CAP: Artificial Intelligence Research Center for Preserving the Environment with a Focus on Underserved Communities

This project establishes an artificial intelligence (AI) research center at Savannah State University (SSU) and provides opportunities for the students to use AI to tackle environmental issues. The AI research center for preserving the environment (AIRCPE) is designed to: 1) improve AI and machine learning (ML) research infrastructure at SSU for future partnerships; 2) provide applied training for the students from underserved communities to develop AI skills; 3) offer career and professional development activities for the students to connect them with the AI market; and 4) conduct use-inspired AI research with a focus on pollution control in three environmental mediums (i.e., water, soil, and air).

This AI center increases the general research capacity at SSU and actively engages students and faculty members with different backgrounds in interdisciplinary research to address real-world environmental problems using AI-powered innovations.

The main project uses a number of AI and ML techniques, such as deep learning and hybrid models, to analyze the available data in these three environmental media, with the goal of pollution control. This project develops tiny machine learning (TinyML) models based on the available data to work on microcontrollers. To develop the TinyML models, deep learning algorithms such as deep neural networks, convolutional neural networks, and long short-term memory networks are trained on the numerous data sets.

The training of the models includes training from scratch using robust data sets and transfer learning, which involves fine-tuning a pre-trained model. TensorFlow is the main machine learning platform for the design of original deep learning models. TensorFlow Lite decreases the size of original deep learning models and makes them adaptable for devices with limited computational capabilities.

TensorFlow Lite Micro, the core-enabling technology for TinyML, is employed to further compress the models so that they work on microcontrollers. Such intelligent embedded systems play a major role in the monitoring of environmental contaminants and building novel treatment technologies.This project is co-funded by the Historically Black Colleges and Universities Undergraduate Program (HBCU-UP), which provides awards to strengthen STEM undergraduate education and research at HBCUs.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.