Compressing RNNs for IoT devices by 15-38x using Kronecker Products.
Urmish ThakkerJesse G. BeuDibakar GopeChu ZhouIgor FedorovGanesh DasikaMatthew MattinaPublished in: CoRR (2019)
Keyphrases
- recurrent neural networks
- management system
- mobile devices
- kronecker product
- smart objects
- cloud computing
- data compression
- embedded systems
- location information
- product information
- mobile applications
- feedforward neural networks
- spatial data
- input output
- smart home
- fuzzy neural network
- intelligent environments
- embedded devices
- echo state networks
- genetic algorithm