Hybrid Client-Server Input Prediction Latency Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Mobile devices face challenges in providing accurate predictive text due to limited computational resources and insufficient network bandwidth, leading to delays in displaying predictive text entries.
Innovation Solution
A mobile device determines the expected user input interval and response latency, deciding whether to perform predictive text entry locally or remotely based on these factors, utilizing a local cache or a remote server for generating predictive text.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If predictive text entry is performed using a large amount of data and computation, then prediction accuracy is improved, but device power consumption and processing burden increase
Solution Approach 1:
The patent segments the predictive text processing into two parts: a lightweight local model that runs continuously on the mobile device for immediate predictions, and a more sophisticated remote model that processes complex queries when needed. This segmentation allows the system to maintain high prediction accuracy while managing power consumption by only activating the resource-intensive remote processing when necessary.
Solution Approach 2:
The patent introduces a server as an intermediary between the mobile device and the predictive text data. The server hosts the computationally intensive prediction models and data, allowing the mobile device to offload heavy processing tasks. This intermediary approach enables accurate predictions without requiring the mobile device to have sufficient computational power or energy reserves.
2Measurement precision
If predictive data is received from remote servers, then prediction accuracy is improved, but network bandwidth requirements increase
Solution Approach 1:
The patent extracts only the essential predictive text data and model parameters from the server, transmitting minimal data packets over the network. Rather than transferring large datasets, the system extracts and transmits only the necessary prediction results or updated model increments, significantly reducing network bandwidth requirements while maintaining prediction accuracy.
3Measurement precision
If predictive data is received from remote servers, then prediction accuracy is improved, but response latency increases
Solution Approach 1:
The patent implements preliminary action by pre-loading prediction models and commonly used predictive data into the local cache on the mobile device during off-peak times or when network conditions are favorable. This allows the device to quickly access prediction data without real-time network delays, reducing response latency while maintaining accuracy through the use of pre-fetched data.
Solution Approach 2:
The patent employs a dynamic approach where the system adaptively switches between local and remote prediction processing based on real-time conditions such as network availability, input complexity, and cache status. This dynamic strategy optimizes the balance between prediction accuracy and response latency by using local processing for immediate responses and remote processing for enhanced accuracy when time permits.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present disclosure includes methods, apparatus, and computer-readable media for receiving one or more user inputs from a user via a user interface, transmitting the one or more user inputs to a remote server, obtaining an expected user input interval and an expected response latency, and displaying, via the user interface, either one or more predicted inputs associated with a content of a local cache or one or more predicted inputs received from the remote server, based at least on whether the expected user input interval is greater than a product of the expected response latency and a factor.