News for hackers. Stuff that matters.
Re:Why Deep Learning Cannot Be Applied to Natural Languages ... by westoncb on Monday January 02, 2017 @03:08PM (#13304008)
Searle's Chinese Room is addressing something different from intelligence; it's concerned with the what happens /within/ the intelligent mind, whereas inputs and outputs are the only things that matter here. To be more specific, in the question of whether deep learning can be used to generate and/or comprehend natural language, we are not concerned with whether the algorithm is conscious of what it's doing as longs as the results are good.
Re:Why Deep Learning Cannot Be Applied to Natural Languages ... by visarga on Monday January 02, 2017 @06:20PM (#13305332)
In order to be conscious it has to be more than a reactive or feedforward system. It has to loop back on itself, like RNNs, and hold internal state.