A Qualitative Logic for Uncertain Evidence and Belief Comparison
| Authors | |
|---|---|
| Publication date | 2025 |
| Host editors |
|
| Book title | Proceedings of the Workshop on AI for Evidential Reasoning |
| Book subtitle | co-located with the 38th International Conference on Legal Knowledge and Information Systems (JURIX 2025) : Turin, Italy, December 9th, 2025 |
| Series | CEUR Workshop Proceedings |
| Event | Workshop on AI for Evidential Reasoning |
| Article number | 2 |
| Pages (from-to) | 8-22 |
| Number of pages | 15 |
| Publisher | Aachen: CEUR-WS |
| Organisations |
|
| Abstract |
We introduce a qualitative logic for comparing strengths of belief and evidential support explicitly, discerning these two comparative notions both syntactically and semantically within a modal logical framework. More precisely, we employ Dempster-Shafer theory (DST) of belief functions to represent uncertain, possibly mutually inconsistent, and incomplete evidence, as well as evidence-based degrees of beliefs. We propose a bi-modal logic that compares propositions in two ways: (1) based on the strengths of belief an evidence-possessing agent has in them and (2) based on the degrees of certainty of the evidence supporting them. (2) is the novel component of the proposed logic, designed to capture a notion of certainty-dominance among sets of evidence, modeled via an Egli-Milner-like order lifting on individual pieces of evidence. We justify this modeling choice, provide key (in)validities of our logic, and establish links to existing modal logics of evidence and belief (functions).
|
| Document type | Conference contribution |
| Language | English |
| Published at |
https://ceur-ws.org/Vol-4157/paper2.pdf
(Final published version)
|
| Other links | |
| Downloads |
paper2-1
(Final published version)
|
| Permalink to this page | |
