Search papers, labs, and topics across Lattice.
This work revisits Shun-ichi Amari's underappreciated formulation of Bayesian duality, establishing a direct theoretical bridge between it and the convex duality inherent to Bayes' rule. By unifying these two perspectives, the authors derive a generalized framework for Bayesian duality that reconciles information-geometric structures with optimization-theoretic updates. The resulting formulation expands the geometric toolkit for analyzing modern probabilistic inference, variational learning, and dual-coordinate optimization in AI systems.
Amari's long-overlooked Bayesian duality and standard convex duality are fundamentally two sides of the same coin, unlocking a unified geometric foundation for probabilistic update rules in AI.
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.
