Videos

This Curious AI Beats Many Games…and Gets Addicted to the TV



Two Minute Papers

Pick up cool perks on our Patreon page:
https://www.patreon.com/TwoMinutePapers

Crypto and PayPal links are available below. Thank you very much for your generous support!
› PayPal: https://www.paypal.me/TwoMinutePapers
› Bitcoin: 13hhmJnLEzwXgmgJN7RB6bWVdT7WkrFAHh
› Ethereum: 0x002BB163DfE89B7aD0712846F1a1E53ba6136b5A
› LTC: LM8AUh5bGcNgzq6HaV1jeaJrFvmKxxgiXg

The paper “Large-Scale Study of Curiosity-Driven Learning” is available here:
Paper – https://pathak22.github.io/large-scale-curiosity/
Blog post – https://blog.openai.com/reinforcement-learning-with-prediction-based-rewards/

We would like to thank our generous Patreon supporters who make Two Minute Papers possible:
313V, Andrew Melnychuk, Angelos Evripiotis, Anthony Vdovitchenko, Brian Gilman, Christian Ahlin, Christoph Jadanowski, Dennis Abts, Emmanuel, Eric Haddad, Eric Martel, Evan Breznyik, Geronimo Moralez, John De Witt, Kjartan Olason, Lorin Atzberger, Marten Rauschenberg, Maurits van Mastrigt, Michael Albrecht, Michael Jensen, Morten Punnerud Engelstad, Nader Shakerin, Owen Skarpness, Raul Araújo da Silva, Rob Rowe, Robin Graham, Ryan Monsurate, Shawn Azman, Steef, Steve Messina, Sunil Kim, Thomas Krcmar, Torsten Reil, Zach Boldyga.
https://www.patreon.com/TwoMinutePapers

Thumbnail background image credit: https://pixabay.com/photo-3774381/
Splash screen/thumbnail design: Felícia Fehér – http://felicia.hu

Károly Zsolnai-Fehér’s links:
Facebook: https://www.facebook.com/TwoMinutePapers/
Twitter: https://twitter.com/karoly_zsolnai
Web: https://cg.tuwien.ac.at/~zsolnai/

Source

Similar Posts

38 thoughts on “This Curious AI Beats Many Games…and Gets Addicted to the TV
  1. It is not curious. it is simply programmed to learn new content. It's programmed to know that it can't learn new content from old content. It's only fake choice is to move to the next unknown content.This is simply what humans do to escape boredom. humans are able to choose the content they interact with. a program has no choices, no free will, no will at all. People work hard to program AI to do things like a human would. Once it completes what people have made and programmed it to do. they then humanize the AI by altering their own perception of the AI.

  2. A Terminator was sent from the future to kill me. Luckily, I just finished watching this episode. So I flicked on my TV and now he’s just standing there, staring in awe into the depths of the flat screen. Thank you, Two Minute Papers, you saved me and the rest of the world!

  3. If the AI is exploring the environment based on visual input, and it found a TV screen, wouldn't it spend a lot of its time trying to learn what a TV was. I.e. making sense of this image that has flashing images on it, and being continually unable to relate the visual information it was seeing to its environment. How does the AI respond to such an alien thing? What if what we're witnessing isn't 'addiction' to TV but petrifaction?

    Imagine exposing a 13th century peasant to a plasma television – they'd probably be equally shocked at this weird contraption in front of them.

  4. cool you gave it curiosity
    now you should give it boredom so it doesn't hang in front of the television looking at random crap all day

  5. I just hope researchers solve the novel stimuli addiction problem before they upload a more powerful successor version of this type of neural network into a strong, bipedal robotic body so the robot does not develop a morbid curiosity addiction.

  6. These things are sure interesting, but scientists should also think about their responsibilities. What a time to be alive? Just wait until the applications are used to suppress us even more.

  7. Curiosity as the only motivator is obviously dangerous in the real world, scine it can lead to immoral experiments for fun.

  8. Lessening the reward if the ai remains inactive for too long might work against the tv addiction. I think that’s how humans deal with this issue.

  9. make the ai curious of the result of it's actions – make it "want" to do something and aim for a result – how it can change the scenery…
    so that looking at the TV will eventually bore the AI.

  10. It's interesting to look at how natural evolution sorted out this problem (mostly) in humans. Specifically, the contrasting desires for novel experiences (curiosity) and the desire to find patterns/similarities in things. If all we're getting is novel experiences with no linking factor, we get bored, but if we're not getting enough novel experiences, we similarly get bored. I know very little about programming, but I expect that the hard part about trying to implement this in an AI would be finding the right balance.

  11. I bet if a desire for agency was added to balance the curiosity, it would partially solve the tv addiction. I know that that's what makes me prefer games to tv.

Comments are closed.

WP2Social Auto Publish Powered By : XYZScripts.com