessible and efficient alternative to traditional methods, making it a valuable tool for those seeking a reliable kitchen aid.
Bluetooth, Cooking aid, Aid for individuals with mobility issues, Egg cracker.
Annabel Shen Tu1, Andrew Park2 , 1Phillips Academy Andover, 180 Main St, Andover, MA 01810, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
Medifact is a mobile application designed to combat the growing problem of medical misinformation by providing users with accurate and trustworthy medical information [1]. This research focused on evaluating the app’s effectiveness through two experiments. The first experiment assessed user perceptions of information accuracy, revealing generally high confidence in the apps content, though with some identified areas for improvement. The second experiment evaluated overall user satisfaction, with high scores for usability and design but noting the need for better information accessibility. Challenges related to data sourcing and system scalability were identified, and potential solutions were proposed. Despite these challenges, Medifact demonstrates significant potential as a reliable resource for public health education [2]. This research highlights the importance of continuous improvement in data validation and user interface design to ensure that Medifact remains a trusted tool for combating medical misinformation.
Health Information, Medicine, Mobile Application, Flutter.
Jiaqi Ji1, Garret Washburn2, 1Dr. TJ Owen’s Gilroy Early College Academy, 5055 Santa Teresa Blvd, Gilroy, CA 95020, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
The technology currently used to monitor school buses and allow for effective communication within them is outdated and does not reflect the technological advances we’ve made as a society. Specifically, it was very common nowadays for parents to be uninformed about the status of their child on a school bus, and the student to be unaccounted for by the driver [2]. To solve this problem, this paper proposes the AI Smart Route mobile application as a solution. The app utilizes facial mesh recognition in order to take attendance and keep track of students getting on and off buses. This allows driver, school, and parents to be informed of whether a student is on the bus. The major technologies used to develop the AI Smart Route mobile include but are not limited too the Flutter framework for the mobile app development, Mediapipe for the facial mesh recognition, and AWS for the back end server hosting [3]. To ensure the mobile application worked as intended, experiments were performed to ensure proper function. The AI Smart Route mobile app is an effective application that surpasses the current methods of taking bus attendance and keeping everyone involved informed of who is on the school bus.
AI , Facial Recognition, App, School Bus.
Farabi Mahmud, Harpreet Singh Chawla, Chia-Che Tsai, EJ Kim, and Abdullah Muzahid, Computer Science and Engineering, Texas A&M University, Texas, USA
Cache side-channel attacks are considered some of the biggest threats to security and privacy in modern computers. Among a multitude of cache side-channel attacks, researchers have recently focused on attacks in large scale multicore architectures, specifically the ones with distributed last level cache (LLC). These attacks capitalize on either physical distance or network congestion to create non-uniform access la tency and subsequently, to leak secrets. In this paper, we propose BOUNDNOC, an efficient, architectural defense mechanism against non-uniform cache access side-channel attacks. BOUNDNOC works by essentially creating an oblivious LLC. When a potentially leaking memory access instruction is executed, BOUNDNOC uses a combination of delay and router bypass mechanisms to eliminate any LLC access (hit) time differences, thereby preventing the side-channel attack. We evaluate two versions of BOUNDNOC- BOUNDNOCDELAY and BOUNDNOCBYPASS. Weimplement the schemes in Gem5 simulator and show that they incur performance overheads of 17.04% or none over the insecure baseline for a range of applications from PARSEC and Rodinia benchmark suites.
Miley Huang1, Zihao Luo2,1Abeka Academy Video Homeschool, 250 Brent Ln, Pensacola, FL 32503, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
This research paper presents the development and analysis of an AI-assisted task management application designed to streamline the process of task creation through both manual and voice inputs [5]. Traditional task management tools often suffer from tedious manual input processes and lack portability, reducing user efficiency and engagement [6]. The proposed app leverages advanced AI technologies to facilitate hands-free task management, allowing users to create tasks via voice commands, even in multitasking scenarios [7]. Experiments conducted to assess the app&aposs performance revealed challenges in voice recognition accuracy, particularly in noisy environments, and a decline in responsiveness under high data loads. However, the app&aposs ability to evolve with AI advancements positions it as a forward-looking solution to the inefficiencies of conventional task management systems. Future improvements focus on enhancing speech recognition accuracy and optimizing data handling processes to ensure a seamless user experience. This work contributes to the ongoing integration of AI into everyday productivity tools.
ChatGPT, Scheduling assistant, Bluetooth, AI.
Jindong Sha1, Jonathan Sahagun2, 120402 Newport Coast Dr, Newport Coast, CA 92657, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
This paper addresses the challenge of designing an automated system for precise motor control, focusing on enhancing accuracy and adaptability in dynamic environments [1]. The project integrates advanced feedback mechanisms with cost-effective sensors and control algorithms to improve system reliability [2]. Two main experiments were conducted: one tested the precision of stepper motors in reaching designated positions, while the other examined the system's response to unexpected input variations. The results indicated that while the system generally performed well, there were areas for improvement, particularly in feedback mechanisms. The paper also compares the project’s methodology with other existing approaches, highlighting the balance between precision, adaptability, and cost-effectiveness [3]. Despite certain limitations, the project successfully demonstrates a functional automated system with potential applications in various fields. This solution is particularly relevant for scenarios where cost-effective, reliable automation is required, making it a valuable contribution to the field.
Automated Motor Control, Feedback Mechanisms, Precision and Adaptability, Cost-Effective Sensors
Miley Huang1, Zihao Luo2,1Abeka Academy Video Homeschool, 250 Brent Ln, Pensacola, FL 32503, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
This research paper presents the development and analysis of an AI-assisted task management application designed to streamline the process of task creation through both manual and voice inputs [5]. Traditional task management tools often suffer from tedious manual input processes and lack portability, reducing user efficiency and engagement [6]. The proposed app leverages advanced AI technologies to facilitate hands-free task management, allowing users to create tasks via voice commands, even in multitasking scenarios [7]. Experiments conducted to assess the app&aposs performance revealed challenges in voice recognition accuracy, particularly in noisy environments, and a decline in responsiveness under high data loads. However, the app&aposs ability to evolve with AI advancements positions it as a forward-looking solution to the inefficiencies of conventional task management systems. Future improvements focus on enhancing speech recognition accuracy and optimizing data handling processes to ensure a seamless user experience. This work contributes to the ongoing integration of AI into everyday productivity tools.
ChatGPT, Scheduling assistant, Bluetooth, AI.
Zixuan Feng1, Edmond You2, 1Arcadia High School, 180 Campus Dr, Arcadia, CA 91006, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
This research paper explores the development of an intelligent music generation application designed to overcome creative stagnation in the music composition process [1]. The app leverages advanced AI techniques, specifically an improved Transformer-XL model, to generate original music based on user inputs, such as text prompts or audio files [2]. The system integrates three major components: a user-friendly interface built with Flutter, a robust backend powered by Python and Firebase for data management, and an AI engine for music generation [3]. Through experiments, the apps performance was evaluated in terms of quality and latency across different input complexities. Results showed that while the AI performs well with simple inputs, it faces challenges with more complex or abstract data, highlighting areas for further optimization. The project demonstrates significant potential in democratizing music creation, providing musicians with an accessible tool to generate and refine musical ideas, ultimately enhancing productivity and creativity in the music industry [3]..
AI-driven music generation, Flutter, Music composition, Algorithmic composition, Music technology.
Alina Duan1 , Joshua Lai2, 1Pacific Academy, 4947 Alton Pkwy, Irvine, CA 92604, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
This paper presents the development and analysis of a productivity app designed to improve task management through gamification and social competition [7]. The app, built using Flutter and Firebase, features a to-do system that allows users to create, track, and organize tasks while competing with others based on task completion [8]. Two key experiments were conducted to evaluate the apps performance: the first assessed synchronization accuracy across different network conditions, revealing significant delays under high latency and intermittent connectivity [9]. The second experiment analyzed the apps ability to handle increasing task volumes, showing a decline in performance as the number of tasks increased. Despite these challenges, the app demonstrates potential in fostering better time management among users, particularly students. Future improvements, including optimizing performance for large data volumes and enhancing cross-platform compatibility, are proposed to further enhance the apps effectiveness and user experience.
Application, Social, Flutter, Firebase.
Fan Lin1, Garret Washburn2, 1Lower Merion High School, 322 Parsons Avenue, Bala Cynwyd, PA 19004, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
The problem of managing multiple open tabs in browsers like Google Chrome is common, often leading to decreased productivity due to the difficulty in navigating and locating specific tabs [1]. This paper proposes a solution through the development of a Chrome extension called Tab Sorter, which organizes open tabs by domain, providing a simple and effective method for users to quickly find and manage their tabs. The extension avoids the complexity of AI-based solutions and focuses on user-friendliness and efficiency. Key components include a background service that listens for tab changes, a popup script that updates the user interface, and a clean, organized frontend. Experiments demonstrated the extension’s efficiency in reducing the time needed to locate tabs and its minimal impact on system resources. Compared to other existing solutions, Tab Sorter offers a streamlined, no-cost alternative that significantly enhances tab management without overwhelming users with unnecessary features [2].
Chrome Extension, Tab Management, Browser Productivity, Domain Sorting, Tab Organization.