Query Planner Agent for Routing Multimedia Query Types
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing multimedia systems face challenges in efficiently categorizing and responding to diverse user queries, including lexical, categorical, exploratory, and multistep queries, often relying on costly large language models for all query types, which can lead to inefficient resource utilization.
Innovation Solution
Implementing a query planner agent using a large language model to classify user queries and route them to appropriate modules (lexical, categorical, exploratory, or multistep query modules) based on query type, optimizing the use of costly LLMs for complex queries.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If large language models are used for all query types, then query response accuracy is improved, but resource utilization deteriorates due to costly LLM consumption on straightforward queries
Solution Approach 1:
The patent segments queries into different types (lexical, categorical, exploratory, multistep) and routes each type to the most appropriate processing module. Simple query types are handled by specialized modules without invoking LLMs, while complex query types are routed to LLM-based modules, thus maintaining accuracy for complex queries while reducing resource consumption for simple ones.
Solution Approach 2:
The patent introduces a query planning module as an intermediary between query reception and processing. This module classifies queries and determines the optimal processing path, acting as a mediator that decides whether LLM resources should be allocated to a particular query, thereby optimizing overall resource utilization while maintaining response quality.
2Device complexity
If a single processing module handles all queries, then device complexity is reduced, but query response time deteriorates due to inefficient resource allocation
Solution Approach 1:
The system is segmented into multiple specialized processing modules (lexical query module, categorical query module, exploratory query module, multistep query module) each optimized for specific query types. This segmentation enables parallel processing paths and reduces the time required for query handling by directing queries to the most efficient module rather than using a single general-purpose module.
Solution Approach 2:
The patent implements dynamic query routing where the processing path is determined based on the specific characteristics of each query. The system adaptively selects the appropriate module for each query type, creating a dynamic allocation of processing resources that optimizes response time while maintaining system manageability through centralized query planning.
Data Source
AI summary
A method is described and includes receiving by a query planner agent of a query response system a user query comprising a request for a response; determining a type of the received user query; identifying one of a plurality of response modules comprising the query response system based on the determined type of the received user query; and forwarding the received user query to the identified one of the plurality of response modules. In example embodiments, the determined type of the received user query comprises one of a lexical query, a categorical query, an explanatory query, and a multistep query, each of which is routed to a different response module.