Description
A Masked Convolution is a type of convolution which masks certain pixels so that the model can only predict based on pixels already seen. This type of convolution was introduced with PixelRNN generative models, where an image is generated pixel by pixel, to ensure that the model was conditional only on pixels already visited.
Papers Using This Method
$\spadesuit$ SPADE $\spadesuit$ Split Peak Attention DEcomposition2024-11-06Obtaining Optimal Spiking Neural Network in Sequence Learning via CRNN-SNN Conversion2024-08-18Ensemble learning for predictive uncertainty estimation with application to the correction of satellite precipitation products2024-03-14Density Matters: Improved Core-set for Active Domain Adaptive Segmentation2023-12-15Uncertainty estimation of machine learning spatial precipitation predictions from satellite data2023-11-13Density Matrix Emulation of Quantum Recurrent Neural Networks for Multivariate Time Series Prediction2023-10-31Efficient quantum recurrent reinforcement learning via quantum reservoir computing2023-09-13PixelRNN: In-pixel Recurrent Neural Networks for End-to-end-optimized Perception with Neural Sensors2023-04-11Quantum Recurrent Neural Networks for Sequential Learning2023-02-07Time-Warping Invariant Quantum Recurrent Neural Networks via Quantum-Classical Adaptive Gating2023-01-19Masked Autoencoders Are Stronger Knowledge Distillers2023-01-01AdaMAE: Adaptive Masking for Efficient Spatiotemporal Learning with Masked Autoencoders2022-11-16Reservoir Computing via Quantum Recurrent Neural Networks2022-11-04Rapid training of quantum recurrent neural networks2022-07-01ConvMAE: Masked Convolution Meets Masked Autoencoders2022-05-08Image Inpainting with Edge-guided Learnable Bidirectional Attention Maps2021-04-25Non-Autoregressive Predictive Coding for Learning Speech Representations from Local Dependencies2020-11-01Cyber Threat Intelligence for Secure Smart City2020-07-26Recurrent Quantum Neural Networks2020-06-25A Formal Hierarchy of RNN Architectures2020-04-18