But not all organisms have brains(nervous system) capable of generating a model of "self"(depersonalization) but continue to protect/care themselves from harm. So what sort of caring or advantage does a "self" model of brain provide us with?
-
-
Replying to @dwindlinghandle
It is the other way around. The more detailed the regulation you want to perform, the more detailed you model must be. Simpler organisms don't have a self model because they cannot afford a sufficiently complex modeling system (large nervous system plus long training period).
2 replies 0 retweets 0 likes -
Replying to @Plinz
Most of the processing in the brain happens at the background. So is the self model involved in some sort of processing? Is the self model just like a visual display unit of a computer displaying data (experience)?
1 reply 0 retweets 0 likes -
Replying to @dwindlinghandle
Most of the self model is unconscious, too. Its content is the set of structures and latent variables that the system uses to predict and regulate its own activity.
1 reply 0 retweets 0 likes -
Replying to @Plinz
Was wondering what "experience" actually does? For eg: when a dog chases a person, the background processing and physiological state changes and the person starts running, the self after some time "is aware" of the state (physiological and thoughts).
1 reply 0 retweets 0 likes -
Replying to @dwindlinghandle @Plinz
If we strip off this self model, dont u think the response of the person would be same?
1 reply 0 retweets 0 likes -
Replying to @dwindlinghandle
Reactive behavior may be similar if you strip out the self model, but not planned behavior. Also, the self model evaluates low level behavior and changes the reactive behavior via training.
1 reply 0 retweets 0 likes -
Replying to @Plinz
Whats your view on the AlphaGo AI Zero programme which "invented" new moves by playing million times over its instances? Here we dont see any self model involved besides training its model and rules of the game.
1 reply 0 retweets 0 likes -
Replying to @dwindlinghandle
AlphaGo’s problems were only related to Go, not to itself.
1 reply 0 retweets 0 likes -
Replying to @Plinz
But it in a trivial and abstract sense, that would't it need an additional model which would 1) evaluate performance of the Go game model. 2) provide "suggestions" to the game model.3) consider the output of the game model to be of itself and doing above (1 and 2).
1 reply 0 retweets 0 likes
There is not much situatedness with respect to the game. The game playing algorithm itself does not have an interesting trajectory, beyond getting better.
-
Show additional replies, including those that may contain offensive content
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.