Evolutionary variational optimization of generative models

J Drefs, E Guiraud, J Lücke - Journal of machine learning research, 2022 - jmlr.org
We combine two popular optimization approaches to derive learning algorithms for
generative models: variational optimization and evolutionary algorithms. The combination is …

[PDF][PDF] A truncated EM approach for spike-and-slab sparse coding

AS Sheikh, JA Shelton, J Lücke - The Journal of Machine Learning …, 2014 - jmlr.org
We study inference and learning based on a sparse coding model with 'spike-and-slab'prior.
As in standard sparse coding, the model used assumes independent latent sources that …

A probabilistic account of exemplar and category generation

A Jern, C Kemp - Cognitive psychology, 2013 - Elsevier
People are capable of imagining and generating new category exemplars and categories.
This ability has not been addressed by previous models of categorization, most of which …

Learning and inference in sparse coding models with Langevin dynamics

MYS Fang, M Mudigonda, R Zarcone… - Neural …, 2022 - direct.mit.edu
We describe a stochastic, dynamical system capable of inference and learning in a
probabilistic latent variable model. The most challenging problem in such models …

Discrete sparse coding

G Exarchakis, J Lücke - Neural computation, 2017 - direct.mit.edu
Sparse coding algorithms with continuous latent variables have been the subject of a large
number of studies. However, discrete latent spaces for sparse coding have been largely …

GP-select: Accelerating EM using adaptive subspace preselection

JA Shelton, J Gasthaus, Z Dai, J Lücke… - Neural …, 2017 - ieeexplore.ieee.org
We propose a nonparametric procedure to achieve fast inference in generative graphical
models when the number of latent states is very large. The approach is based on iterative …

Are V1 simple cells optimized for visual occlusions? A comparative study

J Bornschein, M Henniges, J Lücke - PLoS computational biology, 2013 - journals.plos.org
Simple cells in primary visual cortex were famously found to respond to low-level image
components such as edges. Sparse coding and independent component analysis (ICA) …

Select-and-sample for spike-and-slab sparse coding

AS Sheikh, J Lücke - Advances in Neural Information …, 2016 - proceedings.neurips.cc
Probabilistic inference serves as a popular model for neural processing. It is still unclear,
however, how approximate probabilistic inference can be accurate and scalable to very high …

Truncated variational expectation maximization

J Lücke - arXiv preprint arXiv:1610.03113, 2016 - arxiv.org
We derive a novel variational expectation maximization approach based on truncated
posterior distributions. Truncated distributions are proportional to exact posteriors within …

Direct evolutionary optimization of variational autoencoders with binary latents

J Drefs, E Guiraud, F Panagiotou, J Lücke - Joint European Conference on …, 2022 - Springer
Many types of data are generated at least partly by discrete causes. Deep generative
models such as variational autoencoders (VAEs) with binary latents consequently became …