$\P$ILCRO: Making Importance Landscapes Flat Again

• URL: http://arxiv.org/abs/2001.09696v2
• Date: Thu, 6 Feb 2020 11:41:02 GMT
• Title: $\P$ILCRO: Making Importance Landscapes Flat Again
• Authors: Vincent Moens, Simiao Yu, Gholamreza Salimi-Khorshidi
• Abstract summary: This paper shows that most of the existing convolutional architectures define, at initialisation, a specific feature importance landscape. We derive the P-objective, or PILCRO for Pixel-wise Landscape Curvature Regularised Objective. We show that P-regularised versions of popular computer vision networks have a flat importance landscape, train faster, result in a better accuracy and are more robust to noise at test time.
• Score: 7.047473967702792
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Convolutional neural networks have had a great success in numerous tasks, including image classification, object detection, sequence modelling, and many more. It is generally assumed that such neural networks are translation invariant, meaning that they can detect a given feature independent of its location in the input image. While this is true for simple cases, where networks are composed of a restricted number of layer classes and where images are fairly simple, complex images with common state-of-the-art networks do not usually enjoy this property as one might hope. This paper shows that most of the existing convolutional architectures define, at initialisation, a specific feature importance landscape that conditions their capacity to attend to different locations of the images later during training or even at test time. We demonstrate how this phenomenon occurs under specific conditions and how it can be adjusted under some assumptions. We derive the P-objective, or PILCRO for Pixel-wise Importance Landscape Curvature Regularised Objective, a simple regularisation technique that favours weight configurations that produce smooth, low-curvature importance landscapes that are conditioned on the data and not on the chosen architecture. Through extensive experiments, we further show that P-regularised versions of popular computer vision networks have a flat importance landscape, train faster, result in a better accuracy and are more robust to noise at test time, when compared to their original counterparts in common computer-vision classification settings.

Related papers

• The Overfocusing Bias of Convolutional Neural Networks: A Saliency-Guided Regularization Approach [11.524573224123905]
  CNNs make decisions based on narrow, specific regions of input images. This behavior can severely compromise the model's generalization capabilities. We introduce Saliency Guided Dropout (SGDrop) to address this specific issue.
  arXiv  Detail & Related papers  (2024-09-25T21:30:16Z)
• Rotation Equivariant Proximal Operator for Deep Unfolding Methods in Image Restoration [68.18203605110719]
  We propose a high-accuracy rotation equivariant proximal network that embeds rotation symmetry priors into the deep unfolding framework. This study makes efforts to suggest a high-accuracy rotation equivariant proximal network that effectively embeds rotation symmetry priors into the deep unfolding framework.
  arXiv  Detail & Related papers  (2023-12-25T11:53:06Z)
• Distance Weighted Trans Network for Image Completion [52.318730994423106]
  We propose a new architecture that relies on Distance-based Weighted Transformer (DWT) to better understand the relationships between an image's components. CNNs are used to augment the local texture information of coarse priors. DWT blocks are used to recover certain coarse textures and coherent visual structures.
  arXiv  Detail & Related papers  (2023-10-11T12:46:11Z)
• Multilayer deep feature extraction for visual texture recognition [0.0]
  This paper is focused on improving the accuracy of convolutional neural networks in texture classification. It is done by extracting features from multiple convolutional layers of a pretrained neural network and aggregating such features using Fisher vector. We verify the effectiveness of our method on texture classification of benchmark datasets, as well as on a practical task of Brazilian plant species identification.
  arXiv  Detail & Related papers  (2022-08-22T03:53:43Z)
• Spatially-Adaptive Image Restoration using Distortion-Guided Networks [51.89245800461537]
  We present a learning-based solution for restoring images suffering from spatially-varying degradations. We propose SPAIR, a network design that harnesses distortion-localization information and dynamically adjusts to difficult regions in the image.
  arXiv  Detail & Related papers  (2021-08-19T11:02:25Z)
• Cross-Descriptor Visual Localization and Mapping [81.16435356103133]
  Visual localization and mapping is the key technology underlying the majority of Mixed Reality and robotics systems. We present three novel scenarios for localization and mapping which require the continuous update of feature representations. Our data-driven approach is agnostic to the feature descriptor type, has low computational requirements, and scales linearly with the number of description algorithms.
  arXiv  Detail & Related papers  (2020-12-02T18:19:51Z)
• An Effective Anti-Aliasing Approach for Residual Networks [27.962502376542588]
  Frequency aliasing is a phenomenon that may occur when sub-sampling any signal, such as an image or feature map, causing distortion in the sub-sampled output. We show that we can mitigate this effect by placing non-trainable blur filters and using smooth activation functions at key locations. These simple architectural changes lead to substantial improvements in out-of-distribution generalization on both image classification under natural corruptions on ImageNet-C and few-shot learning on Meta-Dataset.
  arXiv  Detail & Related papers  (2020-11-20T22:55:57Z)
• Lightweight Modules for Efficient Deep Learning based Image Restoration [20.701733377216932]
  We propose several lightweight low-level modules which can be used to create a computationally low cost variant of a given baseline model. Our results show that proposed networks consistently output visually similar reconstructions compared to full capacity baselines.
  arXiv  Detail & Related papers  (2020-07-11T19:35:00Z)
• Steering Self-Supervised Feature Learning Beyond Local Pixel Statistics [60.92229707497999]
  We introduce a novel principle for self-supervised feature learning based on the discrimination of specific transformations of an image. We demonstrate experimentally that learning to discriminate transformations such as LCI, image warping and rotations, yields features with state of the art generalization capabilities.
  arXiv  Detail & Related papers  (2020-04-05T22:09:08Z)
• Two-shot Spatially-varying BRDF and Shape Estimation [89.29020624201708]
  We propose a novel deep learning architecture with a stage-wise estimation of shape and SVBRDF. We create a large-scale synthetic training dataset with domain-randomized geometry and realistic materials. Experiments on both synthetic and real-world datasets show that our network trained on a synthetic dataset can generalize well to real-world images.
  arXiv  Detail & Related papers  (2020-04-01T12:56:13Z)
• Contextual Encoder-Decoder Network for Visual Saliency Prediction [42.047816176307066]
  We propose an approach based on a convolutional neural network pre-trained on a large-scale image classification task. We combine the resulting representations with global scene information for accurately predicting visual saliency. Compared to state of the art approaches, the network is based on a lightweight image classification backbone.
  arXiv  Detail & Related papers  (2019-02-18T16:15:25Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.