Phase transitions in the mini-batch size for sparse and dense neural networks.
Raffaele MarinoFederico Ricci-TersenghiPublished in: CoRR (2023)
Keyphrases
- phase transition
- batch size
- neural network
- random constraint satisfaction problems
- batch mode
- constraint satisfaction
- poisson process
- single item
- batch processing
- satisfiability problem
- np complete
- randomly generated
- random instances
- upper bound
- cellular automata
- finite horizon
- training process
- online learning
- active learning