We present an introductory end-to-end data science project that involved first-year students to implement a machine learning project. Students collected pictures of more than 20 common glassware in a Chemistry laboratory to perform a supervised transfer learning project. They utilized four pre-trained neural networks, Inception-V1, Inception-V3, ResNet-50, and ResNet-101 to distinguish between the glassware. Students learned basic concepts of machine learning, neural networks and applications of machine learning in Chemistry and related domains as part of this experience. This project was carried out using the Wolfram language and the pre-trained neural networks were obtained from the Wolfram Neural Net Repository. The classification results achieved 99% accuracy in the identification of glassware and provided a great experience for students to experience advanced computing resulting in a highly tangible output. Students compared the efficacy of the different neural networks and analyzed the overall performance using multiple commonly used metrics. The project was implemented in the author’s Introduction to Scientific Computing course that provides students with an introduction to Wolfram language and its applications across multiple scientific domains. We believe that this project is highly accessible and adaptable to students and instructors at other institutions using Wolfram language or other programming languages.

Image recognition for laboratory glassware: A data-science and machine learning project for chemistry students

ACS Spring Meeting 2021

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.