Ai gf

>>8475691
shit idea

File: FF-18-10-2016-STD.jpg (168KB, 1024x768px) Image search: [Google]

168KB, 1024x768px

gee, what cud go wrong when u teach artefecial intelligance to thinkk?

>>8475691
So you're doing natural language processing the hard way? Your approach needs too many humans.

There are much better ways to do this.

>>8475742
Like what? I dont mean to put too much emphasis on tl;dr'ing, i also want him to be capable of creating novel content, among other things.

>>8475742
Also, for the record, of course i want to get it to the point where he assembles his vocabulary, use a search engine to resolve queries about each new word, repeat ad infinitum- with the ultimate goal of being able to take that data and use it to assemble what would basically be a database of predictions and profiles.

>>8475756
>>8475751
And how do you plan to accomplish this?

>>8475691
Looks interesting OP. What language are you programming it in?

>>8475760
Eking it out as we speak.

>>8475789
C#. Its all just string manipulation.

>>8475760
To be more specific- when it learns a new word, it assigns an int to each possibility (eg is it a proper noun) within the script: 0 is unknown, 1 is yes, 2 is no.

>>8475856
How does it learn a new word? What is the algorithm for doing so?

Second, even if you know which words are nouns, verbs, and what not, this information is pretty useless for understanding sentences. This is because the meaning of words depends on the contexts in which they are used.

If you just want to determine if words are nouns, verbs, etc just use a dictionary

If not take a look at some of deepmind's most recent stuff
https://deepmind.com/research/publications/