1
$\begingroup$

I'm wanting to conduct game theoretic analyses of ongoing conflict situations (e.g. the US/North Korea negotiations; Syrian conflict; etc) as reported in the news media. I believe that AI may help me do this by helping me to pick out from the text: the parties involved; the issues over which they are in conflict; the choices they have; their preferences. However I'm not sure whether to approach this using 'modern' 'deep learning' approaches or to try something along the lines of the classic work by Schank, deJong etc. who used the notion of scripts (and sketchy scripts) in their work with conceptual dependency approaches. Does anyone have comments, suggestions that may guide my work please?

$\endgroup$
  • $\begingroup$ I would use pattern matching and frames with slots. It is much easier to see what's going on, and it will work without needing masses of training data. Schank and Riesbeck's Inside Computer Understanding includes a chapter on POLITICS by Jaime Carbonell, which sounds like what you have in mind. Another good source is Bart Kosko's Fuzzy Thinking where he describes using Fuzzy Cognitive Maps for modelling political situations. $\endgroup$ – Oliver Mason Mar 4 at 13:47
  • $\begingroup$ That sounds most encouraging Oliver. As it happens I opened my new copy of the book you mention yesterday and I see the material you mention. I contacted Roger Schank today and he came right back (!!) also with the suggestion to also see his stuff on stories. So I'm quite excited. Many thanks. $\endgroup$ – Jim Bryant Mar 4 at 14:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.