abstract pattern matching

28 views
Skip to first unread message

Ben Goertzel

unread,
Dec 30, 2016, 10:08:25 AM12/30/16
to opencog, Zarathustra Goertzel
This 2002 thesis appears to explain the "right" way to do pattern
matching on graphs (he touches on labeled graphs, but the approach
should work for labeled hypergraphs too) ...

http://sqrl.mcmaster.ca/~kahl/Publications/RelRew/RelRew_TR2002-03.pdf

-- in this approach pattern matching rules are expressed without
reference to particular nodes or links, so all matching by default is
"up to homomorphism"

Among many other things, this would sorta do alpha-conversion
automagically behind the scenes...

Not suggested for immediate implementation, just putting it out there
as something to consider if we ever want to pretty much rebuild the
pattern matcher from the ground up ;p ...

-- Ben

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

“I tell my students, when you go to these meetings, see what direction
everyone is headed, so you can go in the opposite direction. Don’t
polish the brass on the bandwagon.” – V. S. Ramachandran
Reply all
Reply to author
Forward
0 new messages