ActInf GuestStream 083.1 ~ "Embodied intelligence", Joshua Bongard
Вставка
- Опубліковано 5 вер 2024
- "Embodied intelligence"
Joshua Bongard
It is still laughably easy to foil AI with adversarial attacks. Why? Because such systems lack embodiment. But dropping deep learners into robots and calling the result Embodied AI misses the mark: Embodiment is about more than just having a body; it is about change. Consider how much the world changes from the perspective of a human as she grows from one cell into 10^13 of them. Grappling with such massive internal change makes grappling with external change, like learning to read or drive, easy by comparison. Thus, to realize safe AI, we must similarly create autonomous technologies whose internal physical changes pretrain them to handle external change, like new tasks or adversarial attacks. I’m excited to discuss a future filled with soft and biological robots that are capable of this “morphological pre-training”, and how it will make them useful, smart, and safe.
www.uvm.edu/ce...
scholar.google...
Xenobots, Biobots, and more!
Active Inference Institute information:
Website: activeinferenc...
Twitter: / inferenceactive
Discord: / discord
UA-cam: / activeinference
Active Inference Livestreams: coda.io/@activ...
12:11 Concerning situation, absolutely, I'm not sure 30:43 on this point, Mike Levin talks about a very basic system that surprised us, and has 6 forms of the operating system, our basal operation system, to move in the biological world, do these hi-brots come with this as part of the biological operating system, just a question, thank you both very much for sharing your time and work Doc Josh, and the active inference team for this presentation, and the moderator Daniel, peace
38:10 Great question/comment