Add and use uncertainty in rules

Here are just the initial instructions: further details will be added during and after explanatory lecture at 12. April and practical work at 13. April.

Your task is to experiment with using numeric confidences and exceptions in rules and ask questions so that these are sensibly used.

To do this, you need to:

• Augment some of your existing (uncertain) rules/facts with numeric confidences,
• Add some new sensible uncertain rules/facts with numeric confidences,
• Augment some of your existing (uncertain) rules/facts with exceptions,
• Add some new sensible uncertain rules/facts with exceptions,
• Ask questions which use these enhanced rules/facts in a sensible way.
• Give a presentation with the example queries, results and explanations of the rules/facts.

Use the gk reasoner for these experiments. Currently only a Linux version of the gk reasoner is available. Please use wsl or a linux server for experiments.

How to give confidences/exceptions and how to get answers?

• For confidences read the paper and look at the examples at confer web page
• For exceptions read the draft paper and look at the gk web page with the gk reasoner and a wealth of examples.

Additionally, it helpful to have a look at the Problog tutorial: a somewhat similar system for confidences.

Useful examples built on our geodata:

• geodata wars and castles: built this during 20 April prax.
• geodata tourism city: built this during 27 April prax. Also used the strategy

{
  "strategy":["query_focus"],
  "query_preference": 1,
  "independence": 100,
  "equality":0,
  "max_answers": 20
}

and the command line

./gk_07 geodata.js  proveevents.js -strategy strategy3.js -seconds 100