Not willing to accept this, now they say that the artificial neural network dreams. Name: Google Deep Dream.
The problem is that the Google Deep Dream images can be compared with dreams by us humans. Or in the general case of an automatic classifier of big data there is no term of comparison. How can we, the pragmatic scientists, know that the output obtained from data (by training a neural network on other data) is full of dreamy dog eyes or not?
If we can’t trust the meaning obtained from big data, by pragmatic means, then we might as well renounce at analytic philosophy and turn towards continental (so called) philosophy.
That is seen as not serious, of course, which means that the ANN dreams, whatever that means. With this stance we transform a kick in the ass of our most fundamental beliefs into a perceived progress.