Skip to content
DraconiticResearch
Search Sources API
203,203 studies31 sourcesUpdated 49 min ago
PreprintCC BY 4.0arXiv

A Generalization of Amari's Bayesian Duality

  • Mohammad Emtiyaz Khan
  • Thomas Möllenhoff
Open at sourcePDFDOICite

Abstract

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.

Related

  1. A Sharp Fisher-Information Inequality for Boolean Pairs: Proof of the Amari--Kobayashi Conjecture
    Jun Chen, Shun Watanabe, Lei Yu
    arXivPreprintarXiv licenceFull textGeneral mathematicsarXiv
  2. Global front solutions converging to an asymptotic traveling wave in a time-fractional Amari model
    Hiroshi Ishii
    arXivPreprintarXiv licenceFull textAnalysis & PDEsarXiv
  3. A Riemannian viewpoint on the Amari–Cencov α-connections and Proudman–Johnson equations
    Martin Bauer, Alice Le Brigant, Cy Maor
    NonlinearityPeer-reviewedCC BY 4.0Full textGeometry & topologyHAL
  4. Promoting Generalization for Exact Combinatorial Solvers via Adversarial Instance Augmentation
    Haoyang Liu, Jie Wang, Yufei Kuang +3
    IEEE Transactions on Pattern Analysis and Machine IntelligencePeer-reviewedLicence not statedFull text13 citedNumerical analysis & scientific computingOpenAlex
  5. Uniform generalization of Goodstein's theorem and Cichon's independence proof
    Gunnar Wilken
    arXivPreprintCC BY-NC-ND 4.0Full textLogic & foundationsarXiv

Cite

BibTeXRIS
BibTeX
@misc{khan2026generalization,
  title = {{A Generalization of Amari's Bayesian Duality}},
  author = {Mohammad Emtiyaz Khan and Thomas Möllenhoff},
  howpublished = {arXiv},
  year = {2026},
  month = sep,
  doi = {10.48550/arxiv.2609.09126},
  eprint = {2609.09126},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2609.09126},
  note = {CC BY 4.0},
}
RIS
TY  - UNPB
TI  - A Generalization of Amari's Bayesian Duality
AU  - Mohammad Emtiyaz Khan
AU  - Thomas Möllenhoff
PY  - 2026
DA  - 2026/09/08
PB  - arXiv
DO  - 10.48550/arxiv.2609.09126
UR  - https://arxiv.org/abs/2609.09126
AB  - 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.
ER  -