Sensitivity-Aware Mixed-Precision Quantization and Width Optimization of Deep Neural Networks Through Cluster-Based Tree-Structured Parzen Estimation.
Seyedarmin AziziMahdi NazemiArash FayyaziMassoud PedramPublished in: CoRR (2023)
Keyphrases
- neural network
- density estimation
- pattern recognition
- optimization algorithm
- tree structure
- optimization problems
- non stationary
- optimization method
- constrained optimization
- fuzzy logic
- error estimation
- quantization error
- global optimization
- sensitivity analysis
- optimization process
- neural nets
- highly non linear
- successive approximation
- precision and recall
- self organizing maps
- knn
- artificial neural networks
- high precision
- back propagation
- estimation algorithm
- spanning tree
- tree structures
- evolutionary algorithm
- tree structured data
- real world