Major news re #deeplearning & its limits: Yoshua Bengio’s lab confirms a key conclusion of Marcus 2001 and 2018: deep learning isn’t efficient enough with data to cope with the compositional nature of language. https://export.arxiv.org/abs/1810.08272
-
If I remember correctly, compositionality is thought to arise from the limited source of symbols (phonemes) and need for lossy compression that is insensitive to corruption by noise.
hunh? who defines compositionality that way? the usual example is at the sentence level, meaning constructed from parts
-
-
These two (or three ;)) views are not mutually exclusive. Plus I don’t think .
@roydanroy tried to define compositionality; merely gave some characteristics and why it’s needed. -
Indeed as was asked why it arises. I often think this way of characterizing phenomena is more fundamental.
End of conversation
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.