Evaluating POWER Architecture for Distributed Training of Generative Adversarial Networks.
Ahmad HesamSofia VallecorsaGulrukh KhattakFederico CarminatiPublished in: ISC Workshops (2019)
Keyphrases
- distributed architecture
- multi agent
- computer networks
- loosely coupled
- wide area network
- hierarchical architecture
- heterogeneous environments
- generative model
- cooperative
- peer to peer networks
- layered architecture
- management system
- scalable distributed
- agent based architecture
- heterogeneous networks
- distributed systems
- multi tier
- multi agent architecture
- master slave
- training set
- highly distributed
- power consumption
- distributed processing
- social networks
- discriminative training
- recurrent networks
- structured peer to peer
- distributed object
- training process
- communication cost
- echo state networks
- grid enabled
- unsupervised learning
- peer to peer
- distributed environment
- complex networks
- network nodes
- multithreading
- heterogeneous systems
- local area network
- agent technology
- network structure
- packet switching
- distributed hash tables
- polynomial neural networks
- training samples