AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
Is the Challenge Worthwhile?
In the case of poet are you predicting whether it's too hard or too easy or are you letting the agent try and kind of measure progress. Do things like early stopping or does the agent actually have to do it? That's when you know if it's toohard or too easy yeah so in that work we took out pretty easy approach which isWe created the environment we tested at the current agents that we had so far but then after a while if you haven't been learning on it I think we can't get out okay.