Hands-on AI Technology Development Experience on Industry-driven Research Problems related to Resilient and Sustainable Forest Product Industries at Mississippi State University, Starkville, MS
This is a U.S. Department of Agriculture (USDA) National Institute of Food and Agriculture (NIFA) funded Research and Extension Experiences for Undergraduates (REEU) project. If you have any questions about this project, please feel free to contact:
Dr. Jason Street, jts118@msstate.edu
USDA AI2F REEU Summer 2025 Program will be between May 27, 2025 and August 1, 2025
at Mississippi State University, Starkville, MS
10-week hands-on research and extension experiences in solving real-world forest-product industry problems using state-of-the-art artificial intelligence (AI) technologies
Weekly seminars/webinars about various research disciplines, field trips, technical writing, presentation skills, graduate school preparation, etc.
Selected students will receive a $6,000 package (including a stipend, housing, transportation, and meal allowances)
Must be enrolled in a community college or 4-year university program
U.S citizen or permanent resident
3.0 or above GPA
Please send the following materials to Dr. Jason Street at jts118@msstate.edu by the deadline of April 21st.
REEU Application Form: Download the Summer 2025 Program Application Form
Resume
Unofficial Transcript(s)
Notification of Selection Extended: April 25, 2025
Faculty Mentors: Marufuzzaman, Street, Wang
Opportunities: Students will gain hands-on experience developing a cost-effective, computationally sound, industry-scale image-based (contact-free) tool to rapidly assess woodchip and sawdust characteristics (e.g., ash, moisture content, particle size distribution) in real-time. The tool will offer an alternative solution to rapidly assess feedstock quality, an ongoing problem for industries that rely on woodchips and sawdust as primary feedstock. The students will learn how to store a large volume of collected images (~10,000 RGB) and feed them into a convolutional neural network-based deep learning (DL) architecture.
Faculty Mentors: Street, Wang, Ragon
Opportunities: The desired moisture of timbers before being treated is approximately 25% to allow for the treatment chemical to easily penetrate the wood cells. However, if the timbers are too wet or too dry, the efficiency of introducing the chemicals into the wood using a treatment cylinder is reduced. Students involved in this project will collect data (e.g., taking images of the timbers and gathering soundwave data), assist with developing and testing deep learning architecture to determine the moisture content of the timbers, create a mobile application, and prepare the extension report. This technology would alleviate the tedious process of physically boring wood to obtain a sample, which is then used to determine the moisture content.
Faculty Mentors: Ragon, Fortuin, Marufuzzaman
Opportunities: Automatically determining and packaging fenceposts with a similar diameter in a realistic, fast-paced industrial environment is an important research problem. The posts must all be cut to an equivalent length but are not peeled to equal diameters. With ongoing manual efforts, industries lose both valuable production and labor time. Students involved in this project will be responsible for gathering data (taking post images, measuring the actual diameter of the posts, etc.) and assisting with writing the code for the deep learning model to predict the classification of the posts.
Faculty Mentors: Fortuin, Tian, Wang
Opportunities: Early detection and quick identification of damaging forest insects can be critical for protecting timber resources and wood quality. Students will gain hands-on experience in developing AI tools for identifying key forest insects from images. Students involved in this research will gather image data from online taxonomic databases as well as utilizing collections at the Mississippi Entomological Museum and the Forest Disturbance, Insect and Conservation Ecology lab to train an AI model in forest insect recognition and identification. Students will develop skills in insect identification and have the opportunity to create a mobile application for forest insect ID for end users to assist in.
Faculty Mentors: Wang, Ragon, Tian
Opportunities: Students will gain experience in using state-of-the-art Generative AI models to create synthetic images and data representations of wood product samples. This will enhance wood quality assessment, defect detection, and new material simulation for the forest products industry. The student will learn how to use AI-based models to augment real-world wood product datasets for training machine learning algorithms, how to improve defect detection and wood classification by generating diverse and realistic wood texture patterns, and how to simulate aging, weathering, and quality variations in wood products for predictive analytics.
Associate Profesor, and Charles R. Stephenson Endowed Professorship
Department of Industrial & Systems Engineering, Mississippi State University
Supply chain optimization with applications in renewable energy, stochastic programming, decomposition methods, solving large scale supply chain network problems and supply chain risk management.
Associate Professor
Department of Sustainable Bioproducts, Mississippi State University
Renewable/alternative energy; Catalysts
Assistant Extension Professor
Department of Sustainable Bioproducts, Mississippi State University
Wood products
Assistant Professor
Forest and Wildlife Research Center
Invertebrate biodiversity in managed forests;
Pesticide risks to beneficial non-target Invertebrates;
Impacts of severe wind disturbances in managed forests;
Associate Professor
Department of Industrial & Systems Engineering, Mississippi State University
Advanced sensing and analytics for system modeling, monitoring, and diagnosis. Applications include manufacturing and healthcare systems.
Assistant Professor
Department of Industrial & Systems Engineering, Mississippi State University
Machine learning modeling and optimization with high data heterogeneity; Image-based machine learning model interpretability; Information fusion; Ensemble learning