Weightless: Lossy weight encoding for deep neural network compression.
Brandon ReagenUdit GuptaRobert AdolfMichael MitzenmacherAlexander M. RushGu-Yeon WeiDavid BrooksPublished in: ICLR (Workshop) (2018)
Keyphrases
- neural network
- data compression
- run length encoding
- arithmetic coding
- lossless coding
- compression scheme
- high compression ratio
- lossy compression
- image compression
- compression ratio
- lossless compression
- compression algorithm
- predictive coding
- wavelet compression
- run length
- back propagation
- weight update
- compressed data
- genetic algorithm
- image quality
- artificial neural networks
- coding method
- fractal image compression
- variable length
- fault diagnosis
- image coding algorithm
- efficient compression
- compression rate
- complexity reduction
- image coder
- wavelet transform
- neural network model
- feed forward
- chain code
- self organizing maps
- image coding
- coding scheme
- high compression
- recurrent neural networks
- activation function
- low memory
- lossless image compression
- low bit rate
- source coding
- neural network is trained
- lempel ziv
- transform coding
- entropy coding
- compressed pattern matching