41530 research outputs found
Sort by
Transitioning the Cal Poly Quarter Horse Enterprise Curriculum: From Quarters to Semesters
No full textThe Cal Poly Quarter Horse Enterprise began in 1978, when student riders brought Cal Poly-bred horses to futurity events to sell them. In 1996, Gene Armstrong brought the sale to Cal Poly, and ever since, the annual performance horse auction has broken records as one of the highest grossing collegiate horse sales in the nation. This sale not only provides funding for the care of over one hundred horses housed at the Equine Center for educational use, but also provides an opportunity for students to gain experience in starting and training young horses, as well as putting on a dynamic public auction event. This valuable opportunity for students is deeply rooted in the agricultural industry, good animal husbandry, and the application of ethical and effective training practices. The complex nature of this program has led to the creation of a detailed curriculum timeline that ensures that equine training outcomes and student event planning are met on tight deadlines. With the impending transition of Cal Poly’s campus to a semester format, the program has been restructured. The Quarter Horse Enterprise benefits from clear curriculum guidelines that help students ensure they are on pace with learning outcomes and meet all learning criteria, from horse training to event production. As such, I identified the need for a new curriculum plan to be developed to reflect altered school session timelines. In collaboration with the instructor responsible for the Quarter Horse Enterprise, Lou Moore-Jacobsen, I am striving to create a day-by-day detailed course outline for each advancing section that will reflect industry best practices and relevant outcomes to ensure that the new structure of the enterprise will allow both students and horses to achieve both program and course learning outcomes in an efficient and comprehensive manner
RISC-V GPU Acceleration on Low-Cost Embedded Systems
No full textThe Vortex project from Georgia Institute of Technology was created to provide an open-source hardware and software GPGPU research platform based on RISC-V. Skybox was introduced as an extension to Vortex to provide dedicated support for 3D graphics rendering acceleration as a complete GPU platform. This work presents contributions to the render output unit of Skybox, including the development of a blend unit, cache bypassing mechanisms, and performance monitoring metrics. With these, Skybox has succeeded in its goal of accelerating graphics rendering on RISC-V platforms. This work also establishes a foundation for adapting Vortex to low-cost embedded platforms such as the PYNQ-Z2 (based on the AMD Zynq 7020 SoC), detailing boot configuration, host processor communication interfaces, and configuration tradeoffs for resource-constrained FPGAs
Teaching Oppositional Resistance to the Weaponization of Artificial Intelligence in Acts of Gendered Violence
No full textMeeting the aims of this special issue to take a uniquely feminist approach to understanding AI in higher education, we offer a critical commentary to help other scholars apply a critical feminist lens when teaching about the promise of artificial intelligence in the classroom. Specifically, we center pornographic deepfakes as a primal, contemporary, and disastrous example of the gendered weaponization of artificial intelligence. This critical commentary has staying power – equipping future teachers with feminist tools to tackle AI (as well as the inevitable future of AI to come) in pedagogical practice
Development of a Colorectal Cancer Spheroid Model to Improve the Accuracy of Preclinical Drug Testing
No full textColorectal cancer is the second leading cause of cancer related deaths worldwide. It currently affects millions of people across the globe and is only expected to increase in impact over the coming years. The most common treatment for metastatic colon cancer is chemotherapy, however, the development of chemotherapeutic drugs is a long and expensive process. A large portion of this development process is spent in preclinical drug testing. However, the models used often lack enough physiological relevance to guarantee the drug’s success in a clinical trial. Due to this gap in testing, researchers seek to develop a more physiologically relevant in vitro tumor model to better represent the interactions of a drug with the tumor microenvironment. The aim of this research was to develop a tri-culture tumor spheroid model that improved upon previous models’ physiological relevance. This was accomplished through a series of experiments. The first aimed to determine the optimal cell density and number of days in culture to develop a baseline tumor spheroid within the targeted size range. The next investigated how the addition of HUVECs would affect tumor spheroid formation and endothelial network formation. The third explored how the addition of HDFs would affect the same parameters. The fourth investigated the viability of spheroids in dense Matrigel to pursue the idea of culturing the spheroids in microfluidic devices. It also investigated the impact of a new method of HDF inclusion, a shell-like layer, on the model’s physiological relevance. The fifth explored the impact of the number of days before transfer as well as the effect of the location of additional HUVECs and HDFs plated in the extracellular matrix on the development of the spheroid model. The first three experiments established the ideal plating density for the spheroid and demonstrated that a tri-culture spheroid best mimics the tumor microenvironment. The fourth experiment revealed that the shell-like plating method improved both tumor spheroid shape and formation of angiogenic sprouts. The fifth experiment further refined the tumor spheroid model and established a protocol for tri-culture tumor spheroids that demonstrated physiologically relevant size, shape, and angiogenic sprouting. However, further work is required to foster these angiogenic sprouts into a perfusable vascular network to further increase the physiological relevance of this model for accurate preclinical drug testing
Accelerating Metal Extrusion Additive Manufacturing Process Development with Citrine Machine Learning Tools
No full textMaterial Extrusion (MEX) is a low-cost approach to metal additive manufacturing that involves extruding a metal-polymer composite filament to create a part comprised of polymer-bound metal powder, which is subsequently debound and sintered to produce a fully metal component. A major barrier to broader implementation of metal MEX is the time-intensive process for determining printing and sintering process parameters. To consistently produce sintered parts that satisfy geometric and materials properties requirements, trial-and-error experimentation is required to develop process parameters, correct shrinkage variability, and mitigate defects. This project aims to leverage Citrine Informatics’ machine learning (ML) tools to accelerate development of key aspects of the metal MEX process, including defect mitigation, shrinkage behavior, and sintering conditions, while still achieving materials properties that meet metal injection molding (MIM) standards. Successfully integrating ML with metallographic characterization and mechanical testing will enhance industrial viability of metal MEX by using predictive modeling to accelerate process parameter optimization, thereby reducing the number of printing and sintering hours required to produce a component that meets design requirements
AI-Powered Accessibility Tracker for Inclusive Public Spaces
No full textThis research project will develop and evaluate a smartphone-based, AI-powered system to crowdsource and analyze accessibility features and barriers in public spaces. Using computer vision and geospatial mapping, the system will identify and categorize issues such as uneven sidewalks, missing or inadequate curb ramps, damaged tactile paving, obstructive overhangs, and the absence of visual or auditory wayfinding cues. The overarching goal is to generate a dynamic, real-time accessibility map that empowers individuals with diverse mobility, sensory, and cognitive needs to navigate public spaces more safely and confidently. The project will integrate technologies and methods from applied machine learning, mobile computer vision, human-centered interface design, and participatory citizen science. The system will support multiple accessibility dimensions, including needs associated with wheelchair and stroller users, people with vision or hearing impairments, neurodiverse individuals, and older adults with endurance or balance limitations. Building on prior published works such as SmartCS (2024), a platform enabling codeless development of ML-powered apps for citizen science, this project will apply similar frameworks for data collection, annotation, and model deployment in an inclusive, community-driven context. The resulting application will feature an interactive, filterable map interface that visualizes both AI-detected and human-reported barriers, enabling both public users and civic planners to understand and prioritize accessibility needs. The student researcher will work closely with the faculty mentor on training and fine-tuning computer vision models based on deep neural networks, collecting and annotating accessibility-related datasets, prototyping and refining the mobile app, and coordinating a pilot study on the Cal Poly campus and in the City of San Luis Obispo (if time permits). Through this experience, the student will gain hands-on skills in inclusive technology development, interdisciplinary research, and ethical community engagement
15cm Gridded Ion Thruster
No full textThis research focuses on optimizing the configuration of a 13cm ion thruster for long term efficiency. Ion thrusters produce thrust for spacecraft when they are in orbit. The thruster needs to be efficient enough to last for the lifetime of the spacecraft as well as through its deorbit procedure. Thruster efficiency is achieved through optimizing the distribution of the high velocity beam of charged particles at the exit that impacts fatigue patterns on voltage grids. The shape of the beam is determined by the placement of magnets on the thruster walls. An experiment to test configurations for increasing efficiency through beam uniformity will be completed with a goal of documenting and publishing test results at an AIAA conference for use in the electric propulsion industry
Reinforcement Learning for Autonomous Parking
No full textThis research project proposes the development of an adaptive reinforcement learning (RL)-based parking system designed to handle complex parking maneuvers that are often unaddressed in existing research. Unlike previous works that focus on isolated maneuvers such as reverse parking or parallel parking, this work presents a flexible framework where multiple parking types (reverse, parallel, diagonal) are addressed through independent agents and then integrated into a cohesive system. The primary gap addressed by this work is the integration of diverse parking maneuvers into a unified RL framework with adaptability to dynamic and varied parking environments. Additionally, this project introduces curriculum learning to enhance training efficiency and domain randomization to improve robustness against environmental variations. This multi-maneuver capability allows for a broader, more scalable application in smart parking systems that can handle different vehicle types and parking scenarios without requiring retraining from scratch. Building on the foundational work of RL-based reverse parking, this project aims to generalize the approach to multiple parking tasks while enhancing robustness through novel training methodologies. The proposed system is trained using the Proximal Policy Optimization (PPO) algorithm within customized simulation environments using OpenAI Gymnasium and HighwayEnv frameworks. A key contribution of this research is the development of an adaptive control architecture that dynamically selects the appropriate maneuver type based on the parking scenario. The project will culminate in a comprehensive simulation-based evaluation demonstrating the system’s ability to effectively handle varied parking tasks with high success rates. The results will inform future extensions to real-world scenarios and contribute valuable insights for integrating diverse control tasks within a single RL framework