Lasal Jayawardena and Prasan Yapa, School of Computing, Informatics Institute of Technology, Colombo 00600, Sri Lanka
Paraphrase generation, a pivotal task in natural language processing (NLP), has tra-ditionally relied on human-annotated paraphrase pairs, a method that is both cost-inefficient and difficult to scale. Automatically annotated paraphrase pairs, while more efficient, often lack in syntactic and lexical diversity, resulting in paraphrases that closely resemble the source sentences. Moreover, existing datasets often contain hate speech and noise, including unintentional inclusion of non-English languages. This research introduces ParaFusion, a large-scale, high-quality English paraphrase dataset developed using Large Language Models (LLM) to address these challenges. ParaFusion augments existing datasets with high-quality data, significantly enhancing both lexical and syntactic diversity while maintaining semantic similarity. It also mitigates the presence of hate speech and reduces noise, ensuring a cleaner, and more focused English dataset. The paper presents one of the most comprehensive evaluations to date, employing a range of evaluation metrics to assess different aspects of the dataset quality. The results underscore the potential of ParaFusion as a valuable resource for improving NLP applications.
Paraphrase Generation, Natural Language Generation, Deep Learning, Large Lan-guage Models, Data Centric AI.
Miguel ´Angel Medina-Ram´ırez, Cayetano Guerra-Artal, and Mario Hern´andez-Tejera, University Institute of Intelligent Systems and Numeric Applications in Engineering, University of Las Palmas de Gran Canarias, Las Palmas de Gran Canarias, Spain
Task-oriented dialogue systems (TODS) have become crucial for users to interact with machines and computers using natural language. One of its key com- ponents is the dialogue manager, which guides the conversation towards a good goal for the user by providing the best possible response. Previous works have proposed rule-based systems (RBS), reinforcement learning (RL), and supervised learning (SL) as solutions for the correct dialogue management; in other words, select the best response given input by the user. This work explores the impact of dataset quality on the performance of dialogue managers. We delve into po- tential errors in popular datasets, such as Multiwoz 2.1 and SGD. For our inves- tigation, we developed a synthetic dialogue generator to regulate the type and magnitude of errors introduced. Our findings suggest that dataset inaccuracies, like mislabeling, might play a significant role in the challenges faced in dialogue management. The code for our experiments is available in this repository: https: //github.com/miguel-kjh/Improving-Dialogue-Management.
Dialog Systems, dialogue management, dataset quality, supervised learn-ing.
Sriraghavendra Ramaswamy, Amazon Development Center (India) Private Limited, Chennai, India
We present a supervised learning approach for automatic extraction of keyphrases from single documents. Our solution uses simple to compute statistical and positional features of candidate phrases and does not rely on any external knowledge base or on pre-trained language models or word embeddings. The ranking component of our proposed solution is a fairly lightweight ensemble model. Evaluation on benchmark datasets shows that our approach achieves significantly higher accuracy than several state-of-the-art baseline models, including all deep learning-based unsupervised models compared with, and is competitive with some supervised deep learning-based models too. Despite the supervised nature of our solution, the fact that does not rely on any corpus of “golden” keywords or any external knowledge corpus means that our solution bears the advantages of unsupervised solutions to a fair extent.
Keyphrase extraction, Supervised learning, Partial Ranking, Domain-agnostic solution, Non-DNN-based model.
Aanchal Varma and Chetan Bhat, Freshworks, India
Recent emergence of large language models (LLMs), particularly GPT variants has created a lot of buzz due to their state-of-the-art performance results. However, for highly domain-specific datasets such as sales and support conversations, most LLMs do not exhibit high performance out-of-the-box. Thus, fine- tuning is needed which many budget-constrained businesses cannot afford. Also, these models have very slow inference times making them unsuitable for many real-time applications. Lack of interpretability and access to probabilistic inferences is another problem. For such reasons, BERT-based models are preferred. In this paper, we present SAS-BERT, a BERT-based architecture for sales and support conversations. Through novel pre-training enhancements and GPT-3.5 led data augmentation, we demonstrate improvement in BERT performance for highly domain-specific datasets which is comparable with fine- tuned LLMs. Our architecture has 98.5% fewer parameters compared to the largest LLM considered, trains under 72 hours, and can be hosted on a single large CPU for inference.<\p>
BERT, LLM, Text Classification, Domain pre-training, NLP applications.
Chen Lin and Piush Kumar Singh and Yourong Xu and Eitan Lees and Rachna Saxena and Sasidhar Donaparthi and Hui Su, Fidelity Investments, 245 Summer Street, Boston, MA 02210
In this paper, we propose using domain adaptation to improve the generalizability and performance of LayoutLM, a pre-trained language model that incorporates layout information of a document image. Our approach uses topic modelling to automatically discover the underlying domains in a document image dataset where domain information is unknown. We evaluate our approach on the challenging RVL-CDIP dataset and demonstrate that it significantly improves the performance of LayoutLM on this dataset. Our approach can be applied to other NLP models to improve their generalization capabilities, making them more applicable in real-world scenarios, where data is often collected from a variety of domains<\p>
LayoutLM, Domain Adaptation, Automatic Domain Discovery, Topic Modelling, RVL-CDIP.
Thushara Manjari Naduvilakandy, Hyeju Jang, and Mohammad Al Hasan, Dept. of Computer Science, Luddy School of Informatics, Computing, and Engineering, Indiana University at Indianapolis, IN
Hypernymy directionality prediction is an important task in Natural Language Pro- cessing due to its significant usages in natural language understanding and generation. Many supervised and unsupervised methods have been proposed for this task, but existing unsupervised methods do not leverage distributional pre-trained vectors from neural language models, as super- vised methods typically do. In this paper, we present a simple yet effective unsupervised method for hypernymy directionality prediction that exploits neural pre-trained word vectors in context, based on the distributional informativeness hypothesis. Extensive experiments on seven datasets demonstrate that our method outperforms or achieves comparable performance to existing unsu- pervised and supervised methods.<\p>
hypernymy directionality prediction, distributional informativeness hypothesis.
Tahsinul Haque Dhrubo, Noshin Tabassum, ASM Tareq Mahmood, Riead Hasan Khan, Farig Yousuf Sadeque, and Muhammad Iqbal Hossain, Computer Science and Engineering, BRAC University, Dhaka, Bangladesh
Modernization uses technology to improve every aspect of our lives. Scientists initially developed tools to speed up communication. Our research seeks to improve the lives of those facing challenges using modern technology. In 1977, machine interpretation of sign language was possible and limited. An experiment matched console English letters to ASL manual set letters reenacted on a mechanical hand. These innovations turn sign language into a gesture-based language. The desired point is emerging. It has begun developing tools to help sign language communicators. This paper’s goal is to facilitate communication between hearing individuals and those who are deaf. The proposed system uses NLP techniques to identify deafness indicators and translate them into an easily understood language, enabling communication between individuals with and without hearing impairments. The BdSL enhanced the dataset. If someone wants to use the model for another language, they must update the dataset. The goal of this research is “Communications for everyone”.<\p>
NLP, Deaf People, BdSL, Sign Language, Communication.
Qizhen Zhao1, Tochi Onuegbu2, 1Shanghai Pinghe School, 261 Huangyang Road, Shanghai, China 201203, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
In a world where music accompanies various tasks, our paper addresses the challenge of understanding the impact of background music on work efficiency. The background problem centers on the lack of precision in existing studies, overlooking individual preferences and work types. Our proposed solution is a Python-based application that evaluates an individuals work efficiency while listening to different music genres [1]. The user-friendly interface incorporates features like music category selection, login options, and real-time statistics tracking [2][3]. Challenges, such as diverse user interactions and limited data, were addressed through a feedback channel for continuous improvement. The application underwent experiments, including regression model evaluations for essay grading and SVM parameter tuning [4]. Results indicated superior performance, emphasizing the relevance of ensemble learning and optimal parameter selection. This application provides a nuanced understanding of how background music influences work efficiency, offering a personalized approach that people can leverage for enhanced productivity and satisfaction in various work scenarios.<\p>
Natural Language Processing, Machine Learning, Efficiency Evaluation, Classifier
Okoro, C. Stanley, Lopez Alexander, Unuriode, O. Austine, Department of Computer Science, Austin Peay State University, Clarksville, USA
In recent years, wildfires have emerged as a global environmental crisis, causing significant damage to ecosystems, and contributing to climate change. Wildfire management methods involve prevention, response, and recovery efforts. Despite advancements in detection methods, the increasing frequency of wildfires necessitates innovative solutions for early detection and efficient management. This study explores proactive approaches to detect and manage wildfires in the United States by leveraging Artificial Intelligence (AI), Machine Learning (ML), and 5G technology. The specific objective of this research covers proactive detection and prevention of wildfires using advanced technology; Active monitoring and mapping with remote sensing and signaling leveraging on 5G technology; and Advanced response mechanisms to wildfire using drones and IOT devices. This study was based on secondary data collected from government databases and analyzed using descriptive statistics. In addition, past publications were reviewed through content analysis, and narrative synthesis was used to present the observations from various studies. The results showed that developing new technology presents an opportunity to detect and manage wildfires proactively. This would save a lot of lives and prevent huge economic loss that is attributed to wildfire outbreaks and spread. Advanced technology can be used in several ways to help in the proactive detection and management of wildfires. This includes the development of the use of AI-enabled remote sensing and signaling devices and leveraging 5G technology for active monitoring and mapping of wildfires. In addition, super intelligent drones and IOT devices can be used for safer responses to wildfires. This forms the core of the recommendation to the fire Management Agencies and the government.
Wildfires, Artificial Intelligence (AI), Machine Learning (ML), 5G technology, remote sensing, drones, and IoT device.
Unuriode, O. Austine, Okoro, C. Stanley, Afolabi, T. Osariemen, Durojaiye, M. Olalekan, Lopez Alexander, Yusuf, Y. Babatunde, Akinwande, J. Mayowa, Department of Computer Science, Austin Peay State University, Clarksville, USA
This study delves into the implications of AI adoption on the labor market. As artificial intelligence (AI) continues to transform industries, it presents a dual impact: job displacement and job creation. AI-driven automation is automating routine and repetitive tasks, which can lead to the displacement of certain roles. However, AI also creates new job opportunities, particularly in AI development and related fields. In this study, we were able to show AIs influence on the human performed tasks. The negative relationship between AI influence and tasks performed by humans shows that AI indeed has a notable and statistically significant adverse impact on human-performed tasks. We discovered that as AI technology advances and becomes more prevalent, certain tasks and roles traditionally carried out by humans are being automated or replaced by machines. Also, we were able to show the relationship between the AI model and human-performed tasks. It was found that AI models exhibit a substantial and statistically significant positive relationship with tasks performed by humans. Our finding suggests a more optimistic outlook for the labor market, where rather than displacing jobs and workers, AI technologies have the potential to enhance their capabilities and create new opportunities.
Artificial Intelligence, Automation, labor market, machine learning.
Mayowa Akinwande, Alexander Lopez, Tobi Yusuf, Austine Unuriode, Babatunde Yusuf, Toyyibat Yussuph and Stanley Okoro, Department of Computer Science, Austin Peay State University, USA
This paper provides a comprehensive examination of the evolution of credit cards in the United States, tracing their historical development, causes, consequences, and impact on both individuals and the economy. It delves into the transformation of credit cards from specialized merchant cards to ubiquitous financial tools, driven by legal changes like the Marquette decision. Credit card debt has emerged as a significant financial challenge for many Americans due to economic factors, consumerism, high healthcare costs, and financial illiteracy. The consequences of this debt on individuals are extensive, affecting their financial well-being, credit scores, savings, and even their physical and mental health. On a larger scale, credit cards stimulate consumer spending, drive e-commerce growth, and generate revenue for financial institutions, but they can also contribute to economic instability if not managed responsibly. The paper emphasizes various strategies to prevent and manage credit card debt, including financial education, budgeting, responsible credit card uses, and professional counselling. Empirical studies support the relationship between credit card debt and factors such as financial literacy and consumer behavior. Regression analysis reveals that personal consumption and GDP positively impacts credit card debt indicating that responsible management is essential. The paper offers comprehensive recommendations for addressing credit card debt challenges and maximizing the benefits of credit card usage, encompassing financial education, policy reforms, and public awareness campaigns. These recommendations aim to transform credit cards into tools that empower individuals financially and contribute to economic stability, rather than sources of financial stress.
Debt, Financial literacy, financial well-being, Economic stability, Credit cards.
Humaira Farid and Volker Haarslev, Concordia University, Montreal, Canada
This paper presents a novel SHOIQ tableau calculus which incorporates algebraic reasoning for deciding ontology consistency. Numerical restrictions imposed by nominals and qualified number restrictions are encoded into a set of linear inequalities. Column generation and branch-and-price algorithms are used to solve these inequalities. Our preliminary experiments indicate that this calculus is more stable and often performs better on SHOIQ ontologies than standard tableau methods.
Description logic, knowledge representation, algebraic reasoning.
Ewuradjoa Mansa Quansah, Saint Petersburg University, Russia
As generative AI systems like ChatGPT gain popularity, empirical analysis is essential to evaluate capabilities. This study investigates ChatGPT’s skills for mathematical calculations through controlled experiments. Tests involving counting numbers, finding averages, and demonstrating Excel methods reveal inconsistencies and errors, indicating lack of true contextual understanding. While ChatGPT can provide solutions, its reasoning shows gaps versus human cognition. The results provide concrete evidence of deficiencies, complementing conceptual critiques. Findings caution against over-reliance on generative models for critical tasks and highlight needs to advance reasoning and human-AI collaboration. This analysis contributes valued grounding amidst hype, urging continued progress so technologies like ChatGPT can be deployed safely and responsibly. Overall, empirical results underscore risks and limitations, providing insights to maximize benefits while mitigating harms of rapidly advancing generative AI.
ChatGPT, Artificial intelligence, AI, Generative AI, large language-based models, experiment.
Hongfan Zhu1, Theodore Tran2, 1YK Pao School, 1251 West Wuding Road, Changning District, Shanghai, China, 2Computer Science Department, California State Polytechnic University, USA
Motion Mentor is a mobile application designed to address a challenge faced by beginner dancers in improving their dancing technique-the development of improper movements and habits when practicing without a teacher’s guidance. Therefore, Motion Mentor offers real-time posture correction and personalized feedback [5]. This method involves using the Mediapipen pose-detection AI model for real time posture detection, combined with advanced algorithms for accurate dance analysis, and Firebase for the storage of data and uploaded videos [6]. Users can access educational content, record their dance performances for feedback, and review their progress. During the experimentation, our system was applied to scenarios involving rapid dance movements to test the accuracy of pose estimation, comparison between the estimated and the actual real-time distance and speed estimation [7]. These scenarios suggested the limitations of our application in different dynamic and lighting conditions, providing insights into areas for improvement. Overall, this solution enhances accessibility and conveniences for all dancers in improving dance technique, offering real-time feedback and educational materials.
Dance Movement, Pose Detection, Video Processing.
Kun Bu and Kandethody Ramachandran, University of South Florida, USA
This paper investigates the sentiments of Twitter users towards the emergent topic of ChatGPT, leveraging advanced techniques in natural language processing (NLP) and sentiment analysis (SA). Our approach uniquely incorporates a dual setting for sentiment analysis: one analyzes the sentiments of original, full-length tweets, while the other first condenses these tweets into succinct summaries before performing sentiment analysis. By employing this dual approach, we are able to offer a comparative analysis of sentiment assessment pre- and post-text summarization, exploring the accuracy and reliability of the summarized sentiments. Central to our methodology is the application of Transformer models, specifically ProphetNet, which facilitates a deeper and more nuanced understanding of the original text. Unlike traditional methods that rely on keyword extraction and aggregation, our approach generates coherent and contextually rich summaries, providing a novel lens for sentiment analysis. This research contributes to the field by presenting a comprehensive study comparing sentiment analysis outcomes between original texts and their summarized counterparts, and examining the effectiveness of different NLP techniques, namely NLTK and the Transformer-based ProphetNet model. The findings offer valuable insights into the dynamics of sentiment analysis in the context of social media and the efficacy of state-of-the-art NLP technologies in processing complex, real-world data.
Sentiment Analysis, Natural Language Processing, Text Summarization, Machine Learning, Twitter Data Analysis, ProphetNet, Transformers.
Henry Muyang Liu1, Jenny Do2, 1LASA, 1012 Arthur Stiles Rd, Austin TX 78721, 2Computer Science Department, California State Polytechnic University, USA
ChatGPT has integrated itself into the academic space in an unprecedented timeframe, as the promise of hours of work done in seconds can outweigh senses of honor and logic [1]. This project determines whether a student is cheating or not by reducing the presence of human decision-making while simultaneously acting as a deterrent for future usage of AI technology [2]. Utilizing databases such as Kaggle, we can procure several samples of human writing in conjunction with ChatGPTs API to generate artificial intelligence instances, which are then stored for usage in machine learning algorithms [3]. Employing powerful Python libraries such as Sklearn and NLTK, we can utilize natural learning processing, the ability for computers to understand human writing, to yield an algorithm that can predict with approximately 96% certainty [4]. The result is a probability ratio, with one side displaying the percentage chance of the sample being human-written, whereas the other displays the likelihood of AI-generated instances. Furthermore, innovation lies in integrating this algorithm with wearable augmented reality technology, allowing users to efficiently scan and assess text elements. This approach amalgamates and helps reduce the delay between text input and response, empowering users to contribute to the decision-making process in identifying academic dishonesty without any loss in efficiency. The result that is shown displays pieces of information to the user that can all play a large role when determining the possibility of cheating, granting the user a role in making the decision along with simple scanning of each text element.
Integrity, AI-driven Detection, Augmented Reality Integration.
Takuto Tsukiyama, Sho Ooi and Mutsuo Sano, Graduate School of Osaka Institute of Technology, Osaka, Japan
People from various professions are involved in the production of anime, including directors, animation directors, character designers, and voice actors/actresses. Specifically, the role of an animation director is of importance in the realm of animation. The animation director serves as the unifying force in shaping the animations style by meticulously reviewing and redrawing the key animations provided by the key animators. The aim of this study is to develop a redrawing system using GLCIC to reduce the workload on animation directors when redrawing original key animation. Specifically, this study devised a system that employs GLCIC to analyze and learn the distinctive drawing styles of individual animation directors from images of their work and subsequently apply those styles to the conversion process. In the experiment, we asked people whose hobby is drawing to experience the developed system, and conducted a qualitative evaluation using a questionnaire and a quantitative evaluation using KLM analysis. As a result, we found that there were issues with ease of modification and UI. Additionally, the KLM analysis revealed that improving the system could reduce work time by a quarter. In the future, we think to improve the system with the aim of increasing work efficiency.
GLCIC, Image Conversion System, Animator Support.
Shogo Saito, Sho Ooi, and Mutsuo Sano, Graduate School of Osaka Institute of Technology, Osaka, Japan
Previous studies have attempted to estimate existing voices from images of animated characters as a way to generate voices suitable for animated characters, but without good results. Therefore, in this study, to link the voice characteristics to match the animation character with the image, we devised a method to analyze the voice s tendency to not be uncomfortable and then establish the ratio of voice learning data based on the analyzed tendency data. Specifically, this study prepares multiple voices for one illustration of an anime character, asks subjects to evaluate the voices, and calculates an evaluation based on the evaluation values. In experiments, we conducted an evaluation experiment using the one-pair comparison method, calculated the distribution of learning data based on the evaluation values obtained, and prepared for the subsequent learning process.
Synthesized speech, Voice generation, Character.
Boo Ho Voon1, Muhammad Iskandar Hamzah1, Teck Weng Jee2, Li Li Lau1, Squiter Macroy Wilson1, Ai Kiat Teo3 Universiti Teknologi MARA, Malaysia, 2Swinburne University of Technology, Kuching, Malaysia, 3SMK DPHA Gapor, Malaysia
The persons with disabilities (PwDs) consistently and regularly need the caring and inclusive healthcare services from the related stakeholders to maintain and even improve their socio-economic well-being. This paper reported the findings based on seven focus group discussions which were participated by the community-based rehabilitation center managers, trainers, and parents) in Malaysia. There were 8-12 participants in each focus group. The discussions were audio-recorded and analyzed accordingly to generate the meaningful themes. The findings suggested six important dimensions of rehabilitation service excellence culture, namely: Trainee orientation, Competitor orientation, Inter-functional coordination, Excellence-driven, Long-term focus, and Employee orientation. These dimensions and their respective items were used to operationalize the multi-item measures to develop the measurement-oriented information system to serve the parents and trainees better. A RehabServE information system will be developed. It is useful to monitor and support the community-based rehabilitation service excellence from the parents’/guardian’s perspective. The dimensional and overall composite scores for RehabServE (Rehabilitation Service Excellence) will be cost effective and convenient for the community-based rehabilitation centers.
Rehabilitation service; Trainee-centered system; Information systems applications
Sofıa Ramos-Pulido, Neil Hernandez-Gress, and Hector G. Ceballos-Cancino, Tecnologico de Monterrey, Av. Eugenio Garza Sada 2501 Sur, Tecnologico, 64849 Monterrey, N.L
Machine learning (ML) algorithms are predictively competitive algorithms with many human-impact applications. However, the issue of long execution time remains unsolved in the literature for high-dimensional spaces. This study proposes combining ML algorithms with an efficient methodology known as the barycentric correction procedure (BCP) to address this issue. This study uses synthetic data and an educational dataset from a private university to show the benefits of the proposed method. It was found that this combination provides significant benefits related to time in synthetic and real data without losing accuracy when the number of instances and dimensions increases. Additionally, for high-dimensional spaces, it was proved that BCP and Linear SVM, after an estimated feature map for the Gaussian radial basis function (RBF) kernel, were unfeasible in terms of computational time and accuracy.
Support vector machine, neuronal networks, gradient boosting, barycentric correction procedure, synthetic data, linear separable cases, nonlinear separable cases, real data.
Isaak Babaev, Todd Packer, Mehdi Ghayoumi, Kambiz Ghazinour, The Advanced Information Security and Privacy Lab, State University of New York, Canton, NY, USA
This paper introduces MAISON, an innovative model designed to combat cyberbullying on social media platforms. Addressing the challenge of anonymity that facilitates such behavior, MAISON integrates advanced user identification policies and employs AI-driven detection mechanisms to effectively identify and mitigate cyberbullying incidents. This approach goes beyond traditional measures, suggesting a combination of technological enhancements and policy reforms, including the use of facial motion vector detection to deter anonymous account creation for malicious purposes. The model emphasizes a holistic strategy, focusing on victim support and resilience, while advocating for robust measures against policy evasion. By aligning with emerging legal frameworks and societal demands for safer digital spaces, MAISON represents a comprehensive solution aimed at reducing both cyberbullying and its offline counterparts, thereby fostering a safer and more responsible online environment.
Cyberbullying, Privacy, Security, Society
Yuhang Zeng1, Jonathan Sahagun2, 1Troy High School, 2200 Dorothy Ln, Fullerton, CA 92831, 2Computer Science Department, California State Polytechnic University, Pomona, CA91768
I wanted to solve this problem due to noticing many new pianists having issues with their hand posture. Therefore, my project solves the problem by making an app that uses a phone’s camera in order to track finger movements [1]. Since smartphones are extremely commonplace, this solution is available for most people. A firebase server is usedin order process images from the app and return numbers back to the app in order to display it [2]. The appalsotracks practicing sessions and displays data about the hands in every session. I had to optimize a lot of parts of theprogram, including determining the frequency in which a frame is sent to be analyzed to the server. I alsoexperimented with testing out the dif erent phone angles as well as the latency of the server. Overall, this is agoodapp for improving piano posture for new pianists.
Piano, Tracking, Computer Vision, Object Detection.
Summer Shen1, Moddwyn Andaya2, 1Saratoga High School, 20300 Herriman Ave, Saratoga, CA 95070, 2Computer Science Department, California State Polytechnic University, Pomona, CA91768
This paper addresses the challenges students face in ef icient essay writing, aiming to improve both productivity andwriting quality. Students often struggle with overwhelming tasks, slow progress, and lack of motivation duringthewriting process. Our proposal combines the Pomodoro Technique, gamification, and an AI-powered writing qualitychecker to of er a holistic solution [8]. The Pomodoro Technique breaks down tasks, while gamification elements, including a playful chicken character, make writing enjoyable. An AI-based writing quality checker provides realtime feedback on grammar, sentence structure, and clarity [9]. Challenges, such as balancing interactivity andef iciency, were addressed through thoughtful design decisions. Experimentation involved applying the application to various writing scenarios, showcasing its adaptability. Resultsdemonstrated enhanced productivity, improved writing quality, and increased user satisfaction. This comprehensiveapproach addresses the shortcomings of existing methodologies and provides users with a valuable tool to navigatethe challenges of essay writing ef iciently and ef ectively.
Essay, Gamification, Unity, Productivity.
Shangbo Wang, Department of Civil Engineering, The University of Hongkong
In recent years, the fields of statistics and machine learning have created numerous methods such as neural networks, decision trees, support vector machines etc. and they have shown their superiority in classification and choice making. However, machine learning models are seen as having a black-box characteristic and a lack of economic interpretation, which have limited their gaining strong popularity among econometricians. In this paper, we propose a generalized model tree which links economic theories of human decision-making such as underlying discrete choice models, to an ensemble soft-decision tree to improve travel mode forecasting performance, overcoming local and global preference heterogeneity without much sacrifice of interpretability and monotonicity. The generalized model tree is a two-stage model, which firstly applies soft-splitting and disjunctions-of-conjunctions rules to a Linear Combination of Compensable Attributes (LCCA) or non-compensable attributes to obtain the probability of each alternative being considered for each decision-maker, and then compensatory models are used at each output node to get the final prediction. We apply the Markov Chain Monte Carlo (MCMC) algorithm to search for the optimum tree by the derived log-likelihood function and improve the AdaBoost algorithm to overcome global preference heterogeneity. We validate the proposed method by using the 2012 California Household Travel Survey dataset (CHTS) and the 2017 National Household Travel Survey dataset (NHTS). We find that by taking into account global preference heterogeneity, the model tree can deliver improved prediction results compared to the standard popular a multinomial logit model (MNL) and the multinomial mixed logit model (MML).
MNL, MML, Discrete Choice Modeling, MCMC, AdaBoost, Travel Mode Forecast.
Trinh Vu Duc Anh1 and Nguyen Truong Thinh2, 1Department of Electrical Engineering, University of South Florida, Tampa City, FL 33620, USA, 2Institute of Intelligent and Interactive Technologies, University of Economics HCMC- UEH, Vietnam
In this paper, an algorithm is developed for the robot to take odometry combined with LiDAR (Light Detection and Ranging) input to perform localization and 3D mapping inside a swiftlet house model. The position of the walls in the swiftlet’s house for calibrating LiDAR data is obtained beforehand and the robot system would superimpose the LiDAR map and swiftlet’s nest to the provided global swiftlet house map. The LiDAR is able to generate a 2D map from point clouds with its 360-degree scan angle. Additionally, it is mounted to a 1 DOF arm for height variation thanks to a Stepper motor to achieve a 3D map from 2D layers. Swiftlet’s nests are detected by differentiating their distinctive shape from the planar concrete wall, recorded by the robot, and monitored until they are harvested. When the robot is powered up, it can localize itself in the global map as long as the calibrating wall is in view in one scan. We evaluate the robot’s functionality in the swiftlet’s cell model with swiftlet’s nest scanned. We propose a bird nest-oriented SLAM system that builds a map of birds’ nests on wood frames of swiftlet houses. The robot system takes 3D point clouds reconstructed by a feature-based SLAM system and creates a map of the nests on the house frame. Nests are detected through segmentation and shape estimation. Experiments show that the system has reproduced the shape and size of the nests with high accuracy.
Intelligent Systems, Recognition, Lidar, Bird’s nest, Monitoring system, SLAM, identified system.
Fidelis Egbuna1 and Chukwuemeka Omerenna, 1Department of Computer Science, IBLT University, Lome. Togo, 2MasteryHive, United Kingdom
The aim is to analyze the sales of a supermarket as well as predict the impact of future sales on profit increase and customers satisfaction in the organization. The technique used for value of purchase is Linear Regression Algorithm, a widely acclaimed Algorithm in the field of Machine Learning. Linear Regression was compared with K- Nearest Neighbors Algorithm as well as with Gradient Descent and Random Forest. The actual data of the year, 2019, was compared to the predicted value and the accuracy of prediction calculated. The results testify to the trustworthiness and accuracy of the different prediction algorithms used. The study showed Random Forest as the best model with the predictions highest accuracy.
Data Visualization, Prediction, Machine Learning, Linear Regression and K-Nearest Neighbors , Random forest and Gradient Descent and Radio Frequency Identification Tag.
Jiaxu Li1, John Morris2, 1Pacific Ridge School, 6269 El Fuerte St, Carlsbad, CA 92009, 2Computer Science Department, California State Polytechnic University, Pomona, CA91768
This paper addresses the challenge of simplifying 3D animation by introducing a Unity package that harnessesartificial intelligence (AI) to convert 2D images or videos into 3D animation frames [2]. The background tothisproblem lies in the arduous and time-consuming nature of 3D animation, which often deters developers and artistsfrom pursuing their creative visions [1]. Our proposed solution leverages AI algorithms to predict 3Dposes andmovements from 2D sources, making animation more accessible and cost-ef ective. Our package utilizes vector mathematics and Unitys capabilities, primarily focusing on establishing the body as ananchor for limb rotations. Challenges included intricate angle calculations and addressing orientationdiscrepancies. We resolved these challenges by refining the AI algorithms and providing user-friendly features [4]. Experimentation involved assessing accuracy, usability, and ef iciency. While accuracy in complex scenariosremains a challenge, user feedback highlighted its potential for ef iciency and time-saving. Ultimately, this tool bridges the gap between 2D and 3D animation, of ering accessibility, cost-ef ectiveness, andstreamlined workflows [3]. Its potential impact on animation and game development makes it a valuable additionfor both professionals and enthusiasts.
Machine Learning, Neural Network , AI, 3D.
Suman Sharma1 and Alexander alexandrov2, 1Data Scientist, Bright Coneection Consulting Inc, Sunnyvale, California, USA, 2MLOps DevOps Engineer, Bright Coneection Consulting Inc, Sunnyvale, California, USA
In the dynamic landscape of the banking industry, the synergistic application of data mining and information retrieval has emerged as a pivotal catalyst for elevating operational efficiency, enhancing customer service, and fostering data-driven decision-making. This abstract delves into a practical banking scenario where these two domains intertwine seamlessly to deliver tangible benefits. In this scenario, a retail bank embarks on a mission to augment customer experiences and fortify defenses against fraudulent transactions. Armed with an extensive repository of customer transaction data, encompassing credit card transactions and customer service interactions, the bank leverages the potential of data mining and information retrieval. Data mining takes center stage through its multifaceted roles such as Fraud Detection and Customer Segmentation, Predictive Analytics, Customer Support Chatbots, Search for Account Information. For customers accessing their online banking accounts, a search feature becomes a beacon of convenience. These feature facilitates swift retrieval of transaction histories. The true efficacy unfolds through the seamless integration of data mining and information retrieval identifying potentially fraudulent transactions in real-time. In response, the information retrieval system springs into action, issuing immediate alerts to customers via SMS or mobile app notifications. These alerts elucidate the issue and provide contact information for timely customer support. Data mining models meticulously assess credit risk using historical customer data. When customers apply for loans or credit cards, the information retrieval system expeditiously retrieves credit assessment results, accompanied by clear explanations for approval or denial decisions. The banking industry benefits immensely from the symbiotic relationship between data mining and information retrieval. While data mining exposes fraud, uncovers customer behavior, and drives predictive decision-making, information retrieval ensures rapid access to vital account information and timely support. This harmonious amalgamation leads to heightened customer satisfaction, minimized operational risks, and informed, data-driven decisions in the realm of banking operations.
Banking industry, Data mining, Information retrieval, Fraud detection, Customer behavior, Predictive decision-making, Customer satisfaction, Operational risks, Data-driven decisions, Symbiotic relationship.
Xiangfeiyang Li1, Jonathan Thamrun2, 1Fairmont Prep Academy, 2200 W Sequoia Ave, Anaheim, CA 92801, 2Computer Science Department, California State Polytechnic University, Pomona, CA91768
In response to the increasing threat of DDoS (Distributed Denial of Service) attacks, this project investigatesfortifying defenses against such malicious invasions. The project incorporates a user-friendly UI featuringtwobuttons: one for uploading captured traf ic files and another for analysis to classify whether it’s a DDoS attack. Thebackground of the problem aspires to a robust and adaptive DDoS detection system to ensure the continuity ofonline services [14]. To resolve this, the project proposes an automated DDoS attack detection mechanismpoweredby Machine Learning and Artificial Intelligence. The application involves two pivotal experiments: the first assessesmodel accuracy, highlighting the Decision Tree as the most promising, while the second focuses on preventingoverfitting during training, and the Random Forest Classifier stands out to this one [15]. The challengesencountered were mitigated through techniques like early stopping and regularization. The model’s applicationacross various scenarios showcased its potential for ef ective real-time DDoS detection.
DDoS Attack, Detection System, Artificial Intelligence, Recognization & Prevention
Arpan Mahara, Jose Fuentes, Christian Poellabauer, Naphtali D. Rishe, Knight Foundation School of Computing and Information Sciences, Florida International University, Florida, USA
Content caching is vital for enhancing web server efficiency and reducing network congestion, particularly in platforms predicting user actions. Despite many studies conducted to improve cache replacement strategies, there remains space for improvement. This paper introduces STRCacheML, a Machine Learning (ML) assisted Content Caching Policy. STRCacheML leverages available attributes within a platform to make intelligent cache replacement decisions offline. We have tested various Machine Learning and Deep Learning algorithms to adapt the one with the highest accuracy; we have integrated that algorithm into our cache replacement policy. This selected ML algorithm was employed to estimate the likelihood of cache objects being requested again, an essential factor in cache eviction scenarios. The IMDb dataset, constituting numerous videos with corresponding attributes, was utilized to conduct our experiment. The experimental section highlights our model’s efficacy, presenting comparative results compared to the established approaches based on raw cache hits and cache hit rates.
Cache Hit, Cache Miss, Content Caching, Machine Learning (ML), Simulation.
Ephrance Eunice Namugenyi, David Tugume, Augustine Kigwana and Benjamin Rukundo, Department of Computer Networks, CoCIS, Makerere University, Uganda
The dynamic landscape of modern agriculture, characterized by a growing reliance on data-driven methodologies, has created an urgent need for innovative solutions to enhance resource utilization. One key challenge faced by local beehive farmers is efficiently managing large data files collected from sensor networks for optimal beehive management. To address this challenge, we propose a novel paradigm that leverages advanced edge computing techniques to optimize data transmission and storage. Our approach encompasses data compression for images and videos, coupled with data augmentation techniques for numerical data. Specifically, we propose a novel compression algorithm that outperforms traditional methods, such as Bzip2, in terms of compression ratio. We also develop data augmentation techniques to improve the accuracy of machine learning models trained on the collected data. A key aspect of our approach is its ability to operate in resource-constrained environments, such as those typically found in local beehive farms. To achieve this, we carefully explore key parameters such as throughput, delay tolerance, compression rate, and data retransmission. This ensures that our approach can meet the unique requirements of beehive management while minimizing the impact on resources. Overall, our study presents a holistic solution for optimizing data transmission and storage across robust sensor networks for local beehive management. Our approach has the potential to significantly improve the efficiency and effectiveness of beehive management, thereby supporting sustainable agriculture practices.
Shihan Fu1, Ang Li2, 1Margarets Episcopal School, 31641 La Novia Ave, San Juan Capistrano, CA92675, 2Computer Science Department, California State Polytechnic University, Pomona, CA91768
Addressing the global challenge of inef icient waste management, my paper introduces an innovative recyclingsolution integrating machine learning, computer vision, and a robotic arm [1]. The background problemrevolvesaround inaccurate waste sorting and the environmental impact of recyclables ending up in landfills. The proposedsolution involves a sophisticated machine learning model for object recognition, a computer vision systemfor realtime detection, and a robotic arm for precise object manipulation [2]. Challenges included optimizing the machinelearning model for diverse materials and enhancing the robotic arms adaptability. Experimentation involved testingthe systems ef iciency in various scenarios, showcasing its ability to recognize and sort recyclables accurately. Theresults demonstrated promising accuracy and adaptability. Ultimately, this solution of ers a practical andautomated approach to waste sorting, reducing environmental impact, and promoting ef icient recycling practices, making it a valuable tool for waste management systems globally [3].
Harry Su1, John Morris2, 1Cate School, 1960 Cate Mesa Road, Carpinteria, CA 93013, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
Acemind aims to address the lack of real-time tennis court availability information, enhacing player experience, promoting community well-being, and making sports easily accessible to not only the wealthy [1]. The technology utilizes Raspberry Pi computers with Pi cameras to record live footage, Firestore Database to store information regarding court status, and a front-end mobile application made with flutter to display information to users [2]. Key challenges include running object detection model, YOLOv5, on the computer seamlessly with the camera, which was solved by adjusting libraries’ versions appropriately and ensuring the proper installation of all packages [3]. The mobile application also struggled to display the correct court’s information, but the problem was fixed with a setState function that updates the bottom popup widget using a variable. During experimentation, YOLOv5 consistently identified humans among distractions commonly found near tennis courts even under suboptimal conditions, proving its resilience to unwanted challenges in inputs [4]. Although Acemind has limitations such as the need of a cellular connection and government permission, it is useful as it presents valuable court information regarding availability without the shortcomings of smart tennis courts and tennis maps, bridging the inequality gap by providing everybody a change to play.
Tennis, AI, Community, Convenience.
Mai M. Goda1, Hassan Mostafa1, 2, and Ahmed M. Soliman1, 1Electronics and Communications Department, Cairo University, Giza, Egypt, 2Zewail City of Science and Technology, 6th of October City, Giza, Egypt
Under the terms of Moore’s law, complementary metal–oxide–semiconductor (CMOS) based technologies encounter design challenges correlated withthe progressive scaling down of the minimum feature size. In addition, recent applications, such as pattern recognition, vector processing, and big data analysis, have large demands which are no longer fulfilled by theconventional computing architecture. Thus, shifting tonew technologies is now a pressing need at the level of the architecture as well as the device level. Eventually, memristors are the most promising devicesthat can fulfill this task. A new dimensional advantage will be given to design novel circuits and systemsthanks to the memristor device. A novel circuit design of a Memristor-based Vernier delay-line Time to Digital Converter (TDC), is proposed on this paper, based on the voltage threshold adaptive memristor (VTEAM) model, as memristor provides low area, low power, and a low threshold voltage.Memristor-based TDC has been designed and implemented with an area size of 0.015?mm?^2, a time resolution of 160 ps, and with a power consumption of 3.6 mw, finally the simulation results are compared to a conventional CMOS-based TDC circuit.
Memristor, VTEAM model, Time-To-Digital Converter (TDC), Memristor-Based TDC, Vernier Delay-Line TDC, D Flip-flop
Olga Dye, Justin Heo, Ebru Celikel Cankaya, Department of Computer Science University of Texas at Dallas
As demand for more storage and processing power increases rapidly, cloud services in general are becoming more ubiquitous and popular. This, in turn, is increasing the need for developing highly sophisticated mechanisms and governance to reduce data breach risks in cloud-based infrastructures. Our research focuses on cloud governance by harmoniously combining multiple data security measures with legislative authority. We present legal aspects aimed at the prevention of data breaches, as well as the technical requirements regarding the implementation of data protection mechanisms. Specifically, we discuss primary authority and technical frameworks addressing least privilege in correlation with its application in Amazon Web Services (AWS), one of the major Cloud Service Providers (CSPs) on the market at present.
Least privilege, attribute-based access control, FedRAMP, zero-trust architecture, condition keys.
Jerry Ku1, Sikang Sun2, Yu Sun3, 1Diamond Bar High School, 21400 Pathfinder Rd, Diamond Bar, CA 91765, 2Department of Computer Science, Purdue University, West Lafayette, IN 47907, 3Computer Science Department, California State Polytechnic University, Pomona, CA 91768
This paper presents ComputeCycle, an innovative solution addressing the pressing issues of obesity and energy consumption through a unique fusion of physical exercise, energy generation, and volunteer computing. The background to this problem lies in the global obesity epidemic and the need for sustainable energy sources. ComputeCycle tackles these challenges by motivating individuals to engage in physical exercise, which simultaneously generates electrical energy to contribute to scientific research through volunteer computing. The proposal outlines the key components of ComputeCycle, including a customized bike system with a DC motor, 3D-printed wheels for efficient energy transfer, and a diode-based energy flow control mechanism. Challenges related to wheel attachment and energy flow direction were resolved through innovative 3D printing and diode integration. Experiments showcased ComputeCycles effectiveness in promoting exercise, energy generation, and volunteer computing, resulting in positive outcomes for health and scientific progress. Our findings emphasize the feasibility of this holistic approach to combat obesity and contribute to sustainable energy solutions. Ultimately, ComputeCycle offers a compelling solution that motivates individuals to improve their health while actively participating in scientific research, making it a valuable and impactful tool for individuals and communities seeking to address both health and environmental challenges.
Energy Generation, Obesity Prevention, Health and Fitness, Internet-Of-Things.
Kaining Yuan1, Jonathan Sahagun2, 1Woodbridge High School, 2 Meadowbrook, Irvine, CA 92604, 2Computer Science Department, California State Polytechnic University, Pomona, CA91768
Microplastic contamination in freshwater ecosystems is a growing environmental concern. This paper introducesMyRiiver, a solar-powered microplastic filtration system, designed to overcome limitations in current methods. Thebackground underscores the urgency of addressing microplastic pollution, emphasizing the need for an ef icient, adaptable, and economical solution. MyRiiver employs a sophisticated multi-layered filtration systemwithout requiring pre-treatment, of ering advantages over existing methodologies. Challenges identified in previousapproaches, such as electrode wear and biofilter maintenance, are addressed through the simplicity of MyRiivers design. Experimental trials showcase its adaptability and superior ef iciency in filtering microplastics as small as 1m. Results demonstrate a significant removal rate, positioning MyRiiver as a practical, scalable, and eco-friendlysolution. The study concludes by asserting MyRiivers potential as a transformative tool for combatingtheescalating global issue of microplastic contamination in freshwater environments.
Water Filter System, IOS, Android, Microplastics, Flutter.
Sakina Oussane, Haroun Benkaouha, and Amir Djouama, LSI Laboratory, USTHB, National School of Artificial Intelligence Algiers, Algeria
The Internet of Things (IoT) is a technology that allows for the connection of physical objects through various digital systems. It is employed in various fields, including medical monitoring. In this context, a specific network called Wireless Body Area Network (WBAN) can be integrated into the IoT infrastructure. This network enables the collection and exchange of health data, the detection of anomalies, and the provision of personalized medical care. The increasing use of WBANs in the medical field is attributed to the numerous advantages they offer in terms of patient monitoring and early detection of health risks. However, researchers face design challenges such as managing energy consumption, ensuring quality of service, and resilience to failures, all while ensuring patient comfort. In this study, we propose a new routing protocol called HEALTH for WBANs. This protocol aims to ensure quality of service in terms of latency and delivery rate while taking into account node energy consumption and temperature. We evaluated the performance of this protocol using the BNS and Castalia frameworks, both based on the Omnet++ simulator. Our simulation results demonstrate that our protocol is effective in conserving energy while maintaining good quality of service.
Wireless Body Network, Energy Efficiency, Quality of Service, Fault Tolerance, Routing protocol, Sensor, Sink.
Ezekiel Ologunde, Department of Computer / Cyber Forensics, University of Baltimore, Baltimore, Maryland, USA
Ransomware attacks have become a significant global threat that targets a broad spectrum of internet and mobile users, with critical cyber-physical systems particularly vulnerable. This unique characteristic of ransomware malware has drawn the attention of security professionals and researchers striving to develop safer, high-assurance systems capable of effectively detecting and preventing such attacks. This paper delves into the recent high-profile data breaches instigated by ransomware, such as those experienced by Colonial Pipeline, Kaseya, and MGM Resorts International. It further explores the role of social engineering in these incidents and emphasizes the importance of understanding the psychological triggers that make individuals susceptible to these tactics. This paper concludes by identifying potential areas for future research, highlighting how social engineering contributes to these attacks, including the impact of emerging technologies on ransomware attacks and the effectiveness of current cyber threat training programs.
Ransomware, Social Engineering, Cloud, Security.
Zhongxuan Xu1, Ziqi Zeng2, Ang Li3, 1Harvard-Westlake Upper School, 3700 Coldwater Canyon Ave, Los Angeles, CA 91604, 2Oaks Christian High School, 31749 La Tienda Rd, Westlake Village, CA 91362, 3Computer Science Department, California State Polytechnic University, Pomona, CA 91768
In response to the challenges faced by the blind and visually impaired, exacerbated by the increasing prevalence of screen-related visual impairments, this proposal aims to address the shortcomings of existing solutions. The proposed solution consists of a cost-effective and secure approach, featuring a suit equipped with sensors and motors,complemented by a mobile app. This integrated system enhances the navigational experience for the visually impaired,offering key features such as safe traversal in any environment, map-guided navigation, and timely warnings of high speed dangers. The suit utilizes time-of-flight sensors and vibration motors to convey crucial information to users,including distance, direction, and warnings, through distinct vibration signals. The mobile app serves as a valuable supplement to the suit, providing features such as traffic sign detection and advanced map navigation [8]. This cohesive blend of innovative hardware and software components establishes a comprehensive solution, addressing the urgent need for effective, affordable, and technologically advanced tools for the blind and visually impaired amidst the era of widespread screen usage.
IoT, Machine Learning, Computer Vision, Sensory Substitution.
Rasha Makhlouf, Department of Industrial Information Systems, Brandenburg University of Technology Cottbus-Senftenberg, Germany
Cloud computing is cementing its presence in IT strategies and CIOs’ priorities across industries. Companies have been heavily adopting different cloud services and integrating them into their IT portfolios. Meanwhile, research has been striving to understand, articulate, and explore the cloud ecosystem. Using 13 expert interviews and 3 case studies, and while applying transaction costs theories, this research contributes to the explicitness of cloud computing. This paper introduces a new component in the cloud computing ontology; Meta-Services. The paper provides a working definition for cloud Meta-Services and explains their inter-relationships with other cloud services and available ontologies. It also discusses four examples of meta-services that emerge during the cloud adoption journey. This discussion and conceptualization of Meta-Services will help cloud customers make more informed decisions about their entire cloud adoption journey. It will also help cloud providers make more realistic promises to their customers.
Meta-Services, cloud computing ontology, transaction cost theory, expert interviews, case studies.
Jack Li1, Karla Avalos2, 1Corona del Mar High School, 2101 Eastbluff Drive, Newport Beach, CA 92660, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
Education plays a pivotal role in fostering innovation and critical thinking. However, existing gaps in public education, particularly in complex subjects like chemistry, present challenges for students [4]. This paper addresses these challenges by proposing a 3D visualization solution aimed at elucidating the intricate atomic structures of molecules [5]. The software developed offers a nuanced understanding of challenging chemistry topics, facilitating enhanced comprehension for learners. Leveraging interactive 3D models, the program provides real-time updates and visualizations, fostering an engaging learning environment. The methodology involves refining and expanding the 3D model to cover various chemistry topics comprehensively. Challenges encountered during development were tackled through iterative improvements, resulting in a robust application. Experimental scenarios demonstrated the tools effectiveness in enhancing user satisfaction and understanding of chemical geometries. The paper concludes that this 3D visualization tool is a valuable resource for educators and learners, offering an innovative solution to bridge gaps in complex subjects and contribute to a more enriched learning experience.
Chemistry, Simulation, Geometry, Repulsion.
Zinuo Xu1, Soroush Mirzaee2, 1Windermere Preparatory School, 6189 Winter Garden Vineland Rd, Windermere, FL 34786, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768
This paper addresses the gap in financial literacy education for teenagers by introducing Landlord Legends, an innovative multiplayer video game designed to impart essential skills in property and money management [1][2]. Grounded in the recognition of underutilized potential in video games for education, the project draws inspiration from the diverse methodologies discussed in "Video Games in Education," "Using a Financial Education Curriculum for Teens," and "How Multiplayer Games Increase Social Closeness." [3] Landlord Legends provides an engaging platform for teenagers to learn and apply financial concepts while fostering social interaction. The challenges in dice interpretation and multiplayer system selection were overcome through meticulous design and research [4]. Experimental results showcased the games effectiveness, yielding increased financial literacy scores and positive participant feedback. Landlord Legends demonstrates the fusion of educational content and gaming, offering a dynamic and enjoyable approach to financial education that addresses challenges faced by teenagers, making it a valuable tool for enhancing financial literacy [5].
Interactive Learning, Educational Technology, Property Management, Multiplayer Games.
Austin Xiao1, Ang Li2, Wyatt Bodle3, 1Troy High School, 2200 Dorothy Ln, Fullerton, CA 92831, 2Computer Science Department, California State Polytechnic University, Pomona, CA 91768, 3California Baptist University, Riverside, CA 92504
Dangerous animal encounters have steadily increased over time and consumption of deadly plants is an important issue [12]. Our paper introduces a new mobile application that addresses the critical need for accurate animal and plant identification and classification to help mitigate safety risks for humans. With up to five million animal attacks reported every year in the United States alone, and over 100,000 cases of toxic plant exposure there is a need and a responsibility to increase awareness of the risks associated with animal and plant ignorance. Our proposed app utilizes innovative classification technologies, offering our users a swift and simple identification of both select plant and animal species. The app will relay information about the potential dangers and general facts about the classified animal [14]. This will help our users to understand the environment they live in and to best prepare themselves against it. Some challenges with this proposal are curating a broad and efficient dataset, there are estimated to be eight million eukaryotic species which is unattainable for one dataset. We then had to decide which valuable information would be best to present without providing unwanted distractions in our user interface. We utilized Google Firebase to ensure secure authentication and data storage while using TensorFlow Lite to power the image classification. We then integrated all of this into flutter to create a friendly user interface and application that can run on both iOS and Android [15]. Once our app was complete, we ran two experiments, one to test the accuracy of our classifications in plants and animals and another to test the effect of lower resolution images on classification accuracy. The experiments shed light on challenges and potential improvements for the application to help improve its efficiency as a tool for users to enhance their awareness, safety and understanding of the environment they live in.
Wildlife Identification, Plant and Animal Classification, Mobile Application, Biodiversity Awareness, Safety Technology.
