That was a great talk really cool stuff! Your coffee example reminded me of Human Compatible “you can’t make coffee if your dead” and it being goal oriented RL. I think the sub goals of resetting the env is the right direction which is awesome but at the same time make me uneasy.
I'm posting my talk from the deep RL workshop here for anyone who didn't catch it: https://www.youtube.com/watch?v=Ik1nS2E4ar4 … The Case for Real-World Reinforcement Learning Or: Why Robinson Crusoe didn't need to be good at chess in the hopes we can all keep it real when it comes to RL
-
-
-
I sometimes think that good AI research really should make people at least *somewhat* uneasy. Obviously we should be thoughtful and responsible, but meaningful, important technological progress is always at least a little threatening.
- Näytä vastaukset
Uusi keskustelu -
-
-
Nice talk, Sergey. I totally agree with you, unsurprisingly.
-
Thank you Yann! That means a lot to me.
Keskustelun loppu
Uusi keskustelu -
-
-
Woow....Great way to think about our ML models. Its not about optimal control but Good enough....
-
Thanks. That part is also somewhat inspired by this rather interesting writeup by Gerald Loeb ("Optimal isn't good enough"): https://viterbi.usc.edu/pdfs/gloeb/81544.pdf …
- Näytä vastaukset
Uusi keskustelu -
-
-
I need to watch that :) so far I still consider it a question of optimal control, just that the problem is incorrectly defined.
Kiitos. Käytämme tätä aikajanasi parantamiseen. KumoaKumoa
-
-
-
Thanks for in depth thinking about hard university!
Kiitos. Käytämme tätä aikajanasi parantamiseen. KumoaKumoa
-
-
-
Nice! Perhaps RL could benefit from giving more space to "idea with proof-of-concept" and "we tried it in the wild" results. Interesting position paper on the harms of overfitting to quantitative results (for HCI): http://grouplab.cpsc.ucalgary.ca/grouplab/uploads/Publications/Publications/2008-UsabilityHarmful.CHI.pdf …
Kiitos. Käytämme tätä aikajanasi parantamiseen. KumoaKumoa
-
Lataaminen näyttää kestävän hetken.
Twitter saattaa olla ruuhkautunut tai ongelma on muuten hetkellinen. Yritä uudelleen tai käy Twitterin tilasivulla saadaksesi lisätietoja.