CSE Teaching Professor Christine Alvardo along with colleagues at three other Southern California universities received a $1.3 million grant from the National Science Foundation (NSF) to investigate how gender, ethnicity, and other factors impact those who want to study computer science. The researchers will study application, admission, and retention data to better understand how demographics and pre-college preparation influence student success, and specifically what factors create or prevent access to a computer science education.
For the first time, computer scientists and researchers with CSE Professor Julian McAuley’s lab demonstrated that systems designed to defeat deepfakes can be deceived. Detectors can be defeated by inserting inputs called adversarial examples into every video frame, which cause artificial intelligence systems to make a mistake. With the spread of fake videos on social media platforms, training detectors against adaptive adversaries is imperative.
Kamalika Chaudhuri, a CSE associate professor, is part of a multi-university team that has won a prestigious U.S. Department of Defense Multidisciplinary University Research Initiative (MURI) Award to develop rigorous methods for human-machine collaborations against adversaries.
Arun Kumar, a CSE professor and Halicioglu Data Science Institute assistant professor, is working to bridge the gap between computing systems and machine learning. Big data is an incredible asset in any industry, but relies on sophisticated data sorting techniques to isolate useful information and to harness its predictive capabilities. These machine learning techniques are often slow and both labor and resource intensive, and it is this disconnect that Kumar and his colleagues are seeking to address through focusing on the deployment issues of scalability and usability.
According to research led by Melissa Gymrek, a professor of Medicine and Computer Science and Engineering, mutations that occur in certain DNA regions may play a significant role in autism spectrum disorders. Called tandem repeats, these mutations are often overlooked by researchers because “they’re generally non-coding– they do not make proteins; their function is unclear; and they can be difficult to analyze,” says Gymrek. The study published in Nature found that on average there were more mutations in ASD children. Using a novel algorithmic tool developed by UC San Diego bioengineering undergraduate student Bonnie Huang, researchers found that tandem repeat mutations predicted to be the most evolutionarily deleterious were found at higher rates in ASD children.
copyright 2020 – Computer Science & Engineering – University of California San Diego
