AI researchers propose new logic for defeasible conditional obligations

This paper introduces a novel two-tiered semantic framework for modeling defeasible conditional obligations, building upon preference-based semantics for deontic logic. It incorporates a nonmonotonic reasoning mechanism to allow for the withdrawal of obligations when new information arises. The framework utilizes two distinct orderings, ideality and normality, to address limitations in previous approaches and establishes a connection with constrained input/output (I/O) logic. AI

IMPACT Introduces a new formal framework for reasoning about obligations, potentially impacting AI systems that require normative reasoning capabilities.

RANK_REASON This is a research paper published on arXiv detailing a new semantic framework for defeasible conditional obligations.

Read on arXiv cs.AI →

• alphaXiv
• arXiv
• CatalyzeX
• DagsHub
• Gotit.pub
• Hansson-Lewis
• Hugging Face
• ScienceCast
• Xavier Parent

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →