Utilizing Online and Open-Source Machine Learning Toolkits to Leverage the Future of Sustainable Engineering
Project Overview
The document explores the impactful integration of generative AI and machine learning in engineering education, specifically within civil and environmental engineering, aligning these technologies with the United Nations Sustainable Development Goals (SDGs). It presents various case studies demonstrating the application of AI tools in the classroom, which engage students through practical projects such as air quality detection, roadside litter detection, automated bird identification, and wildlife camera trap monitoring. These initiatives not only enhance hands-on learning but also address real-world environmental challenges. Additionally, the document discusses the difficulties faced when teaching machine learning concepts to non-computer science students, underscoring the necessity for innovative pedagogical strategies to effectively convey complex topics. By showcasing the successful incorporation of AI in educational settings, the findings suggest a promising avenue for enhancing student engagement and fostering a deeper understanding of sustainable practices through technology.
Key Applications
Environmental Monitoring and Species Detection
Context: Environmental Engineering classrooms targeting students interested in environmental science, conservation technology, and community science. The applications are designed for both freshman and junior-level engineering students, as well as students studying conservation and community engagement.
Implementation: Utilizing machine learning tools such as Edge Impulse, alongside various sensor technologies including Raspberry Pi, accelerometers, and camera traps, to collect and analyze data related to air quality and wildlife. Students engage in hands-on projects that involve data collection, training models, and using applications like Merlin Bird ID for species identification.
Outcomes: Students gain practical experience in data collection, analysis, and conservation technology, enhancing their understanding of environmental issues and biodiversity. Achievements include high accuracy in litter detection and community science participation, fostering awareness of waste management and wildlife conservation.
Challenges: Challenges include limited access to equipment, the need for data labeling, dependence on open-source data, and ethical considerations regarding data collection and security for camera placements.
Implementation Barriers
Educational Barrier
Challenges in teaching machine learning to non-computer science students due to varying levels of programming knowledge and difficulty in understanding complex machine learning libraries and tools.
Proposed Solutions: Provide structured coding guidelines, clear infrastructure recommendations, and encourage the use of simple, digestible platforms that do not hide complexity.
Resource Barrier
Limited access to powerful computing devices for students outside of computer science.
Proposed Solutions: Utilize cloud services that offer free credits for educational purposes.
Project Team
Andrew Schulz
Researcher
Suzanne Stathatos
Researcher
Cassandra Shriver
Researcher
Roxanne Moore
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Andrew Schulz, Suzanne Stathatos, Cassandra Shriver, Roxanne Moore
Source Publication: View Original PaperLink opens in a new window
Project Contact: Dr. Jianhua Yang
LLM Model Version: gpt-4o-mini-2024-07-18
Analysis Provider: Openai