Improvements to Context Based Self-Supervised Learning

2017/11/30 by T. Nathan Mundhenk, Daniel E. Ho, Daniel Ho +1
Computer Science · Mathematics · #Advanced Neural Network Applications #Artificial intelligence #Computer science #Context (archaeology) #Discriminative model #Domain Adaptation and Few-Shot Learning #Generalization #Machine learning #Mathematics #Multimodal Machine Learning Applications #Pascal (unit) #Pattern recognition (psychology) #Skew #Software portability #cs.CV #cs.LG #cs.NE

paper · pdf · doi:10.1109/cvpr.2018.00973

Accepted paper at CVPR 2018

arxiv created 2018/03/28 · openalex publication_date 2018/06/01 · arxiv updated 2021/07/06 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/05

Abstract

We develop a set of methods to improve on the results of self-supervised learning using context. We start with a baseline of patch based arrangement context learning and go from there. Our methods address some overt problems such as chromatic aberration as well as other potential problems such as spatial skew and mid-level feature neglect. We prevent problems with testing generalization on common self-supervised benchmark tests by using different datasets during our development. The results of our methods combined yield top scores on all standard self-supervised benchmarks, including classification and detection on PASCAL VOC 2007, segmentation on PASCAL VOC 2012, and "linear tests" on the ImageNet and CSAIL Places datasets. We obtain an improvement over our baseline method of between 4.0 to 7.1 percentage points on transfer learning classification tests. We also show results on different standard network architectures to demonstrate generalization as well as portability. All data, models and programs are available at: https://gdo-datasci.llnl.gov/selfsupervised/.

Citations

• Unsupervised Visual Representation Learning by Context Prediction
• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• The Pascal Visual Object Classes (VOC) Challenge
• High-Performance Neural Networks for Visual Object Classification
• Multi-column deep neural networks for image classification
• Discriminative Unsupervised Feature Learning with Exemplar Convolutional Neural Networks
• ImageNet classification with deep convolutional neural networks
• Data-dependent Initializations of Convolutional Neural Networks
• Unsupervised Learning by Predicting Noise
• Multi-task Self-Supervised Visual Learning
• Deep Residual Learning for Image Recognition
• Colorization as a Proxy Task for Visual Understanding
• Caffe: Convolutional Architecture for Fast Feature Embedding
• Learning to See by Moving
• Unsupervised Learning of Visual Representations using Videos