Dacheng Tao

In the world, it is predicted that one in ten automobiles will be self-driving by 2030. Designing computer vision systems that can recognize barriers to prevent collisions at least as successfully as a human driver is essential to putting these autonomous cars on the road.

The "brains" of self-driving vehicles are neural networks, which are collections of AI algorithms inspired by neurological processes that occur in the human cerebral cortex. Neural networks are created for computer vision applications by Dacheng Tao, an Australian computer scientist at the University of Sydney. Moreover, he is developing algorithms and models that can analyze films taken by moving cameras, such as those seen in self-driving cars.

The environment can be modeled extremely effectively using neural networks, according to Tao, head of the UBTECH Sydney Artificial Intelligence Institute, a collaboration between the University of Sydney and the world's largest robotics business UBTECH.

For a five-year effort utilizing deep learning methods to enhance moving-camera computer vision in autonomous machines and vehicles, Tao received an Australian Laureate Fellowship in 2017. Deep learning, a subtype of machine learning, employs neural networks to create systems that can "learn" as they analyze their own data.

Tao's study has produced more than 40 journal articles and conference papers since it began in 2018. He is one of Australia's most highly cited computer scientists and one of the most prolific researchers in AI research output from 2015 to 2019, according to the Dimensions database. According to Google Scholar's database, Tao's publications have received over 42,500 citations since 2015. He received the Australian Museum's Eureka Award for Excellence in Data Science in November 2020.

A neural network was trained by Tao and his colleagues in 2019 to create 3D worlds using a motion-blurred picture, similar to one that might be acquired by a driving automobile. The researchers were able to retrieve what they refer to as "the 3D world concealed behind the blurs" by paying attention to details like motion, blurring effect, and depth at which it was recorded. The research could improve how self-driving cars interpret their environment.