As far as I know, OpenCog doesn't have any language capabilities (yet), and there is currently no development work in this direction. There are several people working on development, but like any large software project, a lot of the work is bogged down on fixing bugs, porting issues, and updating the documentation. They do have a traditional rule based parser (RelEx) and sentence generator which I don't believe is integrated into AtomSpace. This is probably not what you want if you are interested in a scalable design for language learning.
The current work in OpenCog is toward developing intelligent characters for games and virtual worlds. This avoids the computationally intensive algorithms required for vision, language, and robotics. There has been some discussion of making a distributed version of AtomSpace but IMHO there are going to be severe performance problems that make scaling to large numbers of processors impractical. Google and Watson require thousands of CPUs to do any significant language processing. I observe a similar trend in text compression. A significant portion of the human brain is also required to process language, suggesting that the problem requires computation on a level similar to that of other hard problems like vision or robotics. This seems counter-intuitive, given that it's only text. The problem is not well understood. Why is it that a human brain sized neural network requiring a 10 petaflop computer can apparently only add numbers at a rate of about one digit per second? On Sun, Sep 23, 2012 at 8:29 PM, Anastasios Tsiolakidis <[email protected]> wrote: > Your answers will arrive soon in the usual mix of "read my book", > "search the list", even in plain English. I would just like to remind > you that you can always benefit more, yourself and others, by having a > go at answering your own questions. I am particularly interested in > what you think could or should " mean for an OpenCog agent to > "understand" dialog"! > > AT > > On Mon, Sep 24, 2012 at 1:18 AM, Eastendtech <[email protected]> wrote: >> Hi, I'm a novice taking up AI as a hobby for the first time. I apologize in >> advance if my questions are too vague/not technical enough for this list but >> I've been an avid learner/lurker for some time. Im more interested in Sentic >> Computing and systems that can infer human emotions. My interests concern >> the natural language processing elements of OpenCog and what it means for an >> OpenCog agent to "understand" dialog. Also, the other question I have is how >> would the learning algorithms and the Probabilistic Logic Network engine >> that's at the heart of opencog mesh with the narrow AI "knowledge >> ontologies" like the CommonSense Computing knowledge base or Cyc, would it >> be possible for a chat-bot type embodied OpenCog mind to use a Knowledge >> Engine as its own internal "Google" to learn things into its AtomSpace and >> what would be the best way to "hook" it to a data source such as that so it >> could make queries when it gets stuck in a conversation or would like more >> info, etc? Final question is, is there anything of use that can be applied >> from CALO (the project that Siri branched off it) into the OpenCog >> architecture? Thanks for taking the time to answer my questions. -- Matt Mahoney, [email protected] ------------------------------------------- AGI Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-c97d2393 Modify Your Subscription: https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-2484a968 Powered by Listbox: http://www.listbox.com
