Paper-Conference

Expected free energy as an information constraint on the Bethe Lagrangian

Active inference selects actions by minimising an expected free energy functional of predicted futures. However, the expectation over …

Wouter M. Kouw

Balancing time and precision with active sensor selection and positioning

High-precision sensing systems often have multiple sensor modalities at their disposal, each trading measurement precision against …

Esther Van Pelt, Alexandru Onose, Ruud Van Sloun, Harm Belt, Wouter M. Kouw

Expected free energy planning with learned external-camera measurement reliability

A robot planning under uncertainty must trade goal progress against whether it will be able to localize itself effectively at potential …

Joost Leliveld, Ömür Arslan, Wouter M. Kouw

Expected free energy-based informative path planning for robotic mars exploration

An autonomous robot efficiently exploring an unknown environment, such as looking for water sources on Mars, faces two simultaneous …

Ajith Anil Meera, Pablo Lanillos, Wouter M. Kouw

A factor graph approach to scalable multi-output Gaussian process regression

Multi-output Gaussian process regression scales cubically in the number of observations times outputs, and dense kernel-matrix methods …

Wouter Nuijten, Albert Podusenko, İsmail Şenöz, Esther Van Pelt, Wouter M. Kouw

Spike-based belief propagation in nonlinear dynamical systems

This paper presents a Bayesian control framework that integrates spike-based dynamics with probabilistic inference for adaptive …

Sepideh Adamiat, Hongye Wang, Wouter M. Kouw, Bert De Vries

Composing non-conjugate factor graphs with closed-form variational inference

We identify five factor-graph primitives: a bilinear factor, an exponential link, a Gamma prior, a Gaussian likelihood, and an equality node, and prove that any model composed from them admits closed-form variational message passing

Mykola Lukashchuk, Kyrylo Yemets, Wouter M. Kouw, Dmitry Bagaev, İsmail Şenöz, Jeff Beck, Bert De Vries

Effects of priors on epistemic uncertainty in autoregressive active inference

Intelligent agents learn to solve tasks through interactions with their environment. We study autoregressive active inference agents …

Esther Van Pelt, Tim N. Nisslbeck, Harm J.W. Belt, Ruud J.G. Van Sloun, Wouter M. Kouw

Bayesian autoregression to optimize temporal Matérn-kernel Gaussian process hyperparameters

We present a probabilistic numerical procedure for optimizing Matérn-class temporal Gaussian processes with respect to the kernel covariance function’s hyperparameters based on Bayesian autoregression.

Wouter M. Kouw

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 M. Kouw, Tim Nisslbeck, Wouter Nuijten

Message passing-based inference in an autoregressive active inference agent