Researchers at North Carolina State University have developed a comprehensive high school curriculum spanning a week. This curriculum aims to assist students in promptly grasping the fundamentals of color chemistry and artificial intelligence, sparking their interest in science and the world around them.
The researchers devised a relatively simple experiment centered on pH levels, which indicate the acidity or alkalinity of a liquid solution, to gauge the effectiveness of teaching students about both artificial intelligence (AI) and chemistry—a notoriously challenging subject.
Utilizing color transformation charts proves to be beneficial for testing pH levels on a test strip. These charts demonstrate that alkaline solutions cause test strips to darken when acidity is present, shifting to yellow and green as acid levels decrease. Conversely, acidic solutions cause the test strips to turn blue and dark green as acid levels decrease and deep crimson in highly alkaline liquids.
The pH scale ranges from 0 to 14, with seven representing neutrality—similar to the pH of tap water. Lower pH levels indicate higher acidity, while higher levels indicate greater alkalinity.
The researchers posed the question, “Can machine learning enhance the precision of reading pH levels compared to visual analysis?” Yang Zhang, an associate professor specializing in cotton architecture, chemistry, and science and a co-author of the study, stated that the AI predictive model trained by students proved to be approximately 5.5 days more accurate than manual readings.
Following the wetting of pH test strips in various common liquids such as beverages, pond water, makeup, etc., individuals used their smartphones to predict the acidity levels creatively. Additionally, the instructors provided test strips with known pH levels obtained through advanced equipment for comparison.
Zhang emphasized the practical implications of such tests, especially in areas with water quality concerns. The simplicity of determining whether the pH level is below 5 or 7, even without sophisticated equipment, was highlighted.
Students inputted data into Orange, a user-friendly machine learning tool suitable for novices without any coding requirements. Machine learning enhanced accuracy by discerning subtle color variations on the test strips corresponding to different pH values, converting images and pH levels into projections.
Upon comparing their manual and machine learning-based pH predictions, students noted that while not perfect, the AI predictions were notably more precise than visual assessments.
Pre- and post-curriculum surveys revealed that students reported increased enthusiasm and knowledge in both science and artificial intelligence. Shiyan Jiang, an associate professor of learning design and engineering at NC State and another co-author, highlighted how students recognized the relevance of cutting-edge technology in addressing real-world challenges and scientific advancements.
This innovative program not only deepens students’ understanding of the fusion of advanced technology and science but also motivates them to explore creative solutions, fostering a profound appreciation for these interdisciplinary intersections.
Zhang also mentioned the potential to expand such educational initiatives to include a broader audience and cover additional concepts in color chemistry, opening up opportunities for more students to benefit from this engaging approach.
