The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks
A Research Paper on Neural Network Pruning
Book •
The lottery ticket hypothesis proposes that dense neural networks contain sparse subnetworks, or 'winning tickets,' that can be trained to achieve comparable accuracy to the full network.
This concept is explored through experiments demonstrating the effectiveness of these subnetworks when identified and trained independently.
The hypothesis has implications for neural network pruning and training efficiency.
This concept is explored through experiments demonstrating the effectiveness of these subnetworks when identified and trained independently.
The hypothesis has implications for neural network pruning and training efficiency.
Mentioned by
Mentioned in 0 episodes
Mentioned by 
Tim Scarfe
The Lottery Ticket Hypothesis with Jonathan Frankle
