cmurobotics
Jeff Schneider
Research Professor
Robotics Institute,
Carnegie Mellon University
February 15, 2019
Self Driving Cars and AI: Transforming our cities and our lives
Abstract: Artificial intelligence and machine learning are critical to reaching full autonomy in self driving cars. I will present two autonomy systems along with the use of machine learning in each of them. I will summarize recent progress in commercializing these systems and make some observations about the potential impact of these systems in our daily life. Some of the biggest remaining challenges include efficiently solving the long tail of unusual events on the road, scaling up from demos to commercially viable systems, and verifying the safety of these AI-based systems. I will finish with thoughts on addressing those issues.
Brief Bio: Dr. Jeff Schneider is a research professor in the Carnegie Mellon University School of Computer Science where his recent research is on machine learning for autonomous systems. He received his PhD in Computer Science from the University of Rochester. He has over 20 years experience developing, publishing, and applying machine learning algorithms in government, science, and industry. He has over one hundred publications and has given numerous invited talks and tutorials on the subject. Jeff is also an entrepreneur. He was a founding member of Uber’s Advanced Technologies Group and spent three years helping to build their self driving car program. Before that, he developed a machine learning based CNS drug discovery system and commercialized it during two years as Psychogenics’ Chief Informatics Officer. Earlier in his career he was the co-founder and CEO of Schenley Park Research, a company dedicated to bringing machine learning to industry. Through his research, commercial, and consulting efforts, he has worked with dozens of companies and government agencies around the world.
Source
Suggestion for "the next thing" — applying robotics to making perennial polyculture scalable.
what happens if there is a power failure and all the traffic lights go out….what strategy is used.
Maybe a multi agent game where the pedestrians and other cars try within limits to interfere with the target car. This should create a lot of tail end events, look at Dota 2 driving a multiplayer game.
What's wrong with the technology working WITH the driver, to warn him when a dangerous situation arises or when his attention is waning, he is falling asleep, etc.? Why are self-driving cars the ONLY solution to the safety issue? I for one will never feel safe in such a vehicle, considering all the many things that could go wrong. But I WOULD appreciate the assistance of some of this new technology while I'm driving.
The real problem with self-driving cars, aside from the very serious safety issues, is the fact that no one really needs or wants them. Aside from the convenience factor, the only other issue is safety, but there is no way to determine whether AV are going to be safer until millions of them are already on the road — far too late to make a difference, sorry.
great talk. the conclusion seems to be that holy grail will likely go through end-to- end reinforcement and imitation learning?
Keren video ny y 😍 subscribe Chanel saya y Bagus utk algorithm Utube
Cmon, gimme some time stamps!
This opens a new pandora box for hacking
He has long left uber