[Prev][Next][Index][Thread]

Re: about current state of the art of LL in AI...



[For new readers that may wonder why this showed up in their
mailboxes... note that the LINEAR mailing list is mirrored to
TYPES. --BCP]

> To: Vladimir Alexiev <vladimir@cs.ualberta.ca>
> From: jpalacio@mia.uv.mx
> could you tell me aboput the current state of the art relating Linear Logic
> in Artificial Intelligence? I want to keep this line of research as much as
> possible.

This question is probably better addressed to the linear mailing list, so I am
forwarding it there. Here are the things that come to my mind. Probably you
already know about some of them. You can find bibliographic references at
http://pecan.srv.cs.cmu.edu/afs/cs/user/iliano/www/linearbib/ 
or http://www.cs.cmu.edu/%7Eiliano/linearbib/linearbib.html. In my opinion,
there is still very little done in this area, and there's potential for much
more. 

- Deductive planning (Masseron+Tollu+Vauzielles; Holdoebler and others at
  Darmstadt).
- Hierarchies with exceptions (Fouquere+Vauzielles).
- Database updates based on the previous work (Vauzeilles+?).
- Explanations (abduction) (Arima).
- NLP (Hodas+Miller), categorial grammars (Morill).
- Connection between LL and Fuzzy logic (Barr
  http://theory.doc.ic.ac.uk/tfm/papers/BarrM/fuzzy.models.ps.Z;  Cardoso+
  Valette+Pradin-Chezalviel; Kreinovich+Nguyen)
- There's a lot of work on LL LP, which isn't necessarily related to AI, but
  it provides programming languages with potential applications to AI that are
  more sophisticated than Prolog. Some of these languages are (no doubt I am
  missing some):
  - LO (Andreoli+Pareschi). Work moved in the direction of multiset-based
    coordination languages and agents
    (http://www.xerox.fr/grenoble/ct/home.html) 
  - hlcc (Sarawsat+Lincoln).
  - ACL (Kobayashi+Yonezawa).
  - LLP (http://bach.seg.kobe-u.ac.jp/llp/).
  - Lolli (Hodas+Miller), Forum (Miller).
  - Lygon (Winikoff+...).
  Of the traditional (and more widely used :-) prolog-based systems, BinProlog
  seems to pay most attention to LL developments.