Member Profile for dgandhi

says...

http://www.logicaldiscourseproject.org/ldpforum/viewtopic.php?f=2&t=9&start=0

idea for you based on your earlier ideas about pre-formatting

vairetube

says...

first off, excellent suggestions as always. i was so set on waiting for wave i didnt even look at appengine. /dumb

"A program to mine available data, which then provides a context which is sufficient to assist the program in an assessment of valid statements, while allowing corrections for invalid assertions made in the source data."

You've got it covered, and so well put I don't have any substantive technical answer as to what the AI needs to be able to perform at the desired level of analysis.

Perhaps start by identifying what is never useful to our own decision making process -- which then starts to define what might be used? All of which depends on the validity of our processes, to be determined by varied opinions and hopefully an eventual logical consensus. Too obvious? Is there a definitive work towards this end? I haven't looked sorry im slacking

have you seen http://www.wolframalpha.com/ ?
and i think i linked this before, but does this tickle your fancy at all:
http://search.cpan.org/~fimm/Lingua-EN-Summarize-0.2/Summarize.pm

i need to look up more details about App Engine but it solves a large problem by having the infrastructure and API's available ... I will make an account and download the SDK. I have the current stable Python "sdk" and from what i've seen so far it is quite powerful in its simplicity... I have no useful personal experience with it.. yet.

I am taking java oriented programming classes and visual basic classes starting late september... i am teaching myself what i can, but i learn .. faster.. in the classroom than on my own. i've got my fingers in a lot of pies but i dont even know what flavor they are yet, so to speak. i don't know enough to know what i would prefer... sad i know. im taking summer quarter this year so other things are hurting my brain and taking up time currently. i have a real open break starting August 22nd.

i can provide a forum you can admin to discuss programming-syntax specific ideas and such, at least... it behooves me to observe at this point.

Forums are up: Go ahead and register so I can set your permissions. http://www.logicaldiscourseproject.org/ldpforum

edit: i forgot to mention, i am taking an intensive foreign language class this summer.... and conversely, i am learning a lot about how English works. I am having certain ideas but ill have to see what i produce from my Notepad ramblings.

vairetube

says...

Awww yea... i had a bit of a breakthrough with the flow chart... i re-organized the flows and components... and it looks really solid, i must say.

the I/O organization was not... correct, and it was bugging me... ill have something to present soon~ you're the only one who remotely knows wtf im talking about... so i had to tell ya~ heh

also, check this out: http://code.google.com/apis/wave/extensions/robots/index.html

shit gets easier every day, i swear... man i wish i could start working more on this now but i still have a lot to learn.

someone will probably beat me to it at this rate... that might even be a relief! LOL! ive refined http://logicaldiscourseproject.org/missionstatement.htm to reflect that i am intent on using these newer API's...

vairetube

says...

The biggest problem i see is that there are so many ways to approach to the problem that all seem to do what is needed, but right now are scattered between researchers and projects all over the world, and i dont have access to them! For instance, you have massively excellent ideas, but would i have even known about them if i didnt come on videosift? so who else might already have another piece to the puzzle that i am completely ignorant of? that is my base starting point, because of my experience level and resources.

you are correct in approaching the linguistic AI component as the heaviest and most critical task to work on. Everything else is trivial in comparison.

The researchers at Cornell started a project in 2004 with a 10 year goal to accomplish basically what we're talking about in terms of analysis, only for a different purpose. I would not like to... spend time on that component when that research will save time and ultimately be better than what i can think of or make with my resources. I could essentially dig the hole, and fill it in with their system... a desire not driven by a particular ... inability... but out of desire for maximum efficiency.

I seem to be my best at unifying already developed systems, so right now... im identifying what components are needed, in what stage, if any, do they exist currently, where do they exist, and how can i use them in this new way.

The database component and GUI is likely where i personally could do the most work, but soon i will have access to the brains here at the CS department, and hopefully they will be interested enough to give me pointers in the right direction for the rest. I am about 7 months out from even getting to that point, so in that time, i am going to keep basically brainstorming and researching all the stuff i can so that the concept is presentable and accessible to new minds.

I can't contribute much at this point besides vision, time, and energy, but i feel like its such a worthy goal that i could easily spend a lifetime on it and consider it well spent. Your ideas make total sense to me, but... i wouldnt have been able to produce them in such logical context as efficiently... so at least, i can identify the need to be educated and educate myself more. I want to see what specific areas i excel at here at school so i feel i have validity to ... waste the time of more brilliant individuals such as yourself.

I mean, you already have such practical solutions for an AI framework, and i haven't even done anything as defined. so it makes me feel .. unworthy of attempting to interact at your level when i know i am not ready... but i know now you have a piece to the puzzle... so progress has been made... perhaps soon i can make the call to arms for the project in a way that is appropriate to the end goal, which is participation of everyone.

That's all i can spew forth ... once again thank you for making me see that it really is important to other people... that alone is enough to keep me going on this. I belive soon i will make a .org website for people to start brainstorming on, with a forum and such... that is how i can most effectively contribute at this point... organizing channels for ideas regarding the various components, as they are identified. My younger brother, for instance, is getting heavy training in the PeopleSoft/oracle suites, so there is another piece right there.

Sorry to be so vague in response to your very specific ideas.. all i can say to them is... YES MOAR PLS!1 and lol you found the same funny definitions for the acronym i did... heh heh heh. There needs to be a clearing house for these ideas... that's where i can come in the soonest, i believe... i would love to watch the conversation develop between more parties with interests and abilities such as yours! ok, well, back to the drawing board to channel some of this excitement productively! i have a large break time coming up in late august, and with that i hope to be in full swing making the website and "conceptual" framework for others to spin off of.

boy this will be neat... thanks D!

vairetube

says...

the long name i want for it is:

Data Capture/Catalogue/Recall System for Facilitating Logical Discourse

...or DCCRS, to make a neat acronym. "Hey, let's check the DCCRS"

but there is probably some more specific or just cooler lingo to toss in. naming it will be almost as fun as working on it!

vairetube

says...

that may be the smartest thing ive ever heard.... i will use your idea as inspiration for projects when i get into upper level AI courses... man that will be a fantastic idea to bounce around with people who know more about implementing it. ill make sure to tell them "Gandhi" told me...

Obama yesterday mentioned that a opensource clearinghouse for virtual college courses is to be funded along with his comm. college stimulus, so that people can better receive education regardless of physical location -- i thought that was pretty nifty. sort of like the open MIT and such, only not school specific or owned.

we're on our way muchas gracias for your help and encouragement.

---------------------

The part that interests me, and that the 'net is missing is:

"facilitate only logical discourse"

Creating a mechanism which allows the users to make any meaningful statement without being able to make statements which have rhetorical effect. If we can find a way to mechanically confine discourse to meaningful statements, without hobbling discussion, then we can trivially attach dis/proof to such statements and facilitate a progression of thought/discourse.

Everything else can be scraped off the 'net.

vairetube

says...

i realize what i want is....basically the internet.... only more user friendly and cross checked... but that is where a new interface, and data storage/access, and security, will need to be refined.

there, i finally finished my thought.

vairetube

says...

Awesome.. thanks

right now I'm thinking of local government implementations, because that's where a lot of stuff can get done... city governments, school board meetings, etc... ive been in those environments where i could really use some back up for my ideas but lose ground by not having ready access to everything i can think of, or partially think of.

Hell, even an IPhone is almost good, but i want a system that works small scale as well as be able to integrate into a larger system; also cost effective, and easy to use... heh heh easy right

You would need local operations centers to employ people to manually tag and adjust data, which isnt all bad.. but an AI to perform the task and take out the human element is preferable to me, as long as it is able to maintain .. integrity.. i guess is the word. that will be a big element.

i think we're trending that way, so likely someone will have done this before i get around to it... but i don't care who does it, i just want it done! it's basically nothing more than a fancy combo of Wikipedia/YouTube with more intense moderation and review. It could be done... we're so close yet so far...

I have had similar thoughts.

I think that this would end up being an AI-hard natural language problem, but if you actually have some plans to implement it I would be happy to help with coding.