Active Inference · Paper · 2026

Robust Belief Sharing in Federated Active Inference: A Recovery-Tested Generalized-Variational Framework for Categorical Contamination-Aware Consensus

Daniel Ari Friedman

Zenodo

Read PDF on Zenodo (version 1.1.0) Publication source Paper folder on GitHub Read extracted text

Overview

The selected Active Fedference 1.1.0 manuscript presents a discrete-categorical connection between active-inference belief sharing and a generalized-Bayes family. It separately assigns claims to source-conditional client updates, a recovery-only heuristic server rule and an objective-backed variational server rule; no robustness guarantee transfers between those axes. The manuscript reports recovery, contamination, sensitivity and socket-backed studies from deterministic artifacts. This catalog refresh does not reproduce those studies or establish general robustness or real-world deployment efficacy.

active inferencefederated learninggeneralised variational inferencebelief sharingrobustnessFedGVI

Overview source: Curated paper metadata.

Methods and contributions

Read the source for the full argument, qualifications, and evidence.

Findings and contributions

  • The reported containment result recovers ordinary belief sharing when robustness is turned off.
  • Client, heuristic-server and variational-server statements are explicitly kept separate rather than inheriting one another's guarantees.

Methods

  • Discrete-categorical generalized-Bayes belief sharing — Described in the selected deposited manuscript; source-text review does not re-execute or independently validate this method.
  • Separate robustness claim owners — Described in the selected deposited manuscript; source-text review does not re-execute or independently validate this method.

Summary sources: Paper metadata and evidence · Extracted source text.

Document downloads

The selected PDF is hosted at Zenodo; local files remain preserved archives.

Current PDF version 1.1.0 is hosted at Zenodo. Local PDFs remain preserved historical archives. Extracted text retains its own source provenance; a matching filename does not establish edition identity.

Citation

Citation metadata follows the unified bibliography.

Friedman, Daniel Ari. 2026. Robust Belief Sharing in Federated Active Inference: A Recovery-Tested Generalized-Variational Framework for Categorical Contamination-Aware Consensus. Zenodo. DOI: 10.5281/zenodo.21864003. URL: https://doi.org/10.5281/zenodo.21864003.
Download bibliography Add to reading list Explore connections

Catalog details and resources

Catalog row210
Citation keyFriedman2026RobustBeliefSharingFederated210

Related works

Ranked by recorded methods, topics, contributors and associated software, with contributor roles and domain-only matches labelled. Metadata overlap is not a citation or a scientific equivalence.

View all Active Inference works, software & media →