The New Approach to Turn the Corner on Safety Critical Situations

2min Snip

00:00

Play full episode

Summary

Transcript

Episode notes

autonomous vehicles can handle many critical road situations, but there are certain scenarios known as corner cases that they struggle with. Most people will drive for a full year and never see anything close to an accident in their own driving. Henry Liu from the University of Michigan developed what we call dance learning approach. What it is is we train an AI agent and so that AI agent can see these safety critical situations very, very often.

00:46 A new test to get autonomous vehicles on the road

Truly autonomous vehicles, ones that don't require a driver to be present and are driven by AI, aren't yet safe for public use. Part of the reason for this is it has been difficult to train them to deal with rare dangerous situations. Now researchers are unveiling a new approach to present lots of these infrequent events to the AI very rapidly, speeding up the training and testing process.

Research Article: Feng et al.

News and Views: Hazards help autonomous cars to drive safely

Video: The driving test for driverless cars

08:23 Research Highlights

How bird-flu is adapting to mammals, and the effect of negative headlines.

Research Highlight: Bird-flu virus makes itself at home in Canada’s foxes and skunks

Research Highlight: It’s bad! Awful! Negative headlines draw more readers

10:43 Why bat research is taking off

Bats are known to tolerate a lot of viruses that are deadly to humans without much issue. With the ongoing pandemic, this has driven researchers to dive more into the world of bats in the hopes of applying bats' tolerance to humans. Reporter Smriti Mallapaty has been writing about this renewed interest and she joined us to tell us more.

News Feature: Bats live with dozens of nasty viruses — can studying them help stop pandemics?

Hosted on Acast. See acast.com/privacy for more information.