Path Sample-Analytic Gradient Estimators for Stochastic Binary Networks

Abstract

In neural networks with binary activations and or binary weights the training by gradient descent is complicated as the model has piecewise constant response. We consider stochastic binary networks, obtained by adding noises in front of activations. The expected model response becomes a smooth function of parameters, its gradient is well defined but it is challenging to estimate it accurately. We propose a new method for this estimation problem combining sampling and analytic approximation steps. The method has a significantly reduced variance at the price of a small bias which gives a very practical tradeoff in comparison with existing unbiased and biased estimators. We further show that one extra linearization step leads to a deep straight-through estimator previously known only as an ad-hoc heuristic. We experimentally show higher accuracy in gradient estimation and demonstrate a more stable and better performing training in deep convolutional models with both proposed methods.

arXiv commentNeurIPS 2020

References in corpus (18)

• PyTorch: An Imperative Style, High-Performance Deep Learning Library
• Categorical Reparameterization with Gumbel-Softmax
• Binarized Neural Networks: Training Deep Neural Networks with Weights and Activations Constrained to +1 or -1
• Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computation
• DoReFa-Net: Training Low Bitwidth Convolutional Neural Networks with Low Bitwidth Gradients
• Expectation Propagation for approximate Bayesian inference
• The Concrete Distribution: A Continuous Relaxation of Discrete Random Variables
• Binarized Neural Networks
• Neural Variational Inference and Learning in Belief Networks
• REBAR: Low-variance, unbiased gradient estimates for discrete latent variable models
• Understanding Straight-Through Estimator in Training Activation Quantized Neural Nets
• Probabilistic Binary Neural Networks
• Techniques for Learning Binary Stochastic Feedforward Neural Networks
• Training Competitive Binary Neural Networks from Scratch
• ARM: Augment-REINFORCE-Merge Gradient for Stochastic Binary Networks
• Normalization of Neural Networks using Analytic Variance Propagation
• Straight-Through Estimator as Projected Wasserstein Gradient Flow
• GO Gradient for Expectation-Based Objectives