Παρακαλώ χρησιμοποιήστε αυτό το αναγνωριστικό για να παραπέμψετε ή να δημιουργήσετε σύνδεσμο προς αυτό το τεκμήριο:
https://ruomo.lib.uom.gr/handle/7000/720
Πλήρης εγγραφή μεταδεδομένων
Πεδίο DC | Τιμή | Γλώσσα |
---|---|---|
dc.contributor.author | Nikolaidis, Spyridon | - |
dc.contributor.author | Refanidis, Ioannis | - |
dc.date.accessioned | 2020-05-21T16:06:43Z | - |
dc.date.available | 2020-05-21T16:06:43Z | - |
dc.date.issued | 2019-05-15 | - |
dc.identifier | 10.1007/978-3-030-20257-6_24 | en_US |
dc.identifier.isbn | 978-3-030-20256-9 | en_US |
dc.identifier.uri | https://doi.org/10.1007/978-3-030-20257-6_24 | en_US |
dc.identifier.uri | https://ruomo.lib.uom.gr/handle/7000/720 | - |
dc.description.abstract | Learnae is a framework proposal for decentralized training of Deep Neural Networks (DNN). The main priority of Learnae is to maintain a fully distributed architecture, where no participant has any kind of coordinating role. This solid peer-to-peer concept covers all aspects: Underlying network protocols, data acquiring/distribution and model training. The result is a resilient DNN training system with no single point of failure. Learnae focuses on use cases where infrastructure heterogeneity and network unreliability result to an always changing environment of commodity-hardware nodes. In order to achieve this level of decentralization, new technologies had to be utilized. The main pillars of this implementation are the ongoing projects of IPFS and IOTA. IPFS is a platform for a purely decentralized filesystem, where each node contributes local data storage. IOTA aims to be the networking infrastructure of the upcoming IoT reality. On top of these, we propose a management algorithm for training a DNN model collaboratively, by optimal exchange of data and model weights, always using distribution-friendly gossip protocols. | en_US |
dc.language.iso | en | en_US |
dc.publisher | Springer Nature Switzerland AG 2019 | en_US |
dc.relation.ispartofseries | Communications in Computer and Information Science book series (CCIS, volume 1000) | en_US |
dc.source | Proceedings of the 20th International Conference on Engineering Applications of Neural Networks (EANN-2019) | en_US |
dc.subject | FRASCATI::Natural sciences::Computer and information sciences | en_US |
dc.subject.other | Decentralized neural network training | en_US |
dc.subject.other | Distributed asynchronous stochastic gradient decent | en_US |
dc.subject.other | Model averaging | en_US |
dc.subject.other | Peer-to-Peer topologies | en_US |
dc.subject.other | Distributed Ledger Technology | en_US |
dc.subject.other | IPFS | en_US |
dc.subject.other | IOTA | en_US |
dc.title | LEARNAE: Distributed and Resilient Deep Neural Network Training for Heterogeneous Peer to Peer Topologies | en_US |
dc.type | Conference Paper | en_US |
dc.contributor.department | Τμήμα Εφαρμοσμένης Πληροφορικής | el |
local.identifier.firstpage | 286 | en_US |
local.identifier.lastpage | 298 | en_US |
local.identifier.eisbn | 978-3-030-20257-6 | en_US |
Εμφανίζεται στις Συλλογές: | Τμήμα Εφαρμοσμένης Πληροφορικής |
Αρχεία σε αυτό το Τεκμήριο:
Αρχείο | Περιγραφή | Μέγεθος | Μορφότυπος | |
---|---|---|---|---|
LEARNAE_FirstPaper.pdf | CAMERA READY | 514,15 kB | Adobe PDF | Προβολή/Ανοιγμα |
Τα τεκμήρια στο Αποθετήριο προστατεύονται από πνευματικά δικαιώματα, εκτός αν αναφέρεται κάτι διαφορετικό.