Anca Dragan is a professor at Berkeley, working on human-robot interaction -- algorithms that look beyond the robot's function in isolation, and generate robot behavior that accounts for interaction and coordination with human beings. Support this podcast by supporting the sponsors and using the special code: - Download Cash App on the App Store or Google Play & use code "LexPodcast" EPISODE LINKS: Anca's Twitter: https://twitter.com/ancadianadragan Anca's Website: https://people.eecs.berkeley.edu/~anca/ This conversation is part of the Artificial Intelligence podcast. If you would like to get more information about this podcast go to https://lexfridman.com/ai or connect with @lexfridman on Twitter, LinkedIn, Facebook, Medium, or YouTube where you can watch the video versions of these conversations. If you enjoy the podcast, please rate it 5 stars on Apple Podcasts, follow on Spotify, or support it on Patreon. Here's the outline of the episode. On some podcast players you should be able to click the timestamp to jump to that time. OUTLINE: 00:00 - Introduction 02:26 - Interest in robotics 05:32 - Computer science 07:32 - Favorite robot 13:25 - How difficult is human-robot interaction? 32:01 - HRI application domains 34:24 - Optimizing the beliefs of humans 45:59 - Difficulty of driving when humans are involved 1:05:02 - Semi-autonomous driving 1:10:39 - How do we specify good rewards? 1:17:30 - Leaked information from human behavior 1:21:59 - Three laws of robotics 1:26:31 - Book recommendation 1:29:02 - If a doctor gave you 5 years to live... 1:32:48 - Small act of kindness 1:34:31 - Meaning of life
Anka Jurgon, a professor at Berkeley, discusses her work on human-robot interaction and creating algorithms that account for coordination with humans. This conversation also includes a brief interruption for sponsor ads.
The guest talks about the transformative experience of meeting a Google self-driving car for the first time, while also discussing their early inspirations in math and computer science.
The speaker expresses their affection for robots, particularly Wally from the animated film, and discusses the difficulties in expressing goals and intentions for robots in real life.
Understanding human behavior is key to anticipating their future actions. Incorporating humans into the state model of robots changes the problem as the robot is no longer the only entity acting in a space.
The challenges of human-robot collaboration arise from the fact that while both agents must act, the robot knows less than the human about what they want. The difficulty lies in making sense of human behavior enough to be able to optimally assist them with autonomy, despite the development of robot capabilities.
Instead of thinking people are too irrational, we can consider that they're rational under different assumptions about the world. People may be operating under a simplified physics model when compared to robots.
Erin Smith discusses the importance of making robots more expressive and the impact it can have on how people react to them.
This podcast discusses how understanding influence and control in interactions between humans and robotics is essential. It highlights how there is often ambiguity in the right of way and how relying on reading cues helps to step outside of this mode of anticipation.
The deployment of autonomous vehicles is still hindered by various challenges including perception problems, planning problems, and control problems in edge cases, despite lighter materials being used. However, sticking to old methods hinder innovation and exploration.
Modeling in AI development is crucial for creating policies that are efficient and safe. While it's possible to learn from scratch, human expertise can be injected to make the system work better.
The human factor in human-robot interactions is key, and while it may be challenging to specify rewards for interaction, having a "dumb" robot that relies on humans to help it out could lead to unintended consequences that must be carefully managed.
The interaction between human and robot, as well as the challenge of programming robots to do what we want, is discussed in this episode. The focus is on reward learning as a means of training robots to perform tasks to our specific requirements.
The robot can collect information from external torques and can even plan its motions by disambiguating through E-stop signals to improve its ability to determine what the user wants.
The speaker discusses how their appreciation for the beauty in life stems from the fact that it is temporary, and how this relates to their work in creating robots that can autonomously interact with humans. They also reflect on the idea of eternal life and the lessons from the show "The Good Place."
The speaker discusses the concept of the vastness of the universe and how it relates to the human experience, referencing a video that visually represents the scale of the universe. They also touch on the multiverse theory.
In this podcast, physicist Dr. Akram Khan-Cheema discusses the multiverse, consciousness and the meaning of life in the context of the fundamental nature of the universe.