Provable Benefits of Overparameterization in Model Compression: From Double Descent to Pruning Neural Networks.
Xiangyu ChangYingcong LiSamet OymakChristos ThrampoulidisPublished in: CoRR (2020)
Keyphrases
- neural network
- probabilistic model
- statistical model
- data sets
- computational model
- high level
- mathematical model
- experimental data
- search space
- decision trees
- control system
- input data
- probability distribution
- prediction model
- greedy algorithm
- management system
- prior knowledge
- artificial neural networks
- expert systems
- multiscale