Title:
METHOD AND DEVICE FOR EXECUTION OF DEEP NEURAL NETWORK IN IOT EDGE NETWORK
Document Type and Number:
WIPO Patent Application WO/2022/203400
Kind Code:
A1
Abstract:
A method and device for execution of deep neural network (DNN) in an internet of things (IoT) edge network are provided. In an embodiment, at least one edge device within communication range of an IoT device are selected. Further, a network for connecting the IoT device with the at least one selected edge device is identified. A split ratio is determined based on an inference time of the DNN and a transmission time required for transmitting output of each layer of DNN from the IoT device to the selected at least one edge device. Finally, a plurality of layers of the DNN are split into a first part and a second part based on the split ratio, and the second part is transmitted to the selected at least one edge device through the identified network. The first part is executed on the IoT device, and the second part is executed on the selected at least one edge device.
More Like This:
Inventors:
KARJEE JYOTIRMOY (IN)
ANAND KARTIK (IN)
BHARGAV VANAMALA NARASIMHA (IN)
S PRAVEEN NAIK (IN)
DABBIRU RAMESH BABU VENKAT (IN)
N SRINIDHI (IN)
NIGAM ANSHUMAN (IN)
JHA RISHABH RAJ (IN)
ANAND KARTIK (IN)
BHARGAV VANAMALA NARASIMHA (IN)
S PRAVEEN NAIK (IN)
DABBIRU RAMESH BABU VENKAT (IN)
N SRINIDHI (IN)
NIGAM ANSHUMAN (IN)
JHA RISHABH RAJ (IN)
Application Number:
PCT/KR2022/004085
Publication Date:
September 29, 2022
Filing Date:
March 23, 2022
Export Citation:
Assignee:
SAMSUNG ELECTRONICS CO LTD (KR)
International Classes:
H04L67/60; G06N3/02; H04L67/12; H04L67/133; H04L67/289
Domestic Patent References:
| WO2018039238A1 | 2018-03-01 |
Foreign References:
| US20190362227A1 | 2019-11-28 | |||
| US20200311546A1 | 2020-10-01 | |||
| US20170337918A1 | 2017-11-23 | |||
| US20210065010A1 | 2021-03-04 |
Attorney, Agent or Firm:
LEE, Keon-Joo et al. (KR)
Download PDF: