Starting a SourceForge project - need team members
Hey guys, I think I'm going to start a SourceForge project that relates to natural language processing. Please see this thread for details - http://www.gamedev.net/community/for...opic_id=367260
and let me know if you're interested (chad.d.johnson _AT_ gmail.com).
> All I'm looking for is help (people) and not criticism.
Like the combined resources of every single University and corporation which has ever worked on this problem?
Maybe you should look around the comp.ai.nat-lang newsgroup for a while as well. At least you'll find a lot more people who know a lot more about what they're talking about.
The reason I don't want to go with a university right now is because they are focused on their stuff, and I would not be able to focus on my stuff as much as I want to. And just like you guys, people would just tell me that I don't know what I am talking about and simply dismiss the idea, mainly because it is indeed a large project and they don't want to even try to get started.
Look people, I don't want to be nasty, but I really do have a plausible idea here. This is the main thing I have thought about for quite a while now, and I've spent many nights researching and contemplating about this. So I am not just another nut who wants to know the difference between DirectX and OpenGL like you're used to.
Thank you for the idea - maybe I will join a newsgroup to discuss my ideas.
Hey the newsgroup is a great idea! I've never used one of those before. I do appreciate your help! THAT is what I've been looking for (I was going to just discuss things with the group at first).
I used to be interested in Natural Language Processing specificaly text. I have a couple of questions.
The link you gave did not work for me.
For what reason do you want to create this?
The next question is actually a statement.
There where a bunch of french men interested in this field a long time ago.
They invented prolog.
So why C/C++?
I don't think you can use a GUI with Prolog. Can you use Prolog code from a C++ program somehow? I guess you could have an external prolog program/module. But Prolog also does not learn the language and I don't think it contextual abilities (not sure).
I don't know why I want to do this, I just do. It fits my way of thinking.
For what reason do you need a GUI to anlyze the contextual meaning of text.
Yes prolog is designed for assist with this.
The syntactic breakdown of a sentence is trivial. I think the contextual database and realworld knowledge is the difficult part. I was just trying to give you ideas on what to research as prolog has these database infrence capabilites (assuming you tell it how to think logically). Maybe you could create an extension to prolog.
A spider that garners information from the net based on contextual catagories then . This would be a major project though and beyond my ability to help.
Prolog is a procedural interpretation language. It allows the creation of heuristics to provide for inference, but I do not know if it allows for fuzzy logic. I imagine it must to be called an infrence.
Good luck, I hope I said somthing meaningful to you.
Thanks for your post and your insight. Sorry, I didn't meant to make it seem like I was trying to shoot down your idea.
I don't think that context should be represented by a database, but rather by the connections between concepts, which are modeled mathematically by the connections between neurons.
Prolog would probably be great for syntactic breakdown. It does seem kind of trivial compared to knowledge representation. Context would be the real challenge. Heuristics...I wonder if you would need to hard-code such rules, or if it's possible to teach them to the computer.
That gives me an idea...maybe I should just use Prolog for now and not worry about language learning. If I could have it do syntactic breakdown and create phrase structures, I could focus more on the knowledge representation. Thanks!
***EDIT***
Does the link work now? http://www.gamedev.net/community/for...opic_id=367260
Last edited by ChadJohnson; 01-05-2006 at 02:28 AM.
The idea is to create a set of basic rules for your heuristics then have the computer
recurse the rules with new information(from the text) to create new heuristics need to interpret the data.
If you are thinking of doing mathmatical neural net processing of the semantic meaning of a language (imo you are out there). From my understanding neural nets in terms of computers are simply matrixs with association function between the sets. I may be speaking like an idiot but to bridge the gap between this and determing the semantics of a piece of text is a waste of time (no offense meant).
After all classic logic with infrence is enough to provide simple feedback about the meaning of a document. Simple program like Elisia can even be done in BASIC, but prolog is designed specifically for this type of thing and it is a language albeit computer language. There are other options as well such as lisp.
(ohh btw you are talking to a bot, just kidding)
