Deep Image: Scaling up Image Recognition

Abstract

We present a state-of-the-art image recognition system, Deep Image, developed using end-to-end deep learning. The key components are a custom-built supercomputer dedicated to deep learning, a highly optimized parallel algorithm using new strategies for data partitioning and communication, larger deep neural network models, novel data augmentation approaches, and usage of multi-scale high-resolution images. Our method achieves excellent results on multiple challenging computer vision benchmarks.

This paper has been withdrawn by the authors due to a mistake related to ImageNet server submissions

References in corpus (14)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• Fine-Grained Visual Classification of Aircraft
• Deep Big Simple Neural Nets Excel on Handwritten Digit Recognition
• One weird trick for parallelizing convolutional neural networks
• CNN Features off-the-shelf: an Astounding Baseline for Recognition
• Some Improvements on Deep Convolutional Neural Network Based Image Classification
• Attention for Fine-Grained Categorization
• Neural Activation Constellations: Unsupervised Part Model Discovery with Convolutional Networks
• Bilinear CNNs for Fine-grained Visual Recognition
• Multi-GPU Training of ConvNets
• Deep convolutional filter banks for texture recognition and segmentation
• Generalized Max Pooling
• The Application of Two-level Attention Models in Deep Convolutional Neural Network for Fine-grained Image Classification

Cited by in corpus (60)

• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• Improved Regularization of Convolutional Neural Networks with Cutout
• LSUN: Construction of a Large-scale Image Dataset using Deep Learning with Humans in the Loop
• Deep Learning with Limited Numerical Precision
• Probabilistic Backpropagation for Scalable Learning of Bayesian Neural Networks
• Acceleration of Deep Neural Network Training with Resistive Cross-Point Devices
• Action Recognition using Visual Attention
• Single Model Deep Learning on Imbalanced Small Datasets for Skin Lesion Classification
• Staleness-aware Async-SGD for Distributed Deep Learning
• Deep Learning in the Automotive Industry: Applications and Tools
• A Joint Convolutional Neural Networks and Context Transfer for Street Scenes Labeling
• Database Meets Deep Learning: Challenges and Opportunities
• Distributed Deep Learning Using Synchronous Stochastic Gradient Descent
• Further advantages of data augmentation on convolutional neural networks
• APAC: Augmented PAttern Classification with Neural Networks
• 8-Bit Approximations for Parallelism in Deep Learning
• Energy-Efficient ConvNets Through Approximate Computing
• SaliencyMix: A Saliency Guided Data Augmentation Strategy for Better Regularization
• GossipGraD: Scalable Deep Learning using Gossip Communication based Asynchronous Gradient Descent
• Analysis and Optimization of Convolutional Neural Network Architectures
• Lets keep it simple, Using simple architectures to outperform deeper and more complex architectures
• Hermite-Gaussian Mode Detection via Convolution Neural Networks
• Optimizing Network Performance for Distributed DNN Training on GPU Clusters: ImageNet/AlexNet Training in 1.5 Minutes
• Deep Neural Networks for Pattern Recognition
• Improving Auto-Augment via Augmentation-Wise Weight Sharing
• Parle: parallelizing stochastic gradient descent
• How useful is photo-realistic rendering for visual learning?
• Faster Asynchronous SGD
• Exploring the Design Space of Deep Convolutional Neural Networks at Large Scale
• Automated detection of oral pre-cancerous tongue lesions using deep learning for early diagnosis of oral cavity cancer
• Deep Action- and Context-Aware Sequence Learning for Activity Recognition and Anticipation
• A Practical Guide to CNNs and Fisher Vectors for Image Instance Retrieval
• Is Saki #delicious? The Food Perception Gap on Instagram and Its Relation to Health
• Social Media Image Analysis for Public Health
• Online Hyper-parameter Learning for Auto-Augmentation Strategy
• Fast Semantic Image Segmentation with High Order Context and Guided Filtering
• Low-cost and high-performance data augmentation for deep-learning-based skin lesion classification
• Targeted Nonlinear Adversarial Perturbations in Images and Videos
• Learning Fine-grained Features via a CNN Tree for Large-scale Classification
• Data augmentation with Symbolic-to-Real Image Translation GANs for Traffic Sign Recognition
• dMath: A Scalable Linear Algebra and Math Library for Heterogeneous GP-GPU Architectures
• Can a CNN Recognize Catalan Diet?
• MGSim + MGMark: A Framework for Multi-GPU System Research
• Stanza: Layer Separation for Distributed Training in Deep Learning
• Learning from Learning Machines: Optimisation, Rules, and Social Norms
• The Penalty Imposed by Ablated Data Augmentation
• Context Augmentation for Convolutional Neural Networks
• A Survey on Large-scale Machine Learning
• Explainability Guided Multi-Site COVID-19 CT Classification
• Recent Advances in Convolutional Neural Network Acceleration
• Learning scale-variant and scale-invariant features for deep image classification
• What augmentations are sensitive to hyper-parameters and why?
• Cross-Domain Image Classification through Neural-Style Transfer Data Augmentation
• Recognition from Hand Cameras
• Efficient Multi-Person Pose Estimation with Provable Guarantees
• On teaching mathematics to gifted students: some enrichment ideas and educational activities
• Hardware Synthesis of State-Space Equations; Application to FPGA Implementation of Shallow and Deep Neural Networks
• ScaleNet: An Unsupervised Representation Learning Method for Limited Information
• Fine-Grained Visual Categorization using Meta-Learning Optimization with Sample Selection of Auxiliary Data
• Deep Sequence Learning for Video Anticipation: From Discrete and Deterministic to Continuous and Stochastic