belief-optimization

Tag

Cards List
#belief-optimization

CHAL: Council of Hierarchical Agentic Language

arXiv cs.AI · 2026-05-14 Cached

This paper introduces CHAL, a multi-agent dialectic framework that treats defeasible argumentation as structured belief optimization for LLM reasoning, using configurable meta-cognitive value systems and a gradient-informed belief revision mechanism.

0 favorites 0 likes
← Back to home

Submit Feedback