USA Professor Receives $155,000 Grant for Artificial Intelligence Research
Posted on September 18, 2025
Dr. Jesse Ables, an assistant professor at the 秘密研究所鈥檚 School of Computing, has been awarded a $155,000 grant from the National Science Foundation to show exactly how and why artificial intelligence determines its decisions.
His funded project combines two critical areas of cybersecurity and artificial intelligence, Intrusion Detection Systems, which safeguard computer networks, and Explainable Artificial Intelligence.
鈥淎 significant drawback of modern artificial intelligence is its inherent opacity. Models like Deep Neural Networks and Large Language Models are so complex that we don't fully understand how or why they make a specific prediction.鈥� Ables said. 鈥淢y research addresses this opacity problem in AI. By generating explanations for a model's predictions, we can begin to understand why it's making certain decisions, both correct and incorrect.鈥�
Ables is developing a new algorithm called Temporal Eclectic Rule Extraction. Its goal is to create explanations for a specific type of powerful AI called a Recurrent Neural Network. The algorithm will show the reason behind its predictions.
鈥淭here is the optimization component of the research project. It's very fast with a small amount of data but quickly slows down as more data is used,鈥� he said. 鈥淚n my previous work, we found we could reduce the number of data samples to a certain extent while still maintaining high accuracy. The goal is to make this algorithm as fast as possible so that it can be used in decision-sensitive domains.鈥�
Ables says the idea of AI making decisions alone is difficult to trust and wants a collaboration between the Explainable AI system and the human user. The system provides the user with predictions and explanations. The human then uses that information, along with their own expertise, to perform their tasks. Humans can even use those insights to alter the AI, improving its decision-making for future tasks.
鈥淚n general, all industries would benefit from Explainable AI,鈥� Ables said. 鈥淗owever, decision-sensitive domains like healthcare, national security and finance would benefit the most from it. These are areas where a single decision can have life-altering effects.鈥�
