# The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks

## TL;DR

This paper introduces the lottery ticket hypothesis, showing that sparse subnetworks within dense neural networks, identified by their initial weights, can be trained independently to achieve comparable or better performance, revealing the importance of initializations.

## Contribution

The paper proposes the lottery ticket hypothesis and an algorithm to find winning subnetworks, demonstrating their effectiveness in training neural networks more efficiently.

## Key findings

- Winning tickets are less than 20% of the original network size.
- Winning tickets learn faster and reach higher accuracy.
- Subnetworks with favorable initializations can be trained independently.

## Abstract

Neural network pruning techniques can reduce the parameter counts of trained networks by over 90%, decreasing storage requirements and improving computational performance of inference without compromising accuracy. However, contemporary experience is that the sparse architectures produced by pruning are difficult to train from the start, which would similarly improve training performance. We find that a standard pruning technique naturally uncovers subnetworks whose initializations made them capable of training effectively. Based on these results, we articulate the "lottery ticket hypothesis:" dense, randomly-initialized, feed-forward networks contain subnetworks ("winning tickets") that - when trained in isolation - reach test accuracy comparable to the original network in a similar number of iterations. The winning tickets we find have won the initialization lottery: their connections have initial weights that make training particularly effective. We present an algorithm to identify winning tickets and a series of experiments that support the lottery ticket hypothesis and the importance of these fortuitous initializations. We consistently find winning tickets that are less than 10-20% of the size of several fully-connected and convolutional feed-forward architectures for MNIST and CIFAR10. Above this size, the winning tickets that we find learn faster than the original network and reach higher test accuracy.