Referring to deep learning as just being good for "perception" really strains credulity. Chess has historically been a prototypical exercise in reasoning, but since variables weren't involved in it's mastery, I guess it's perception now?https://twitter.com/GaryMarcus/status/1068897657530138629 …
-
That is a very liberal definition of variable binding. That would encompass all planning procedures (e.g. s_{t+n} is bound to the state predicted in n steps). If wading into model-based RL is all you wanted, then every DRL researcher is already on board. Mission accomplished?
-
-
DQN is model-free, but much of the work in the field (often by the same people) has involved model-based methods. (Tangent: you could even make a case for DQN's replay buffer being a model http://proceedings.mlr.press/v37/vanseijen15.html …)
-
certainly agreed. whether you can build an adequate model w/out variables is perhaps an open question analogous to what i just wrote about mcts
- 2 more replies
New conversation -
-
-
You are in a node (variable) and there is a number of alternatives for the next move. You switch to a new node (variable changed) by selecting an action. Thus you manipulated a variable. Does deep learning make you that blind?
-
Look, nodes! I guess Q-learning was symbolic manipulation all along. Seriously though, if the variable binding process is as simple as s_1 <- state-after-s_0, then its not terribly useful to use the same terminology as, say, evaluating a context-free grammar.pic.twitter.com/fReehI3DiK
- 1 more reply
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.