A Generalization of Amari's Bayesian Duality
Paper generalizes Amari's Bayesian duality by connecting it to a convex duality of Bayes' rule.
The authors revisit Amari's less-known work on Bayesian duality from information geometry. They connect Bayesian duality to a convex duality formulation of Bayes' rule and present a generalization of it. The paper is purely theoretical and discusses relevance for modern AI, with no experiments or model releases.
- Connects Amari's Bayesian duality to convex duality of Bayes' rule
- Proposes a generalization of Bayesian duality
- Discusses implications for modern artificial intelligence
Full article57 words · extracted from arxiv.org · click to collapse
Amari's contributions to information geometry and machine learning are well known. Here, we revisit Amari's work on Bayesian duality which has not received as much attention. We connect Amari's Bayesian duality to a convex duality of Bayes' rule. Using this connection, we present a generalization of Amari's Bayesian duality and discuss its relevance for modern artificial intelligence.
Text extracted automatically; images, tables and formatting may be missing. Original: https://arxiv.org/abs/2609.09126