Reinforcement learning is a paradigm that will eventually be superseded. We just haven't figured out what the new, more generally useful, paradigm is yet. When we do, there's going to be a revolution. It will be very interesting.
Show this thread
-
@fchollet please say more about your last two items: Why is separation between the agent and the environment a problem? What is the distinction between behavior programs and behavior program generation that we need?
1) An agent is not a static set of possible actions and reward variables. Its affordances change over time: the environment becomes part of the agent. A clever agent will actively seek to gradually *own (absorb) more of the environment* over time.
-
-
2) this emphasizes task-specific skill, which is useless (no one needs a program that plays Pacman) as opposed to algos capable of acquiring arbitrary skills. Which is why deep RL still achieves close to 0 generalization after all these years: generalization was never encouraged.
-
You must have a more general (!) notion of generalization in mind. AlphaGo certainly can generalize to Go positions that it has never seen during training. That is the form of generalization we have been trying to achieve for many years.
- Show replies
New conversation -
-
-
I think there are interesting developments in the future wrt the "boundary" between agent and environment. Embodiment, offloading cognition onto the environment, tool use---none of these are fully captured by the agent/env diagram.But it doesn't preclude them either.
-
This is a great illustration -- tool use, offboard cognition, etc. are great examples of how agency is artifactual. We're transforming parts of our environment into parts of the circuit of agency. Need a fully embedded view of agency.
- Show replies
New conversation -
-
-
Do you think there are any frameworks of agency which properly embed the agent within an environment & allow this change in capacity/boundary? For me this is an issue with how Friston's FEP work is framed; I think the Causal Entropic Forces generalizes over this boundary, helps.
-
This Tweet is unavailable.
- Show replies
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.