lottery-ticket-reading-list
The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks
The first lottery ticket hypothesis paper.
Stabilizing the Lottery Ticket Hypothesis
Applying lottery ticket to large-scale networks, with different tricks, e.g. weight rewinding.
Drawing early-bird tickets: Towards more efficient training of deep networks
Lottery ticket for efficient (hardware-wise) training. Tickets emerge at the early stage of the training process when large learning rates are used. Can be combined with channel-pruning/quantization.
Successfully Applying the Stabilized Lottery Ticket Hypothesis to the Transformer Architecture
Applying stabilized lottery ticket to Transformer, i.e. the "Stabilizing" paper.
Playing the lottery with rewards and multiple languages: lottery tickets in RL and NLP
Comparing Rewinding and Fine-tuning in Neural Network Pruning
The State of Sparsity in Deep Neural Networks
Lottery tickets (matching networks) don't exist at initialization.