What does fault tolerant deep learning need from MPI?
Vinay AmatyaAbhinav VishnuCharles SiegelJeff DailyPublished in: EuroMPI/USA (2017)
Keyphrases
- deep learning
- fault tolerant
- fault tolerance
- distributed systems
- message passing
- parallel algorithm
- unsupervised learning
- parallel implementation
- unsupervised feature learning
- shared memory
- load balancing
- mental models
- parallel computing
- general purpose
- weakly supervised
- deep architectures
- massively parallel
- machine learning
- digital libraries
- learning algorithm
- viewpoint
- feature selection