I want the neural network to be able to remember, but a perceptron can only remember something during training, but I want the neural network to adapt to new conditions without retraining, for example, if I said that my name is Nikolai, it remembered it, or if green mushrooms were useful in the game before, but then became poisonous, it stopped eating them, and then, if they returned to normal, it started eating them again. I thought about LSTM, but this is a recurrent neural network, and people seem to be abandoning them in favor of transformers?
6
-
$\begingroup$ Very interesting question (+1). A few ideas come to mind….Have you thought about a transformer with “external memory”, or Neural episodic memory ? $\endgroup$– Robert LongCommented Nov 29 at 18:47
-
$\begingroup$ @RobertLong, -? $\endgroup$– Nikolai VorobievCommented Nov 30 at 21:16
-
$\begingroup$ Hi Nikolai 🙂 Does that mean “no” ?🤔 $\endgroup$– Robert LongCommented Nov 30 at 21:58
-
$\begingroup$ @RobertLong, site just opened strangely and instead of an answer there was only your name and a dash "– Robert Long ", I do not know why $\endgroup$– Nikolai VorobievCommented Nov 30 at 21:59
-
$\begingroup$ Nikolai, it must be the StackExchange gremlins 🤷♂️🙂Can you read my comment now ?🤔 $\endgroup$– Robert LongCommented Nov 30 at 22:38
|
Show 1 more comment