Joint Optimization of Dimension Reduction and Mixed-Precision Quantization for Activation Compression of Neural Networks.
Yu-Shan TaiCheng-Yang ChangChieh-Fang TengYi-Ta ChenAn-Yeu WuPublished in: IEEE Trans. Comput. Aided Des. Integr. Circuits Syst. (2023)
Keyphrases
- dimension reduction
- joint optimization
- neural network
- principal component analysis
- feature extraction
- high dimensional
- high dimensional data
- high dimensional problems
- sparse representation
- singular value decomposition
- feature space
- compression ratio
- linear discriminant analysis
- image compression
- feature selection
- high dimensionality
- low dimensional
- dimensionality reduction
- pattern recognition
- dimension reduction methods
- random projections
- cluster analysis
- unsupervised learning
- transform coding
- multiple description coding
- preprocessing
- manifold learning
- entropy coding
- face recognition
- nearest neighbor
- image processing
- signal processing
- sparse metric learning