Multi-Agent Task Processing Under Local Compute Limits
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The processing power of locally deployed data processing models is limited by the number of parameters, which exceeds the computing resources of electronic devices, leading to insufficient capacity to meet user needs.
Innovation Solution
A method involving multiple intelligent agents with different processing capabilities are used to process a target task, generating initial results, which are then processed in a first or second mode based on the number of different results to obtain a final result, combining their capabilities without increasing model parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Power
If the number of model parameters is increased to enhance processing power, then the processing capacity is improved, but the computing resources consumed by the electronic device increase
Solution Approach 1:
The patent divides the task processing into multiple stages with different intelligent agents having specialized capabilities. Instead of using a single large model with all parameters, the system segments the processing function across multiple agents (e.g., retrieval agent, generation agent, verification agent), each handling specific sub-tasks. This segmentation allows the system to achieve high processing capacity without requiring all parameters to be loaded into the device simultaneously.
Solution Approach 2:
The patent introduces a task scheduling agent as an intermediary that coordinates between different intelligent agents. This mediator manages the workflow, selects appropriate agents based on task requirements, and orchestrates the processing stages. The intermediary enables efficient resource utilization by dynamically assigning tasks to available agents without requiring all processing capabilities to be present in the device at once.
2Power
If a single large data processing model is deployed locally to provide strong processing capacity, then the processing capability is improved, but the device complexity increases due to resource limitations
Solution Approach 1:
The patent implements a universal task scheduling agent that can coordinate multiple types of intelligent agents (retrieval agents, generation agents, verification agents, etc.). This multi-functional scheduler handles diverse task types by dynamically selecting and coordinating appropriate agents, reducing the need for device-specific optimization for each task type. The system achieves high processing capacity through agent composition rather than requiring a single monolithic model.
Solution Approach 2:
The patent creates a dynamic agent selection and coordination mechanism where the task scheduling agent adapts its behavior based on task requirements. Instead of a static model configuration, the system dynamically assembles different combinations of intelligent agents according to the specific task at hand. This dynamic approach allows the system to optimize resource usage and reduce device complexity by only loading and coordinating necessary agents for each task.
3Measurement precision
If multiple intelligent agents with different processing capabilities are used to process tasks, then the processing accuracy is improved through capability combination, but the system complexity increases
Solution Approach 1:
The patent incorporates feedback mechanisms where verification agents review the outputs of generation agents, and the task scheduling agent monitors overall task progress. This feedback loop allows the system to correct errors and improve accuracy by having specialized verification agents check the work of generation agents. The feedback mechanism manages complexity by providing structured correction channels rather than requiring complete redesign of the processing architecture.
Solution Approach 2:
The patent merges the capabilities of multiple intelligent agents through coordinated processing stages. The task scheduling agent combines the outputs of retrieval agents, generation agents, and verification agents into a final integrated result. This merging process achieves high processing accuracy by synthesizing information from multiple specialized agents while managing complexity through a unified coordination framework that handles the integration systematically.
Data Source
AI summary
A task processing method includes obtaining a target task, determining a plurality of intelligent agents with different processing capabilities based on the target task, using the plurality of intelligent agents to process the target task to obtain a plurality of first results, each intelligent agent outputting a first result, and processing the plurality of first results in a first processing mode or a second processing mode based on a number of different first results to obtain a final result of the target task.