SIGMA: A Sparse and Irregular GEMM Accelerator with Flexible Interconnects for DNN Training.
Eric QinAnanda SamajdarHyoukjun KwonVineet NadellaSudarshan SrinivasanDipankar DasBharat KaulTushar KrishnaPublished in: HPCA (2020)
Keyphrases
- training process
- high dimensional
- avoid overfitting
- parallel implementation
- neural network
- sparse data
- training algorithm
- training data
- test set
- input output
- power consumption
- training phase
- data mining
- power dissipation
- training set
- case study
- training examples
- sparse representation
- feature selection
- information systems
- machine learning
- data sets