Andres,

(and also Linas and anyone else interested...)

I have refreshed my memory on clustering for unsupervised POS
learning... this was the approach I had fiddled with long ago,

http://www.cs.rhul.ac.uk/home/alexc/papers/eacl2003.pdf

https://github.com/ninjin/clark_pos_induction

I note that Spitkovsky (at Google) uses a similar method in his more
recent work on unsupervised part of speech learning,

https://web.stanford.edu/~jurafsky/goldtags.pdf

These guys are doing clustering on sparse vectors derived via
co-occurrence of various sorts -- they're not using
dimension-reduction; though Spitkovsky is doing some dependency
parsing...

This is at bottom just EM clustering, but it's used in a way that's
nicely customized for part of speech induction...

This paper finds that Fuzzy C-Means outperforms EM on classifying
word2vec output vectors, in a somewhat different context:

http://dm.snu.ac.kr/static/docs/TR/SNUDM-TR-2016-11.pdf

-- Ben






-- 
Ben Goertzel, PhD
http://goertzel.org

"I am God! I am nothing, I'm play, I am freedom, I am life. I am the
boundary, I am the peak." -- Alexander Scriabin

-- 
You received this message because you are subscribed to the Google Groups 
"opencog" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/opencog.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/opencog/CACYTDBe4Zn40R2dNCp%3DYS0XfOrZOzCGiHSqhSO5s9drrcc8huQ%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to