Paper-Conference

Composing Non-Conjugate Factor Graphs with Closed-Form Variational Inference

Stacking probabilistic building blocks into deeper architectures typically breaks closed-form inference. We show that closed-form …

Mykola Lukashchuk, Kyrylo Yemets, Wouter Kouw, Dmitry Bagaev, İsmail Şenöz, Jeff Beck, Bert de Vries

The Quotient Bayesian Learning Rule

This paper introduces the Quotient Bayesian Learning Rule, an extension of natural-gradient Bayesian updates to probability models …

Mykola Lukashchuk, Raphaël Trésor, Wouter Nuijten, İsmail Şenöz, Bert de Vries

Active Inference is a Subtype of Variational Inference

Automated decision-making under uncertainty requires balancing exploitation and exploration. Classical methods treat these separately …

Wouter Nuijten, Mykola Lukashchuk

A Message Passing Realization of Expected Free Energy Minimization

We present a message passing approach to Expected Free Energy (EFE) minimization on factor graphs, based on the theory introduced in …

Wouter Nuijten, Mykola Lukashchuk, Thijs van de Laar, Bert de Vries

Message passing-based inference in an autoregressive active inference agent

We present the design of an autoregressive active inference agent in the form of message passing on a factor graph.

Wouter Kouw, Tim Nisslbeck, Wouter Nuijten

Message passing-based inference in an autoregressive active inference agent

ExponentialFamilyManifolds.jl: Representing exponential families as Riemannian manifolds

ExponentialFamilyManifolds.jl implements exponential family natural parameter spaces as Riemannian manifolds, enabling geometric …

Mykola Lukashchuk, Dmitry Bagaev, Albert Podusenko, İsmail Şenöz, Bert de Vries

Online Bayesian system identification in multivariate autoregressive models via message passing

We propose a recursive Bayesian estimation procedure for multivariate autoregressive models with exogenous inputs based on message passing in a factor graph

Tim Nisslbeck, Wouter Kouw

Improved Depth Estimation of Bayesian Neural Networks

This paper proposes improvements over earlier work by Nazareth and Blei (2022) for estimating the depth of Bayesian neural networks. …

Bart van Erp, Bert de Vries

Riemannian Black Box Variational Inference

We introduce Riemannian Black Box Variational Inference (RBBVI) for scenarios lacking gradient information of the model with respect to …

Mykola Lukashchuk, Wouter Nuijten, Bvdimitri, İsmail Şenöz, Bert de Vries

Coupled autoregressive active inference agents for control of multi-joint dynamical systems

We propose an active inference agent to identify and control a mechanical system with multiple bodies connected by joints. This agent …

Tim Nisslbeck, Wouter Kouw