{"ID":2863568,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2509.24716","arxiv_id":"2509.24716","title":"Discrete Variational Autoencoding via Policy Search","abstract":"Discrete latent bottlenecks in variational autoencoders (VAEs) offer high bit efficiency and can be modeled with autoregressive discrete distributions, enabling parameter-efficient multimodal search with transformers. However, discrete random variables do not allow for exact differentiable parameterization; therefore, discrete VAEs typically rely on approximations, such as Gumbel-Softmax reparameterization or straight-through gradient estimates, or employ high-variance gradient-free methods such as REINFORCE that have had limited success on high-dimensional tasks such as image reconstruction. Inspired by popular techniques in policy search, we propose a training framework for discrete VAEs that leverages the natural gradient of a non-parametric encoder to update the parametric encoder without requiring reparameterization. Our method, combined with automatic step size adaptation and a transformer-based encoder, scales to challenging datasets such as ImageNet and outperforms both approximate reparameterization methods and quantization-based discrete autoencoders in reconstructing high-dimensional data from compact latent spaces.","short_abstract":"Discrete latent bottlenecks in variational autoencoders (VAEs) offer high bit efficiency and can be modeled with autoregressive discrete distributions, enabling parameter-efficient multimodal search with transformers. However, discrete random variables do not allow for exact differentiable parameterization; therefore,...","url_abs":"https://arxiv.org/abs/2509.24716","url_pdf":"https://arxiv.org/pdf/2509.24716v3","authors":"[\"Michael Drolet\",\"Firas Al-Hafez\",\"Aditya Bhatt\",\"Jan Peters\",\"Oleg Arenz\"]","published":"2025-09-29T12:44:05Z","proceeding":"cs.LG","tasks":"[\"cs.LG\",\"cs.AI\",\"cs.RO\"]","methods":"[\"Transformer\",\"Variational Autoencoder\"]","has_code":false}