CNN Filter DB: An Empirical Investigation of Trained Convolutional Filters

Abstract

Currently, many theoretical as well as practically relevant questions towards the transferability and robustness of Convolutional Neural Networks (CNNs) remain unsolved. While ongoing research efforts are engaging these problems from various angles, in most computer vision related cases these approaches can be generalized to investigations of the effects of distribution shifts in image data. In this context, we propose to study the shifts in the learned weights of trained CNN models. Here we focus on the properties of the distributions of dominantly used 3x3 convolution filter kernels. We collected and publicly provide a dataset with over 1.4 billion filters from hundreds of trained CNNs, using a wide range of datasets, architectures, and vision tasks. In a first use case of the proposed dataset, we can show highly relevant properties of many publicly available pre-trained models for practical applications: I) We analyze distribution shifts (or the lack thereof) between trained filters along different axes of meta-parameters, like visual category of the dataset, task, architecture, or layer depth. Based on these results, we conclude that model pre-training can succeed on arbitrary datasets if they meet size and variance conditions. II) We show that many pre-trained models contain degenerated filters which make them less robust and less suitable for fine-tuning on target applications. Data & Project website: https://github.com/paulgavrikov/cnn-filter-db

significantly reduced PDF size in v2; Accepted as ORAL at IEEE/CVF Conference on Computer Vision and Pattern Recognition 2022 (CVPR)

References in corpus (5)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Distilling the Knowledge in a Neural Network
• Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms
• Convolutional Neural Networks for Medical Image Analysis: Full Training or Fine Tuning?
• Measuring Robustness to Natural Distribution Shifts in Image Classification

Cited by in corpus (6)

• E Pluribus Unum Interpretable Convolutional Neural Networks
• Adversarial Robustness through the Lens of Convolutional Filters
• Convolutional Neural Networks Do Work with Pre-Defined Filters
• On the Privacy Risks of Deploying Recurrent Neural Networks in Machine Learning Models
• Revisiting Hidden Representations in Transfer Learning for Medical Imaging
• Enhancing Generalization in Convolutional Neural Networks through Regularization with Edge and Line Features