Campfire: Compressible, Regularization-Free, Structured Sparse Training for Hardware Accelerators.
Noah GamboaKais KudrolliAnand DhootArdavan PedramPublished in: CoRR (2020)
Keyphrases
- compressive sensing
- sparsity regularization
- early stopping
- structured sparsity
- mixed norm
- real world
- high dimensional
- signal processing
- training examples
- test set
- elastic net
- sparse data
- training set
- avoid overfitting
- sparse approximation
- sparse coding
- sparse representation
- supervised learning
- stochastic gradient descent
- regularization methods
- training algorithm
- training process
- training samples
- model selection