LoCo: Local Contrastive Representation Learning

Abstract

Deep neural nets typically perform end-to-end backpropagation to learn the weights, a procedure that creates synchronization constraints in the weight update step across layers and is not biologically plausible. Recent advances in unsupervised contrastive representation learning point to the question of whether a learning algorithm can also be made local, that is, the updates of lower layers do not directly depend on the computation of upper layers. While Greedy InfoMax separately learns each block with a local objective, we found that it consistently hurts readout accuracy in state-of-the-art unsupervised contrastive learning algorithms, possibly due to the greedy objective as well as gradient isolation. In this work, we discover that by overlapping local blocks stacking on top of each other, we effectively increase the decoder depth and allow upper blocks to implicitly send feedbacks to lower blocks. This simple design closes the performance gap between local learning and end-to-end contrastive learning algorithms for the first time. Aside from standard ImageNet experiments, we also show results on complex downstream tasks such as object detection and instance segmentation directly using readout features.

arXiv commentNeurIPS 2020

References in corpus (19)

• A Simple Framework for Contrastive Learning of Visual Representations
• Representation Learning with Contrastive Predictive Coding
• Improved Baselines with Momentum Contrastive Learning
• SGDR: Stochastic Gradient Descent with Warm Restarts
• Momentum Contrast for Unsupervised Visual Representation Learning
• One weird trick for parallelizing convolutional neural networks
• MMDetection: Open MMLab Detection Toolbox and Benchmark
• Contrastive Multiview Coding
• Training Deep Nets with Sublinear Memory Cost
• Large Batch Training of Convolutional Networks
• Towards Biologically Plausible Deep Learning
• The Reversible Residual Network: Backpropagation Without Storing Activations
• Direct Feedback Alignment Provides Learning in Deep Neural Networks
• How Auto-Encoders Could Provide Credit Assignment in Deep Networks via Target Propagation
• Meta-Learning Update Rules for Unsupervised Representation Learning
• Putting An End to End-to-End: Gradient-Isolated Learning of Representations
• Backpropamine: training self-modifying neural networks with differentiable neuromodulated plasticity
• Greedy Layerwise Learning Can Scale to ImageNet
• Learning to Remember from a Multi-Task Teacher

Cited by in corpus (8)

• Contrastive Representation Learning: A Framework and Review
• Hard Negative Mixing for Contrastive Learning
• Efficient Self-supervised Vision Transformers for Representation Learning
• Local plasticity rules can learn deep representations using self-supervised contrastive predictions
• Parallel Training of Deep Networks with Local Updates
• Contrastive Learning of Global-Local Video Representations
• Meta-Learning Bidirectional Update Rules
• You Never Cluster Alone