language modeling is not unsupervised learning.
-
ok, i buy that. lm by training to predict a hidden word is not unsupervised imo.
Matrix factorization: unsupervised. Using a missing word as a label to be predicted: supervised. Same task, very similar results, but meaningfully different ways to conceptualize the solution.
-
-
Isn't "predicting missing word..." unsupervised in the sense that no human annotation is required to learn? Or is self-supervised a better word?
-
Seeing your output labels during training ("supervision"), vs. How expensive it is to get these labels ("annotation"), are two orthogonal matters. In "next word prediction" you have the label, but it is cheap to obtain.
- Show replies
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.