3 ms·
Is DeepMind’s new reinforcement learning system a step toward general AI?
- hatmatrix 5y agoClassic example for Betteridge's Law of Headlines
- dane-pgp 5y ago> DeepMind’s reinforcement learning agents might have become the masters of the virtual XLand. But their simulated world doesn’t even have a fraction of the intricacies of the real world. That gap will continue to remain a challenge for a long time. I guess the question is how the learning curve of a world increases as you add more complexity/realism to it. If one of these agents was placed in an environment like Minecraft, I wonder how quickly it would learn that building a pickaxe is beneficial for survival. It wouldn't be fair to expect it to brute force (let alone intuit) what the correct crafting recipe is for one of those, but it could be told when it has the necessary ingredients, and given a reward for each new type of item it holds.