Decoupled Neural Interfaces using Synthetic Gradients

by   Max Jaderberg, et al.

Training directed neural networks typically requires forward-propagating data through a computation graph, followed by backpropagating error signal, to produce weight updates. All layers, or more generally, modules, of the network are therefore locked, in the sense that they must wait for the remainder of the network to execute forwards and propagate error backwards before they can be updated. In this work we break this constraint by decoupling modules by introducing a model of the future computation of the network graph. These models predict what the result of the modelled subgraph will produce using only local information. In particular we focus on modelling error gradients: by using the modelled synthetic gradient in place of true backpropagated error gradients we decouple subgraphs, and can update them independently and asynchronously i.e. we realise decoupled neural interfaces. We show results for feed-forward models, where every layer is trained asynchronously, recurrent neural networks (RNNs) where predicting one's future gradient extends the time over which the RNN can effectively model, and also a hierarchical RNN system with ticking at different timescales. Finally, we demonstrate that in addition to predicting gradients, the same framework can be used to predict inputs, resulting in models which are decoupled in both the forward and backwards pass -- amounting to independent networks which co-learn such that they can be composed into a single functioning corporation.


page 1

page 2

page 3

page 4


Understanding Synthetic Gradients and Decoupled Neural Interfaces

When training neural networks, the use of Synthetic Gradients (SG) allow...

Feed-Forward On-Edge Fine-tuning Using Static Synthetic Gradient Modules

Training deep learning models on embedded devices is typically avoided s...

Benchmarking Decoupled Neural Interfaces with Synthetic Gradients

Artifical Neural Network are a particular class of learning system model...

Time-Adaptive Recurrent Neural Networks

Data are often sampled irregularly in time. Dealing with this using Recu...

Sample Complexity Bounds for Recurrent Neural Networks with Application to Combinatorial Graph Problems

Learning to predict solutions to real-valued combinatorial graph problem...

Learning Hierarchical Information Flow with Recurrent Neural Modules

We propose ThalNet, a deep learning model inspired by neocortical commun...

Long Timescale Credit Assignment in NeuralNetworks with External Memory

Credit assignment in traditional recurrent neural networks usually invol...

Code Repositories



view repo


Reference implementation of decoupled training with synthetic gradients.

view repo


Implements Decoupled Neural Interfaces using Synthetic Gradients

view repo


Decoupled Neural Interfaces with NNabla

view repo