paper-with-me

Papers

On the advantages of stochastic encoders

2021-02-18 · ICLR Workshop Neural_Compression 2021 5 · Lucas Theis, Eirikur Agustsson

Stochastic encoders have been used in rate-distortion theory and neural compression because they can be easier to handle. However, in performance comparisons with deterministic encoders they often do worse, suggesting that noise in the encoding process may generally be a bad idea. It is poorly understood if and when stochastic encoders do better than deterministic encoders. In this paper we provide one illustrative example which shows that stochastic encoders can significantly outperform the best deterministic encoders. Our toy example suggests that stochastic encoders may be particularly useful in the regime of "perfect perceptual quality".

📄 PDF Abstract BibTeX arXiv:2102.09270

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Stochastic Ratio Matching of RBMs for Sparse High-Dimensional Inputs

2013-12-01 · NeurIPS 2013 12 · Yann Dauphin, Yoshua Bengio

Sparse high-dimensional data vectors are common in many application domains where a very large number of rarely non-zero features can be devised. Unfortunately, this creates a computational bottleneck for unsupervised fe…

text-classificationText ClassificationVocal Bursts Intensity Prediction

From Variational to Deterministic Autoencoders

2019-03-29 · ICLR 2020 1 · Partha Ghosh, Mehdi S. M. Sajjadi, Antonio Vergari, Michael Black 외

Variational Autoencoders (VAEs) provide a theoretically-backed and popular framework for deep generative models. However, learning a VAE from data poses still unanswered theoretical questions and considerable practical c…

DecoderDensity Estimation

Robust Regression via Model Based Methods

2021-06-20 · Armin Moharrer, Khashayar Kamran, Edmund Yeh, Stratis Ioannidis

The mean squared error loss is widely used in many applications, including auto-encoders, multi-target regression, and matrix factorization, to name a few. Despite computational advantages due to its differentiability, i…

modelMulti-target regressionregression

Model Selection for Bayesian Autoencoders

2021-06-11 · NeurIPS 2021 12 · Ba-Hien Tran, Simone Rossi, Dimitrios Milios, Pietro Michiardi 외

We develop a novel method for carrying out model selection for Bayesian autoencoders (BAEs) by means of prior hyper-parameter optimization. Inspired by the common practice of type-II maximum likelihood optimization and i…

modelModel SelectionRepresentation Learning

Learning Generative Factors of EEG Data with Variational auto-encoders

2022-06-04 · Maksim Zhdanov, Saskia Steinmann, Nico Hoffmann

Electroencephalography produces high-dimensional, stochastic data from which it might be challenging to extract high-level knowledge about the phenomena of interest. We address this challenge by applying the framework of…

EEGElectroencephalogram (EEG)