Adaptive Neural Network Selection for Distributed Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing distributed client-server processing systems face challenges in balancing processing tasks between client and server devices due to network quality and resource constraints, leading to data loss and increased latency, particularly in intensive tasks like speech recognition and image processing.
Innovation Solution
The implementation of adaptive artificial neural network (ANN) selection techniques, where client computing devices determine operating parameters to select appropriate ANNs that distribute processing tasks between client and server devices, optimizing computation, communication, and latency by using bottleneck layers for efficient data transmission.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If the client computing device handles a greater amount of processing task due to low network quality, then network bandwidth requirements are reduced, but processing accuracy and capability deteriorate
Solution Approach 1:
The processing task is segmented into multiple portions distributed between client and server devices. The client device performs initial processing locally to reduce network bandwidth requirements, while the server device performs additional processing to maintain accuracy. This segmentation allows the system to adapt to varying network conditions while preserving processing quality.
2Loss of energy
If compression is applied to data transferred to the server computing device, then network bandwidth requirements are reduced, but data loss occurs which negatively affects processing results
Solution Approach 1:
The client computing device performs preliminary processing of the data before transmission to the server. This preliminary action extracts and processes essential features locally, reducing the amount of data that needs to be compressed and transmitted over the network, thereby minimizing both bandwidth consumption and data loss.
3Power
If the server computing device handles intensive processing tasks, then processing capability is improved, but network latency increases
Solution Approach 1:
The processing task is divided into segments that can be executed in parallel across client and server devices. The client device performs initial processing locally without waiting for server response, reducing latency. The server device handles intensive processing of segmented data, maintaining high processing capability while minimizing network round-trip time.
4Loss of energy
If more processing is performed at the client computing device, then network bandwidth and latency are reduced, but device energy consumption increases
Solution Approach 1:
The system dynamically changes the distribution parameter of processing tasks between client and server devices based on real-time conditions such as network quality, device battery status, and processing requirements. This parameter adjustment allows optimization of energy consumption by performing more processing locally when beneficial and offloading to server when appropriate.
Data Source
AI summary
Computer-implemented techniques can include obtaining, by a client computing device, a digital media item and a request for a processing task on the digital item and determining a set of operating parameters based on (i) available computing resources at the client computing device and (ii) a condition of a network. Based on the set of operating parameters, the client computing device or a server computing device can select one of a plurality of artificial neural networks (ANNs), each ANN defining which portions of the processing task are to be performed by the client and server computing devices. The client and server computing devices can coordinate processing of the processing task according to the selected ANN. The client computing device can also obtain final processing results corresponding to a final evaluation of the processing task and generate an output based on the final processing results.