Picture of me in the Smart Food and Agricultural Engineering (SAFE) Lab
Author: Samuel Vinson | Major: Biological Engineering, Spanish | Semester: Spring 2025
Inconsistent beef grading practices have long posed economic challenges, created inefficiencies within the industry, and undermined consumer trust. Addressing these critical issues, my honors thesis explores how cutting-edge artificial intelligence can enhance the consistency and accuracy of beef marbling assessments across multiple processing facilities. Specifically, my research utilized unsupervised domain adaptation regression, a sophisticated machine learning method designed to bridge performance gaps among diverse grading environments.
This approach significantly reduces grading variability by minimizing the subjective inconsistencies often observed with traditional human evaluators. Historically, grading methods have relied heavily on visual inspection by USDA-certified graders, making the system vulnerable to fatigue, bias, and environmental differences between facilities. By employing convolutional neural networks, particularly ResNet-50 architecture, my project achieved improved generalization across facilities with varied lighting, equipment, and operational protocols. Statistical analyses confirmed substantial improvements over conventional techniques, underscoring the practical potential of this innovative approach.
My choice of this research topic was deeply influenced by my personal background. Growing up visiting my grandparents’ cattle farm in Siloam Springs provided a firsthand view of the challenges associated with inconsistent beef grading. This personal connection, coupled with my academic interests in artificial intelligence, agriculture, and computer vision, inspired me to seek a research mentor. Connecting with Dr. Dongyi Wang, who was actively working on a USDA-funded research project, allowed me to align my academic goals with meaningful, real-world applications.
Through this project, I gained valuable insights into the historical development and current challenges of beef grading technologies, reinforcing my understanding of artificial intelligence and its real-world implications. My programming skills notably advanced as I navigated through software frameworks such as Python and PyTorch, essential tools in contemporary AI research. One of the most significant hurdles I encountered was mastering domain adaptation, a relatively advanced topic which is typically covered in graduate coursework. Overcoming this obstacle involved extensive independent research, intensive review of academic literature, and iterative experimentation, significantly enhancing my analytical and technical capabilities.
My mentor, Dr. Wang, was integral to my research journey, consistently providing support and expertise throughout the project. His guidance extended beyond technical assistance, encouraging me to pursue external funding opportunities and pushing me to expand my professional network. Collaboration with peers in the Smart Food and Agricultural Engineering (SAFE) Lab also provided essential support, fostering a collaborative environment and facilitating a deeper understanding of practical research challenges.
A major milestone of my research will be presenting my findings at the ASABE Annual International Meeting in Toronto, Canada, this summer. Participating in this international conference offers a unique opportunity to share my work with industry leaders and academics, gather invaluable feedback, and connect with professionals in biological and agricultural engineering from around the world.
Looking forward, I am excited about applying the knowledge and techniques acquired through this research during my summer internship at Walmart. As a Process and Quality Engineer, the skills developed through domain adaptation methodologies will significantly contribute to my role, specifically in improving operational efficiency and consistency across Walmart’s vast network of distribution centers. Overall, this research experience has fundamentally shaped my academic journey, reinforcing my commitment to addressing real-world challenges with innovative engineering solutions.