If dropout limits trainable depth, does critical initialisation still matter? A large-scale statistical analysis on ReLU networks.
Arnu PretoriusElan Van BiljonBenjamin van NiekerkRyan EloffMatthew ReynardSteven D. JamesBenjamin RosmanHerman KamperSteve KroonPublished in: CoRR (2019)
Keyphrases
- statistical analysis
- small scale
- real life
- depth map
- fully automatic
- statistical analyses
- network structure
- clinical data
- network analysis
- computer networks
- network size
- heterogeneous networks
- network design
- community detection
- depth information
- real time
- sensor networks
- high resolution
- case study
- artificial intelligence