Information Processing Reusing Existing Responses to Reduce Computation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional generative AI technologies require significant processing costs to generate responses to prompts due to the need for continuous computation and data generation.
Innovation Solution
An information processing apparatus that includes a reception unit, determination unit, and providing unit to determine if a response to a new prompt is available using existing information, and provides existing information as a response when possible, reducing the need for continuous computation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If generative AI is used to generate response information for each prompt, then the response quality and accuracy are improved, but the processing cost increases significantly
Solution Approach 1:
The system pre-generates multiple candidate response information items and stores them in advance before they are needed. When a prompt is received, the system retrieves pre-generated responses that match the prompt conditions without performing new generative AI computations, thereby maintaining response accuracy while eliminating the high processing costs associated with real-time generation.
Solution Approach 2:
Instead of generating unique response information for each prompt through computationally expensive generative AI processes, the system creates and stores multiple copies of candidate response information that can be reused across different prompts. This copying approach preserves response quality while dramatically reducing the need for repeated high-cost generation operations.
2Speed
If generative AI continuously generates new responses, then the responsiveness to new prompts is improved, but the processing time and computational resources increase
Solution Approach 1:
Candidate response information is generated and stored in advance before actual prompts are processed. This preliminary generation eliminates the need for time-consuming comput ation during query processing, enabling instant retrieval and significantly improving response speed while reducing processing time losses.
Solution Approach 2:
The system skips the time-consuming generative AI computation step by using pre-generated candidate responses. Instead of rushing through the generation process for each prompt, the system directly retrieves appropriate responses from stored candidates, effectively skipping the computational bottleneck and achieving rapid response delivery.
3Use of energy by moving object
If existing information is used to provide responses, then the processing cost is reduced, but the applicability to new prompts may be limited
Solution Approach 1:
The system generates and stores multiple candidate response information items that can serve multiple different prompts. Each candidate response is designed to be universally applicable to various prompt conditions, allowing the same response to answer multiple different questions or requests, thereby enhancing prompt applicability while maintaining cost efficiency through reuse of existing information.
Solution Approach 2:
The system varies parameters of candidate response information such as the number of candidates, diversity of response types, and matching criteria to improve adaptability. By adjusting these parameters, the system can better match existing responses to new prompts with different requirements, expanding the applicability of stored information to a broader range of prompt scenarios.
Data Source
AI summary
An information processing apparatus according to the present application includes a reception unit, a determination unit, and a providing unit. The reception unit receives a query including a new prompt sent from a user or a query for obtaining the new prompt. The determination unit determines whether or not a response to the new prompt is available by using existing information that is already-existing information. The providing unit provides, when it is determined by the determination unit that the response to the new prompt is available by using the existing information, the existing information to the user as response information indicating the response to the new prompt.