Pascal Hitzler builds a more accountable AI
Many people use AI for restaurant recommendations, studying for their next quiz or maybe they don’t even know they’re using it. But how does it actually work?
For Pascal Hitzler, university distinguished professor of computer science and Lloyd T. Smith Creativity in Engineering chair, his research focuses on making AI trustworthy while also giving the public a deeper understanding of it.
“Currently, a lot of development in AI comes out of machine learning,” Hitzler said. “The innate problem is that its performance can only be assessed statistically.”
To understand models, Hitzler’s lab looks at AI networks the same way that neuroscientists study the brain. Understanding the artificial neurons tells them about the trustworthiness of a given model.
“By understanding the internal mechanisms, we get additional evidence that helps us make a decision whether we should trust the system or not,” he said.
Hitzler is creating safer, more trustworthy AI by giving models a way to verify their own internal logic with hard data. He applies his research to autonomous crop sprayers that correct errors before an incident occurs. Also, he is a founding member of the Kansas Legislative Artificial Intelligence Task Force, where he gives his input on a variety of implications of AI in Kansas.
As a land-grant institution, K-State is always searching for ways to make a positive impact on Kansas communities. Hitzler achieves this by using his AI expertise in the government setting and as the co-director of the Institute for Digital Agriculture and Advanced Analytics (ID3A). ID3A is K-State’s interdisciplinary institute that aims to integrate digital technologies to solve agricultural challenges. His work through this organizations has helped Kansas farmers take the next step toward making quicker, more informed decisions on the farm.
By Will Hudelson
Curated from Mapping the future: K-State computer science positions Kansas for its AI era | Kansas State University