External Entity-Summary Database for Accurate, Fresh LLM Responses
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current generative models, such as large language models (LLMs), often generate inaccurate or hallucinated responses due to outdated training data, leading to decreased user trust, elongated interactions, and increased resource consumption.
Innovation Solution
An external entity-summarization database is built to store summarized descriptions of entities, which are generated using a first generative model, allowing LLMs to leverage up-to-date information through in-context learning, reducing the risk of hallucinations and resource consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If LLMs are trained on outdated training data, then the model can be trained efficiently with available data, but the generated responses become inaccurate and hallucinated
Solution Approach 1:
The system performs preliminary actions by collecting and summarizing current information about entities from external data sources before the LLM generates responses. This pre-collected information is stored in a database and retrieved during interactions, ensuring the LLM works with up-to-date data without requiring constant retraining.
Solution Approach 2:
An intermediary information retrieval system is introduced between the user query and the LLM. This intermediary searches external data sources, retrieves relevant current information, and provides it to the LLM as context, allowing the model to generate accurate responses without being retrained on new data.
2Reliability
If LLMs are updated frequently with fresh data, then the responses remain accurate and up-to-date, but the computational cost and time increase
Solution Approach 1:
The system performs preliminary actions by collecting and summarizing current information about entities from external data sources before the LLM generates responses. This pre-collected information is stored in a database and retrieved during interactions, ensuring the LLM works with up-to-date data without requiring constant retraining.
Solution Approach 2:
Instead of copying the entire training process, the system creates lightweight summaries of current entity information from external sources. These summaries are stored in a database and retrieved as needed, providing fresh data without the computational overhead of full model retraining.
3Reliability
If LLMs are updated frequently with fresh data, then the responses remain accurate and up-to-date, but the computational resources and energy consumption increase
Solution Approach 1:
Instead of copying the entire training process, the system creates lightweight summaries of current entity information from external sources. These summaries are stored in a database and retrieved as needed, providing fresh data without the computational overhead of full model retraining.
Solution Approach 2:
The system extracts only the necessary current information about entities from external data sources and stores it separately in a database. This extraction approach allows the LLM to access fresh data without processing or retraining on entire datasets, significantly reducing computational energy consumption.
4Reliability
If LLMs generate responses without external information, then the system operation is simple, but the responses become hallucinated and inaccurate
Solution Approach 1:
An intermediary information retrieval system is introduced between the user query and the LLM. This intermediary searches external data sources, retrieves relevant current information, and provides it to the LLM as context, allowing the model to generate accurate responses without being retrained on new data.
Solution Approach 2:
The system performs preliminary actions by collecting and summarizing current information about entities from external data sources before the LLM generates responses. This pre-collected information is stored in a database and retrieved during interactions, ensuring the LLM works with up-to-date data without requiring constant retraining.
Data Source
AI summary
Implementations relate to generating an external entity-summarization database to store summarized descriptions for feature(s) of entities (e.g., brand A laptop, brand B laptop, brand C SUV, etc.) that belong to one or more entity classes (e.g., laptop, SUV) using a first generative model. Implementations relate to utilizing the external entity-summarization database in generating a response responsive to a user query seeking information (e.g., recommendation of one or more entities) of a particular entity class. For instance, a summarized description for features of one or more entities that belong to the particular entity class can be identified based on querying the external entity-summarization database using a query-based embedding generated for the user query. The summarized description can be used to generate the response responsive to the user query.