Jump to content
  • Advertisement

Archived

This topic is now archived and is closed to further replies.

Kylotan

Resources on natural language generation?

This topic is 5537 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

From what I learned, most of the natural language processing research is into processing existing bodies of text,and extracting information from it. However, from a gaming point of view, I''m more interested in the reverse operation - taking formal data and information and producing text from it. In particular, I want to be able to generate interesting descriptions of rooms, similar to those in MUDs or interactive fiction/text adventures, based on data such as the properties of the room, which rooms are adjacent, and so on. At a basic level, it''s pretty easy to generate a set of phrases with variable ''placeholders'', and substitute random words in as appropriate. This works, but is not interesting, and certainly doesn''t come across as ''natural''. Also, it requires something of a brute force approach in that someone has to write lots of these boilerplate phrases in order for there to be enough that they seem fairly unrepetitive to the end user. I can patch it up with an ''optimiser'' that adds endophora (for example), but I don''t think this would go far enough. So basically I have 2 questions: the first is whether there are any research papers or articles out there that address this kind of thing. The second is just for comments on my approach and ways to improve it. I appreciate this is on the boundaries of AI,but since I am interested in any ''rigorous'' NLP information as well as more informal methods, I thought I''d ask here first, possibly following up in the Game Design forum later. [ MSVC Fixes | STL Docs | SDL | Game AI | Sockets | C++ Faq Lite | Boost
Asking Questions | Organising code files | My stuff | Tiny XML | STLPort]

Share this post


Link to post
Share on other sites
Advertisement
Look into Roger Schank''s research in the ''70s. I can''t remember the name of the module which did the text generation, but it worked with SAM, PAM and POLITICS.

___________________________________

"The outside of a horse is good for the inside of a man" - Winston Churchill

Share this post


Link to post
Share on other sites
Your best bet is to go to your university library and get your hands on the book Inside Computer Understanding, and from there, proceed to read Schank''s other books if you find them as well. That book and his other books essentially explain to you the philosophy and explicit mechanisms to do this, beginning with how to internally represent the meaning of something before generating text to convey that meaning.

___________________________________

"The outside of a horse is good for the inside of a man" - Winston Churchill

Share this post


Link to post
Share on other sites

  • Advertisement
×

Important Information

By using GameDev.net, you agree to our community Guidelines, Terms of Use, and Privacy Policy.

Participate in the game development conversation and more when you create an account on GameDev.net!

Sign me up!