User Equipment Neural Network Data Packet Descriptor
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional wireless communication methods are not designed to efficiently transmit neural network coefficients or partial computational results, limiting the effectiveness of distributed neural network processing across networking nodes.
Innovation Solution
A user equipment (UE) with a processor and transmitter that generates and transmits neural network computation results in data packets, including descriptors with parameters and settings, using specific protocols and Quality of Service (QoS) indicators to ensure efficient communication with base stations, mapping neural network communication QoS Flow to data radio bearers, and configuring communication sessions for optimal performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional wireless communication methods are used to transmit neural network coefficients or partial computational results, then the transmission can be performed using existing infrastructure, but the transmission efficiency and delivery performance are insufficient for distributed neural network processing
Solution Approach 1:
The patent introduces specialized parameters and settings in the data packet descriptor to characterize neural network computation results, including neural network type, number of layers, size of computation results, level of computation results, sequence number, and time stamp. These parameter changes enable the wireless communication system to efficiently handle and prioritize neural network data transmissions.
Solution Approach 2:
The patent introduces a specialized data packet structure with a descriptor as an intermediary element between the neural network computation results and the wireless transmission channel. The descriptor contains specific parameters that mediate the transmission process, enabling the base station to properly route and prioritize the neural network data.
2Measurement precision
If neural network computation results are transmitted with detailed descriptors and parameters, then the delivery performance and processing accuracy are improved, but the overhead and complexity of the communication protocol increase
Solution Approach 1:
The patent segments the data packet into distinct components: a packet header with an indicator field, a descriptor with specific parameters, and a data payload containing the neural network computation results. This segmentation allows each component to serve its specific function while maintaining overall system efficiency.
Solution Approach 2:
The patent designs a universal data packet structure that can carry various types of neural network computation results through the use of a descriptor with multiple parameters. The same packet structure can accommodate different neural network types, layers, and computation levels, making the protocol versatile without requiring multiple specialized formats.
3Loss of time
If Quality of Service indicators are configured for neural network computation results, then the delivery performance and latency are improved, but the configuration complexity and signaling overhead increase
Solution Approach 1:
The patent configures Quality of Service parameters and data radio bearer mappings in advance through RRC signaling before the actual neural network computation results need to be transmitted. This preliminary configuration ensures that when data needs to be transmitted, the appropriate QoS parameters and bearers are already in place, reducing latency without requiring complex real-time configuration.
Data Source
AI summary
A user equipment includes a processor and a transmitter. The processor performs a neural network computation to generate neural network computation results. The neural network computation results are intermediate data of the neural network computation. The intermediate data are the neural network computation results of computation nodes in partial layers of the neural network computation. The transmitter transmits a data packet to a base station to perform computation of computation nodes in remaining layers of the neural network computation. The data packet includes the neural network computation results, a packet header, and a descriptor. The descriptor includes parameters and settings of the neural network computation results. The parameters and settings include at least two of a neural network type, number of layers in the neural network, a size of the neural network computation results, level of the neural network computation results, a sequence number, and a time stamp.