Jayanth Ramakrishnan and Parvendan Rangaswamy, Department of Computer Science and Engineering, SRM University-AP, Amaravati, Andhra Pradesh, India
Large Language Models (LLMs) have gotten really good at generating code, and we usually measure thatwith Pass@k. But Pass@k only tells you how often the model gets it right on its best try. It says nothingabout how much the outputs vary or whether they’re structurally consistent across multiple runs. That’s abig problem for real-world use, where you typically call the model just once.So we came up with the Stochastic Stability Index (SSI). It’s a combined metric that looks at three things:whether the code is syntactically valid (Msyn), how similar its structure is to a correct solution using thenormalised Zhang–Shasha Tree Edit Distance on Abstract Syntax Trees (Mast), and whether it actuallyworks when run in a safe sandbox (Mfun). SSI directly penalises variation from one sample to the next,giving you a single reliability score that reflects actual deployment risk, not just best-case performance.We tested SSI on HumanEval and MBPP using three top code generation models—Qwen, DeepSeek-Coder-v2, and Yi-Coder-9b—at three temperatures (T ∈{0.2, 0.5, 0.8}), generating 20 independent samples perproblem (that’s n = 96,840 programs total). What we found is a Creativity–Stability Paradox that shows upacross all models and both benchmarks: Pass@10 goes up as you raise the temperature, but SSI goes downas outputs become more variable. For Qwen on HumanEval, Pass@10 jumps from 86.8% to 94.0%, whileSSI falls from 0.6833 to 0.6563. A one-way ANOVA on structural similarity across every model–benchmarkcombination (in all cases p > 0.05) shows that the drop comes almost entirely from functional inconsistency,not structural divergence. That tells us something important about how to design better stability metrics. Thebottom line: if you really care about production-ready code generation, you need reliability-aware metricsalongside Pass@k.
Large Language Models,Code Generation,Stochastic Stability Index,Abstract Syntax Trees,Pass@k,Inference Temperature,Benchmark Evaluation,Output Variance
Zijing Dai1, Tyler Boulom2, 1Arnold O. Beckman High School, 3588 Bryan Ave, Irvine, CA 92602 2Woodbury University, 7500 N Glenoaks Blvd, Burbank, CA 91504
Deciding what to cook from available ingredients and judging when food is done are everyday obstacles that contribute to food waste and undercooking risk. This project presents ScanChef AI, a low-cost Raspberry Pi device with a camera and display that uses Google’s Gemini multimodal model to identify ingredients, generate recipes, and monitor cooking doneness in real time, paired with a Flutter companion app that communicates over a lightweight local HTTP server. Key challenges — constrained edge hardware, unreliable network responses, and unstructured AI output — were solved by offloading inference to the cloud, using background threads with short timeouts and exponential backoff, and prompting the model for strict, easily parsed formats. In experiments the system recognized common ingredients with 88.8% mean accuracy and tracked cooking doneness within a 5.5-point mean absolute error. By making capable, hands-free kitchen AI affordable and practical, ScanChef AI lowers the barrier to confident home cooking.
Large Language Models,Code Generation,Stochastic Stability Index,Abstract Syntax Trees,Pass@k,Inference Temperature,Benchmark Evaluation,Output Variance
Harry Zhang1and Rodrigo Onate2, 1JSerra Catholic High School, 26351 Junipero Serra Rd, San Juan Capistrano, CA 92675, 2California State University, Fullerton, 800 N State College Blvd, Fullerton, CA 92831
Soccer video review is a critical feedback tool for athlete development, but manually finding goals in match footage is prohibitively time-consuming for the amateur and youth coaches who make up the largest share of the sport’s 265 million players. This work introduces GoalVision, a Flutter mobile application and Flask Python backend that automate the task using a dual YOLOv11 pipeline — a pretrained YOLOv11n model for the ball and a custom model fine-tuned on the SoccerNet dataset for the goal — together with a novel depth-from-scale heuristic that interprets ball-size growth as goal-line crossing. Three design challenges around false positives, resolution invariance, and long-running backend jobs were addressed through cooldown logic, frame-area normalization, and Firestorestreamed progress. Across 50 annotated clips, the system achieved an F1-score of 0.89 at default settings and maintained ~0.87 F1 from 480p to 4K, demonstrating that a single phone can deliver previously inaccessible automated goal analysis at essentially zero hardware cost.
Soccer goal detection, YOLOv11, Computer vision, Object detection, Depth estimation, Mobile application
h6 style="color:black;font-family:classic wide,sans-cserif;font-size:20px">An AI-Supported Mobile Platform to Enhance Dementia Patient Engagement and Caregiver Wellness Using PhotoBased Reminiscence Puzzles and Real-Time Care CoordinationJiaxuan Wang1, Bobby Nguyen2, 1Mercersburg Academy, 100 Academy Drive, Mercersburg, PA 17236, 2University of California, Riverside, 900 University Ave, Riverside, CA 92521
Alzheimer disease affects millions of patients and burdens nearly 13 million unpaid caregivers in the United States alone, yet nonpharmacological support tools such as reminiscence therapy remain fragmented across paper memory books, standalone puzzle apps, and separate scheduling tools. MemoryWell proposes a single AI-supported mobile platform that turns family photos into interactive reminiscence puzzles, records positive “Shining Moments,” tracks caregiver mood and wellness, and coordinates care through a shared, real-time Care Circle. The system links a Flutter client, a Firebase real-time backend, and an OpenAI GPT-4o-mini-AI layer that generates personalized encouragement, pattern insights, and wellness tips. Implementation challenges included calibrating puzzle accessibility, constraining AI-generated content, and securing a shared real-time backend. Two experiments showed that puzzle difficulty grows super-linearly with grid size, validating the shipped 3×3 default, and that AI content reliability degrades directly with API failure rate due to an absence of retry logic. MemoryWell offers a more integrated, personalized alternative to prior single-purpose reminiscence, wellness, and coordination tools.
Dementia Care, Reminiscence Therapy, Mobile Health (mHealth), Artificial Intelligence, Large Language Models
h6 style="color:black;font-family:classic wide,sans-cserif;font-size:20px">Medical Software Validation in the AI Era 2026: A Case Study of Vectors AI System Medical PlatformShay Ginsbourg1, 1Vectors AI System, Vectors.co.il, Israel
This paper explores Artificial Intelligence (AI) and Natural Language Processing (NLP) applied to medical software validation. Focusing on Vectors AI System’s WPS-Medical platform, an internal AI-assisted Software Validation and Verification (V&V) system, we demonstrate how local large language models and structured AI agents support complex regulatory processes. Key benefits include improved traceability, enhanced documentation efficiency, and better adherence to standards like IEC 62304, ISO 14971, and FDA 21 CFR Part 820. Early internal pilots indicate efficiency gains while preserving human expert oversight. This paper provides educational insights for validation engineers and medical device developers navigating AI-enabled systems.
Artificial Intelligence; Natural Language Processing (NLP); Medical Software Validation (V&V); Software 3.0 Paradigm; Karpathy Wiki Model; Agentic LLMs; IEC 62304 Compliance; Requirements Traceability Matrix (RTM); Self-Validating Autonomous Systems; Automated Regulatory Gap Analysis; Privacy-Preserving AI
Evariste Ntaryamira1, Grace Murhula Kabi2, 1Université Paris 8, Saint-Denis, France, 2Institut Supérieur de Commerce (ISC), Bukavu, République Démocratique du Congo
Educational Management Information Systems in Sub-Saharan Africa collect data from thousands of schools each year to feed the indicators required by Sustainable Development Goal 4 and the Continental Education Strategy for Africa. A recurring problem, however, undermines this process: as data travels up the administrative pyramid, each echelon that handles a file reformats it, reinterprets column names, and silently alters values. By the time a record reaches the national level, it often says more about the chain of people who handled it than about the school that produced it. This paper introduces a multi-layer semantic mediation architecture that addresses the problem at its root. Positioned between school-level data sources and the national server in a Hub-and-Spoke topology, the architecture comprises six formally defined processing layers: ingestion, statistical profiling, adaptive schema matching, coherence validation, incremental dictionary learning, and transactional synchronization. We formalize the mediation process as a structure-preserving functor between schema categories and prove its correctness under composition. Three original algorithms are proposed and analyzed: AdaptiveSchemaMatch, which combines TF-IDF lexical analysis, distributional divergence, and value-set overlap into a weighted similarity measure with gradientbased weight adaptation; IncrementalDictLearn, a reinforcement-inspired dictionary learner with provable convergence; and CoherenceGuard, a multi-tier consistency engine. We further show how cross-lingual sentence embeddings and lightweight autoencoders can augment the matching and anomaly detection capabilities in data-rich settings, while remaining deployable on resourceconstrained devices. The architecture is designed for offline-first, low-connectivity environments and scales to national-level volumes through a formally analyzed batch-parallel processing model. We describe the deployment context within the PAQABU project (UNESCO, Burundi) and the eSchool platform (Democratic Republic of Congo), and discuss the implications for educational data governance across the continent.
Semantic mediation, schema matching, educational data, Hub-and-Spoke, EMIS, incremental learning, data quality, Big Data, deep learning, Sub-Saharan Africa
Kazim Ali Korejo, Ana Maria Martinez-Enriquez, and Faaiza Zulfqar, Computer Science, Cinvestav-IPN, Av Instituto Politecnico Nacional 2508 Zacatenco, 07360, Mexico City, Mexico
Generative adversarial networks have emerged as a significant paradigm in Deep learning for Natural Language Processing, enabling improvements in multilingual learning, semantic modeling, text generation, and multimodal synthesis, to name a few. Our research aims to analyze the underlying learning mechanisms and the architectures used for different applications, such as paraphrase generation, text-toimage synthesis, machine translation, data augmentation, sentiment analysis, and noise reduction. Besides, our study analyzes the integration of Transformer with multilingual adaptation, focusing on limitations such as training instability, collapse mode, unreliable evaluation, and semantic inconsistency. Lastly, we outline opportunities for the way forward involving retrieval-augmented, explainable multimodal semantic alignment, symbolic reasoning, and hybrid large language model frameworks for more reliable, semantically grounded language generation.
GANs, Large Language Models, Multi-Agent, Generative Adversarial Networks, GAN-Variant.
Zihan Xu1 and Jason Moya2 1BASIS International School Hangzhou, No. 9 Yulin Road, Shangcheng District, Hangzhou, 310016, China 2California State Polytechnic University, Pomona, 3801 W Temple Ave, Pomona, CA 91768
Young athletes face elevated rates of anxiety, stress, and performance pressure, yet rarely access traditional sportpsychology services because of stigma, cost, and time scarcity during peak training periods. Athlete Emotion Manager is a Flutter and Firebase mobile application that consolidates four evidence-supported techniques — box breathing, guided visualization, grounding, and curated educational articles — into a single self-regulation platform built around three components: a Firebase-backed authentication system, a longitudinal mood-tracking and visualization system that surfaces a personalized trend chart, and a progress-and-streak engagement system designed to reinforce consistent practice without manipulative patterns. A ten-day in-the-wild trial with 24 competitive athletes showed that low-friction exercises such as box breathing dominated actual usage, and that more than 75% of participants rated every delivered technique as somewhat or very helpful. The application demonstrates that a focused, mood-aware mobile tool can meaningfully extend the reach of athlete mental-wellness support.
Athlete mental health, Sports psychology, Emotional self-regulation, Mobile health (mHealth), Firebase.
Vivienne Zhai1and Quincy Stokes2 1St. Margarets Episcopal School, 31641 La Novia Ave, San Juan Capistrano, CA 92675 2University of California, Irvine, Irvine, CA 92697
Many athletes struggle with recovery, rehabilitation, and long term wellness after injuries or physical strain. Athlete Well was created to provide athletes with recovery guidance, wellness education, and progress tracking in one organized platform. The app includes three major systems: recovery plans, progress tracking, and educational wellness resources. These systems help users follow rehabilitation exercises, monitor improvement over time, and learn about injury prevention, mobility, nutrition, rest, and mental wellness [1]. One challenge during development was organizing different recovery resources into a simple and easy-to-use interface. To evaluate the app, I conducted a survey experiment where users tested Athlete Well and rated its usability and helpfulness. The results showed that most users found the app organized, understandable, and useful for recovery support [2]. Athlete Well is important because it encourages athletes to recover safely, understand their bodies better, and build healthier long-term recovery habits instead of focusing only on athletic performance.
Sports Recovery, Rehabilitation Tracking, Athlete Wellness, Mobile Health Application
Junde Zhu1, Rodrigo Onate2, 1Army and navy academy, 2605 Carlsbaad Blvd Caarlsbad, CA 92008, California State University, Fullerton, 800 N State College Blvd, Fullerton, CA 92831
Food waste remains a significant problem because many people are unsure how to use ingredients already available in their kitchens [1]. This often results in unnecessary grocery purchases, wasted food, and unhealthy dining habits. To address this issue, this project developed YummySnap, a mobile application that uses artificial intelligence to identify ingredients from photographs and generate personalized recipes with nutritional information. The system was built using Flutter, OpenAI image analysis and language models, and local storage through shared_preferences [2]. Three major components were implemented: ingredient recognition, recipe generation, and recipe history management. Several challenges were considered, including image quality, user interface design, and data persistence. Experimental testing showed that image clarity strongly affects ingredient detection accuracy, while user preferences significantly improve recipe relevance. These findings demonstrate that combining computer vision with personalized AI recommendations can create a practical tool for reducing food waste and encouraging healthier home cooking.
Artificial Intelligence, Health, Simple, Waste management
Haoyu He1, Jonathan Sahagun2, 1Sage Hill School, 20402 Newport Coast Dr, Newport Coast, CA 92657, 2California State University, Los Angeles, 5151 State University Dr, Los Angeles, CA 90032
Compulsive “doomscrolling” harms attention and wellbeing, yet infinite feeds remove the natural cues that would prompt a user to stop. Focusrail is a wearable inertial sensor, worn on the hand, that detects the motion of scrolling and alerts a connected phone in real time. An LSM6DSOX IMU streams accelerometer and gyroscope data at 50 Hz over Bluetooth Low Energy to a Flutter application, which runs a 12-sample sliding-window detector and fires a gentle alert with a cooldown to avoid fatigue, logging each event for weekly review. Key challenges — separating scrolling from ordinary motion, reliable low-power streaming, and alert fatigue — were addressed through dual acceleration/gyroscope thresholds, BLE with auto-reconnect, and a suppression interval. Experiments showed scrolling is cleanly separable from resting, typing, and walking, and that the chosen gyroscope threshold sits at the optimum of the precision–recall curve. By cueing the behavior the instant it begins, Focusrail enables more intentional phone use.
Doomscrolling, Wearable sensing, Inertial measurement unit (IMU), Bluetooth Low Energy (BLE), Flutter
Andrew Park, University of California, USA, Jenny Zhao, Flintridge Sacred Heart Academy, USA
Children remain among the most vulnerable members of society, and current safety tools are poorly suited to younger children who do not own smartphones or who cannot reliably advocate for themselves in an emergency. We present Watchful Trustline, a wearable safety necklace system that pairs a Particle Boron 404X cellular microcontroller, a BNO085 inertial measurement unit, and a PA1010D GPS receiver with a Flutter companion application backed by Firebase Authentication and Realtime Database. The necklace continuously streams location every 30 seconds and publishes immediate alerts when its accelerometer detects a free-fall to impact pattern, and the application surfaces both data streams to the guardian in real time. We tested the system across urban and remote environments, observing end-to-end fall-alert latencies between three and twenty-two seconds and GPS error consistently under twelve meters. Watchful Trustline offers families a child-specific, always-on safety tool that smartphones and consumer trackers cannot replace.
Child Safety, Flutter, GPS, Safety Necklace, IoT
Jason Moya, California State Polytechnic University, USA, Tyler Kaiyang Chen, Northwood High School, USA
Falls are the leading cause of injury among adults over 65, and their consequences worsen sharply when help is delayed. This paper presents StrideGuard, a low-cost wearable system that both detects falls and predicts instability before they occur. A QT Py ESP32-S3 microcontroller samples an LSM6DSOX accelerometer at 100 Hz, detecting falls through a freefall-plus-impact threshold algorithm and scoring gait instability with a sliding-window analysis of movement smoothness and irregularity. Alerts travel over Bluetooth Low Energy to a Flutter application that raises alarms, charts live instability scores, and logs events to Firebase for caregiver review. In evaluation, the instability score separated simulated shuffling from normal walking (mean 6.21 versus 0.90) with zero false warnings after debouncing, and the fall detector identified 85% of sixty instrumented falls with only a 15% false-alarm rate on its most challenging control activity. StrideGuard shows that predictive, automatic fall monitoring is achievable at consumer prices.
Fall detection, Gait instability, Wearable sensors, Accelerometer, LSM6DSOX, ESP32-S3
Jiaxuan Wang1, Bobby Nguyen2, 1Mercersburg Academy, 100 Academy Drive, Mercersburg, PA 17236 2University of California, Riverside, 900 University Ave, Riverside, CA 92521
Alzheimer’s disease affects millions of patients and burdens nearly 13 million unpaid caregivers in the United States alone, yet nonpharmacological support tools such as reminiscence therapy remain fragmented across paper memory books, standalone puzzle apps, and separate scheduling tools. MemoryWell proposes a single AI-supported mobile platform that turns family photos into interactive reminiscence puzzles, records positive “Shining Moments,” tracks caregiver mood and wellness, and coordinates care through a shared, real-time Care Circle. The system links a Flutter client, a Firebase real-time backend, and an OpenAI GPT-4o-mini-AI layer that generates personalized encouragement, pattern insights, and wellness tips. Implementation challenges included calibrating puzzle accessibility, constraining AI-generated content, and securing a shared real-time backend. Two experiments showed that puzzle difficulty grows super-linearly with grid size, validating the shipped 3×3 default, and that AI content reliability degrades directly with API failure rate due to an absence of retry logic. MemoryWell offers a more integrated, personalized alternative to prior single-purpose reminiscence, wellness, and coordination tools.
Dementia Care, Reminiscence Therapy, Mobile Health (mHealth), Artificial Intelligence, Large Language Models
Arianna Lai1, Han Tun Oo2, 1University High School, 4771 Campus Dr, Irvine, CA 92612, 2Purdue University, 610 Purdue Mall, West Lafayette, IN 47907
Small and medium-sized enterprises contribute a disproportionate share of global pollution yet lack accessible tools to evaluate materials, locate sustainable suppliers, and avoid greenwashing. GreenLaunch is a Flutter mobile application that addresses these gaps by combining a curated Cloud Firestore database of material sustainability ratings, an OpenAI GPT-4o-mini analyzer that returns structured eco-evaluations for arbitrary materials, a Google Places-driven map of nearby recycling centers, organic stores, and eco-friendly businesses, and a community feed with notifications and reporting moderated through Firebase [13]. Key challenges — LLM hallucination, geolocation accuracy, and feed data quality — were mitigated through prompt constraints, fallback default coordinates, and authenticated reporting workflows [14]. Across two experiments, the AI analyzer agreed with the curated baseline within 0.31 points on a 1–10 scale, and the Places search returned an average of 56 sustainable resources per city, demonstrating that the design is robust enough for everyday use by prospective and established green entrepreneurs.
Green Entrepreneurship, Mobile Application, Flutter, Firebase, Large Language Models
Richard Luo1, Tyler Boulom2, 1La Jolla Country Day School, 9490 Genesee Avenue La Jolla, CA 92037, 2Woodbury University, 7500 N Glenoaks Blvd, Burbank, CA 91504
Standard alarms only make noise and can be silenced from bed, so snoozing fragments sleep and worsens sleep inertia, a problem linked to poorer alertness and long-term health risks. This project, Rumble Pulvinar, is a smart alarm that cannot be dismissed until the user physically leaves their pillow. A velostat pressure sensor under the pillow is read by an ESP32-S3 running CircuitPython, which computes a pressure percentage relative to a calibrated baseline and streams it over Bluetooth Low Energy to a Flutter app that owns the alarm [9]. Noise was tamed with sample averaging, exponential smoothing, and debouncing; reliability was secured with a self-healing BLE link; and cheating was prevented with a get-up countdown that resumes the alarm if the user lies back down. Detection separated occupied from empty pillows with about 99% accuracy, and the enforced alarm cut average time-in-bed from about 17 minutes to 2.4, demonstrating a simple, low-cost way to actually get people out of bed.
Smart alarm, Sleep inertia, Velostat pressure sensor, Piezoresistive sensing, Bluetooth Low Energ
Copyright © EDTECH 2026