Thank you to everyone for your responses. It's been a fun weekend delving into some of this stuff.
Guyren, I particularly enjoyed working through some of your suggestions which inevitably led me to your OpenCyc talk at SDRuby (http://www.sdruby.org/podcast/76) - good stuff :) I am wondering if you still agree with the term AI being somewhat of a misnomer and that "Simulated Intelligence" is more descriptive of the work being done. I couldn't agree more. Have you played around with ConceptNet5 at all? http://conceptnet5.media.mit.edu/ I am finding the "ConceptuallyRelatedTo" relationships between phrases most useful. p.s. is it just me or are there no dates associated with the SDRuby podcasts? I think it's essential to know *when* a talk was given so we know how dated the material is. > > Some things that might be useful here: > > 1. Cyc (in the form probably of OpenCyc). I did a presentation on accessing > Cyc from JRuby maybe a year ago now. Among other things, Cyc provides the > most sophisticated ontology available. You didn't say what this is for, but > if you need some ability to deal with similes or automatically determining > things like subject areas, Cyc's ontology and inference capabilities may be > useful. Cyc's ontology also provides relations between its terms and the > contents of… > 2. Wordnet. This is a large computerized dictionary that provides similes, > word groupings and the like; > 3. OpenCalais (opencalais.com). This does an amazing job of entity extraction > (identifying "things": people, places, events etc) in natural language > documents (so it's for parsing documents, not for interpreting individual > sentences, say). > 4. If this is a commercial exercise you have a budget for, I know a number of > world-class computational linguists I can put you in touch with. > -- -- SD Ruby mailing list [email protected] http://groups.google.com/group/sdruby --- You received this message because you are subscribed to the Google Groups "SD Ruby" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/groups/opt_out.
