Have you looked into reinforcement learning? Recall the systems developed to beat humans in games such as Go and Chess? My guess is that eventually (if not already in the works) these language models will incorporate reinforcement learning methods to advance another stage. Question is, what goals they will have the system seek?
Have you looked into reinforcement learning? Recall the systems developed to beat humans in games such as Go and Chess? My guess is that eventually (if not already in the works) these language models will incorporate reinforcement learning methods to advance another stage. Question is, what goals they will have the system seek?