How does Dung's Grounded Semantics framework work in practice? I got this from slides of an AI course but can't figure it out:
Grounded Semantics is said to minimize amount of arguments IN (green) while Preferred Semantics tries to maximize this. I understand the Preferred Semantics but don't see how all arguments would be defensible according to Grounded Semantics. Wouldn't C always be OUT and D always IN?
If anyone could help me out on this it'd be much appreciated!
You might find the paper at the link,
https://pdfs.semanticscholar.org/21a1/a88328f8e274816a70061666ee2ded6f8235.pdf
to be helpful here. There is a definition on page 6 of what it means to defend an argument, and, it seems to justify the statement of your diagram. What is more cryptic with your graphic is that the grounded semantics is $\emptyset$ to the best that I can tell -- no nodes are highlighted with green. So, while node C is defensible, it is not defensible with a non-conflicting set of arguments. Nodes A and B conflict with one another. So, the grounding set must be empty.
Hopefully, you will find an expert. I am just a guy who can read definitions.
Thanks for your question -- interesting subject.