18368790. COMMUNICATION-AWARE INFERENCE SERVING FOR PARTITIONED NEURAL NETWORKS (Cisco Technology, Inc.)
COMMUNICATION-AWARE INFERENCE SERVING FOR PARTITIONED NEURAL NETWORKS
Organization Name
Inventor(s)
Myungjin Lee of Bellevue WA US
Gustav Adrian Baumgart of Seattle WA US
Ramana Rao V.R. Kompella of Foster City CA US
COMMUNICATION-AWARE INFERENCE SERVING FOR PARTITIONED NEURAL NETWORKS
This abstract first appeared for US patent application 18368790 titled 'COMMUNICATION-AWARE INFERENCE SERVING FOR PARTITIONED NEURAL NETWORKS
Original Abstract Submitted
In one implementation, a device generates outputs of nodes in a upstream layer of a partitioned neural network. The device assigns priorities to each of the outputs of the nodes. The device selects, based on the priorities, a subset of the outputs to send to a remote device. The device sends, via a computer network, the subset of the outputs to the remote device for input to a downstream layer of the partitioned neural network.