We’re all used to robots that fail when their environment changes unpredictably. Our robotic system is adaptable enough to handle unexpected situations not seen during training, such as being prodded by a stuffed giraffe:pic.twitter.com/wBoh1nt9Kv
-
-
Prikaži ovu nitHvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
-
-
-
One person on the project used to have consistently better results on the robot. For a while, we couldn't figure out why. It turned out that his laptop was faster, and it incurred less latency on the robot, which in turn gave better results.
-
"This AGI doesn't work at all!" "Try closing your browser tabs" "Oh hey, it works!"
- Još 1 odgovor
Novi razgovor -
-
-
Sometimes we were unaware that our robot is partially broken because the neural network could compensate for it. The model worked just fine with broken fingers or defected sensors.
- Još 5 drugih odgovora
Novi razgovor -
-
-
Pah! That's a piece of cake. Can it be trained to do a Rubik cube the proper way - by peeling the stickers off and resticking them to complete the puzzle? That's the real test.
-
No, the proper way is to disassemble the cube (remove all the corners and edges from the core with the middle of all the faces), and then reassemble it.
- Još 1 odgovor
Novi razgovor -
-
-
That's further evidence that deeply intelligent robots and virtual assistants will be created. I am proud that my organization is forward-thinking about safety and policy. We have to embed human values in AI systems, and we have to figure out how to distribute the benefits of AI.
- Još 1 odgovor
Novi razgovor -
-
- Kraj razgovora
Novi razgovor
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.
