Skip to main content

Improving the future of AI

By Joe McAdory

The future of artificial intelligence includes autonomous vehicles, doctors and surgeons, and even helping authorities identify criminals. Though AI is transforming disciplines from transportation to policing to healthcare, AI-based systems often make mistakes when facing scenarios that they have never seen before.

Assistant professor Anh Nguyen earned a $460,736 Faculty Early Career Development Award by the National Science Foundation for his five-year project, “Harnessing external knowledge to improve computer vision robustness, explainability, and user accuracy.”

Funds allow his project to create the first K-6 artificial intelligence club in Alabama, explore creative ways to implement AI research into a new course at Auburn University, establish collaborative efforts with industry partners and educate external audiences by publishing reproducible code, public-oriented videos and explore new ways AI can reach better decisions, and ultimately provide better outcomes. 

Poor AI decision-making can have disastrous consequences.

The solution? Build AI-based systems that harness external sources of knowledge to make more informed and accurate decisions. How? Utilize large text and image datasets, including Wikipedia and Google, for image classifiers to leverage and make better decisions. Humans often rationalize data before making decisions. Why not AI? The research goal is to shift the paradigm, where AI-based systems make deliberate decisions via critical thinking.

Nguyen will also explore means to design AI systems that users can better understand, learn thought processes and determine how and why conclusions are made. Then they can debug, assess them and improve them in human-in-the-loop applications.