Picture for Nihat Ay

Nihat Ay

Max Planck Institute for Mathematics in the Sciences, Leipzig, Germany

Analyzing Multimodal Integration in the Variational Autoencoder from an Information-Theoretic Perspective

Add code
Nov 01, 2024
Viaarxiv icon

A Concise Mathematical Description of Active Inference in Discrete Time

Add code
Jun 11, 2024
Viaarxiv icon

On the Fisher-Rao Gradient of the Evidence Lower Bound

Add code
Jul 20, 2023
Viaarxiv icon

Inversion of Bayesian Networks

Add code
Dec 20, 2022
Viaarxiv icon

Invariance Properties of the Natural Gradient in Overparametrised Systems

Add code
Jun 30, 2022
Viaarxiv icon

Natural Wake-Sleep Algorithm

Add code
Aug 15, 2020
Figure 1 for Natural Wake-Sleep Algorithm
Figure 2 for Natural Wake-Sleep Algorithm
Figure 3 for Natural Wake-Sleep Algorithm
Figure 4 for Natural Wake-Sleep Algorithm
Viaarxiv icon

On the Locality of the Natural Gradient for Deep Learning

Add code
May 21, 2020
Figure 1 for On the Locality of the Natural Gradient for Deep Learning
Figure 2 for On the Locality of the Natural Gradient for Deep Learning
Figure 3 for On the Locality of the Natural Gradient for Deep Learning
Figure 4 for On the Locality of the Natural Gradient for Deep Learning
Viaarxiv icon

Comparing Information-Theoretic Measures of Complexity in Boltzmann Machines

Add code
Jul 30, 2017
Figure 1 for Comparing Information-Theoretic Measures of Complexity in Boltzmann Machines
Figure 2 for Comparing Information-Theoretic Measures of Complexity in Boltzmann Machines
Figure 3 for Comparing Information-Theoretic Measures of Complexity in Boltzmann Machines
Figure 4 for Comparing Information-Theoretic Measures of Complexity in Boltzmann Machines
Viaarxiv icon

Information Theoretically Aided Reinforcement Learning for Embodied Agents

Add code
May 31, 2016
Figure 1 for Information Theoretically Aided Reinforcement Learning for Embodied Agents
Figure 2 for Information Theoretically Aided Reinforcement Learning for Embodied Agents
Figure 3 for Information Theoretically Aided Reinforcement Learning for Embodied Agents
Figure 4 for Information Theoretically Aided Reinforcement Learning for Embodied Agents
Viaarxiv icon

Geometry and Determinism of Optimal Stationary Control in Partially Observable Markov Decision Processes

Add code
Feb 13, 2016
Figure 1 for Geometry and Determinism of Optimal Stationary Control in Partially Observable Markov Decision Processes
Figure 2 for Geometry and Determinism of Optimal Stationary Control in Partially Observable Markov Decision Processes
Figure 3 for Geometry and Determinism of Optimal Stationary Control in Partially Observable Markov Decision Processes
Figure 4 for Geometry and Determinism of Optimal Stationary Control in Partially Observable Markov Decision Processes
Viaarxiv icon