Seminar to offer better understanding on AI technology and real-world uses

Matt Swayne
February 27, 2019

UNIVERSITY PARK, Pa. — Gaining a better understanding on how artificial intelligence — AI — actually works and seeing how it is currently being used in real-world situations may help bridge the gap between the fears and hype that surround the technology, according to a Penn State researcher who will present a seminar on AI in March.

Lee Giles

C. Lee Giles

IMAGE: Penn State

C. Lee Giles, the David Reese Professor of Information Sciences and Technology and an associate of the Institute for CyberScience, will offer a talk about two of the most popular forms of AI methodologies during March’s CyberScience Seminar. The seminar will be held from 1:30 to 3 p.m. on Thursday, March 14, in Room 233B of the HUB-Robeson Center at University Park.

“AI seems to divide us into two camps. While some believe it will transform the world, others fear it will take over it. What we really need is a better understanding of how AI works,” said Giles. “Understanding how AI works on actual problems allows us to have a better idea of both AI’s power as well as its limitations.”

Giles said the seminar will cover enterprise AI; machine learning, which is already used in numerous applications and technologies; and deep learning, a machine learning model loosely based on the brain’s own neural networks, that is growing in use.

“Deep learning is a fairly new type of machine learning model that is becoming more popular because of how well it performs,” said Giles. “Within the last ten years, deep learning has begun to significantly outperform other machine learning methods. First, deep learning began to outperform machine learning in applications, such as helping computers recognize images and pictures, and, more recently, it’s been used to create technology that can help computers translate different languages.”

He added that deep learning teams are continuing to make progress on solving problems that require large data, and news of deep learning teams winning problem-solving competitions is making headlines regularly. Giles’ introduction to deep learning will also cover how it works and explain some of the technology’s limitations.

“This may help researchers currently using machine learning for their own problems, as well as give researchers ideas about how they can successfully use the technology in future work,” said Giles.

Giles said his talk will briefly review stateful and stateless deep learning — also known as recurrent or dynamic networks and feedforward networks — classes of deep learning architecture and explain how developers are combining both classes to create hybrid systems.

The CyberScience Seminars are a series of sessions held during each spring and fall semesters to show how researchers at Penn State are using data science to explore our world and solve some of the biggest social challenges of our time.

Registration is required. For more information, visit the ICS event page.

(Media Contacts)

Last Updated March 28, 2019