PUBLISHED IN ,

Artificial Intelligence Research at Duke

From healthcare to criminal justice to environmental sustainability, Duke researchers and students use the tools of artificial intelligence to assist with various important societal problems

Graphic of human brain with computer parts.
Duke researchers and students use the tools of artificial intelligence to assist with various important societal problems from healthcare to criminal justice to environmental sustainability
By Duke Today Staff

Artificial Intelligence research at Duke covers everything from health to enhancing photos to machine learning. See what some Duke researchers are doing in the field.

AI for Everyone Was An Academic Gateway

This slide from an AI from Everyone session shows an example of how artificial intelligence can be used to estimate human poses.

During Winter Breakaway, David Carlson, an assistant professor in Civil and Environmental Engineering taught “AI for Everyone,” which included an introduction to the math and computations underlying machine learning and artificial intelligence.

Duke Students Taught A Computer to Detect Covid-19 In Lung Scans

Duke students Anmol Warman ’22 and Pranav Warman ’20 have trained a computer to spot the telltale signs of COVID-19 in lung scans and rule out other infections that look similar to the human eye.

Duke student brothers worked together in a home office in Tampa to develop a machine learning system to help clinicians spot the telltale ‘ground glass opacities’ in the lung scans of potential Covid patients.

AI And The Democratization Of Healthcare

Brain scans.

Because no two brains are alike, machine learning is being used to help neurosurgeons home in on the precise area where the electrode should go to treat Parkinson’s disease with deep brain stimulation.

Accurate Neural Network Computer Vision Without The 'Black Box'

New research offers clues to what goes on inside the minds of machines as they learn to see. A method developed by Cynthia Rudin's lab reveals how much a neural network calls to mind different concepts as an image travels through the network’s layers.

Researchers could better troubleshoot neural networks if they’d only show their work! Well, now a Duke team has developed a mechanism they call ‘concept whitening’ that helps them peer inside the black box while the algorithm is teaching itself new concepts.

This AI Birdwatcher Lets You 'See' Through The Eyes Of A Machine

A Duke team trained a computer to identify up to 200 species of birds from just a photo. Given a photo of a mystery bird (top), the A.I. spits out heat maps showing which parts of the image are most similar to typical species features it has seen before.

A deep neural network developed at Duke learned to identify 200 kinds of birds by studying nearly 12,000 images while the researchers watched ‘heat maps’ that showed what features of the photos it was using to learn one warbler from another.

Artificial Intelligence Makes Blurry Faces Look More Than 60 Times Sharper

Digital faces from clear to blurry.

Duke AI researchers used a machine learning tool called a ‘generative adversarial network’ to enable an algorithm to make educated guesses about what the face in a blurry photograph looks like in focus. But, be advised, it isn’t a tool to help facial recognition software. 

Duke Launches Graduate Certificate in AI for Product Innovation

Graphic of a brain with binary code in the background.

In May 2020, the Pratt School of Engineering introduced a 15-month online learning program for working professionals that taps into our world-class machine learning research and teaching.

Duke, As Seen Through Fauvism and Machine Learning

Electrical and computer engineering alumna Shixing Cao is experimenting with combining her training and her love of art. She's applying the painting styles of the masters to photos of Duke scenery, using the machine-learning approach developed by Leon Gat

Electrical and computer engineering alumna Shixing Cao has been applying the painting styles of the masters to photos of Duke scenery, using a machine-learning approach.