Reinforcement Learning
Discussions center on the use, effectiveness, and application of reinforcement learning (RL) to the problem in the HN post, including comparisons to supervised learning, DeepMind approaches, and RL techniques like RLHF.
Activity Over Time
Top Contributors
Keywords
Sample Comments
reinforcement learning goes along way too.
Advertising: if you're interested in RL, subscribe to https://www.reddit.com/r/reinforcementlearning/ !
seems like they gunning for deepmind rl based ideas.
Maybe it's some reinforcement learning.
It looks like the wrong AI for the problem. RL should be more successful.
Has anyone attempted yet to run a reinforcement learning strategy on it?
Why use RL for this instead of plain old supervised learning?
They're clearly taking a reinforcement learning approach. Just a few thousand more returns and you'll be good to go!
More likely reinforcement learning, I think.
Isn't RL the algorithm we want basically?