M.S. in Computer Science — Agile

A Katz School researcher and several colleagues are developing an AI-powered wearable device that can monitor illicit drug use in individuals with substance use disorder.

In a research paper, “Precision Polysubstance Use Episode Detection in Wearable Biosensor Data Streams,” Dr. Honggang Wang, chair of the Katz School’s Computer Science and Engineering Department, describes a method called RP-STREAM, which is an algorithm that can identify when an individual is using drugs by analyzing data collected from a wearable device. 

To develop the model, the researchers collected data from 15 people who had used cocaine and were fitted with a wearable biosensor called Affectiva Q, which can measure various physiological changes such as electrodermal activity, physical activity and body temperature. To figure out when these people used cocaine, the researchers evaluated patient notes, urine tests and data from the device. 

“When it comes to detecting drug use, some studies have been successful in finding out when people use drugs, but they usually look at big chunks of time,” said Dr. Wang. “This new method is different because it can figure out when someone uses drugs in smaller timeframes. It’s like looking at a movie frame by frame instead of the whole movie at once.” 

Dr. Wang presented the team’s findings at the Institute of Electrical and Electronics Engineers (IEEE) International Conference on Connected Health: Applications, Systems and Engineering Technologies, with colleagues Joshua Rumbut and Hua Fang of the University of Massachusetts Dartmouth and University of Massachusetts Chan Medical School, and Edward Boyer of Ohio State University and Harvard Medical School.

Read the full story

Katz School researchers received the Emerging Research Award at the Future Technologies Conference for their work on a machine learning algorithm that could reduce the number of traffic accidents involving self-driving cars.

In their paper, “LaksNet: An End-to-End Deep Learning Model for Self-Driving Cars in Udacity Simulator,” Youshan Zhang, assistant professor of artificial intelligence and computer science, and Lakshmikar Polamreddy, a master’s candidate in artificial intelligence, describe their convolutional neural network (CNN) model for self-driving cars, which aimed to address the limitations of previous work in this area.

The LaksNet model uses images and steering angles collected from a Udacity simulator—an open-source simulator for training and testing self-driving, deep-learning algorithms. The Udacity simulator includes a virtual representation of a car and its surroundings, allowing users to implement and test their algorithms for tasks, like perception, decision-making and control, providing a safe environment for learning and experimenting with self-driving technologies.

“Our approach involved building and training end-to-end, machine-learning models using extensive sets of data, typically in the form of images collected from cameras,” said Zhang. “These models were trained to drive vehicles in a way that minimized accidents.”

Read the full story

Dr. Youshan Zhang, assistant professor of computer science and artificial intelligence, and Jialu Li of Cornell University have created a novel noise removal method that could benefit the hearing impaired and improve the listening experience for audiophiles everywhere.

In their paper, “BirdSoundsDenoising: Deep Visual Audio Denoising for Bird Sound,” the researchers described how they created a deep visual audio denoising (DVAD) model using a dataset of 15,300 bird sounds—varying in length from 1 second to 15 seconds—that strips out the background noise, in this case natural sounds like wind and rain, to produce clean bird sounds.

The researchers presented their model in January at the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) conference in Hawaii. Dr. Zhang said the model is robust enough to apply to human speech, especially to background noise that is particularly damaging to speech intelligibility for people with difficulty hearing.

“Our DVAD model can first denoise the background noise and then increase the volume of the low voice,” he said.

In a novel twist, the researchers turned the audio of the bird sounds into a series of images; used a photo editing tool that eliminates the original background of an image without compromising its integrity; created a segmentation model to edit out the noisy parts of the image; and then applied an algorithm to produce the “denoised,” or clean bird sounds.

“To the best of our knowledge, we are the first to transfer audio denoising into an image segmentation problem,” said Dr. Zhang. “By removing the noise area in the audio image, we can realize the purpose of audio denoising.”

Read the full story