LLM Digital Assistant Graphical Explanation Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing enterprise systems struggle to present complex data insights from machine learning systems in a user-friendly and intuitive manner, as they lack the ability to generate graphical representations within digital assistants, making it difficult for users to understand and interpret data effectively.
Innovation Solution
A digital assistant system that leverages large language models (LLMs) to generate text and code for graphical representations, integrating them into graphical user interfaces (GUIs) within digital assistants, enabling the presentation of explanatory text and graphical data representations in response to user queries.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If enterprise systems use traditional data presentation methods, then system complexity is reduced, but user understanding and interpretation of data becomes difficult
Solution Approach 1:
The patent introduces an intermediary component (explanation generation module) that translates complex machine learning data and operations into human-friendly explanations. This mediator converts technical data processing steps into natural language descriptions, making the system accessible to non-experts without increasing apparent complexity for users.
Solution Approach 2:
The patent replaces traditional mechanical data presentation methods (tables, charts) with intelligent text generation using large language models. This substitution enables dynamic, context-aware explanations that adapt to user needs, significantly improving data interpretability while the underlying complexity is managed through automated AI processes.
2Ease of operation
If graphical representations are integrated into digital assistants, then user-friendly data presentation is improved, but system complexity increases
Solution Approach 1:
The patent creates a universal explanation generation framework that can handle multiple types of machine learning data and queries through a single integrated system. The large language model serves multiple functions: generating explanations, creating graphical representations, and adapting to different user contexts, thereby improving usability without proportionally increasing system complexity.
Solution Approach 2:
The system implements self-service capabilities where the digital assistant automatically generates appropriate explanations and graphical representations based on user queries without requiring manual configuration. The AI model autonomously determines the best presentation format and content, reducing the operational burden on users while managing complexity through automated decision-making.
3Loss of information
If LLM-generated text and graphical representations are provided together, then user understanding is enhanced, but processing requirements increase
Solution Approach 1:
The patent applies partial action by generating explanations and graphical representations selectively based on user needs and query complexity. The system assesses each request and provides appropriate levels of detail and visualization, avoiding unnecessary processing while ensuring sufficient information is provided for effective data interpretation.
Solution Approach 2:
The system performs preliminary analysis of user queries to determine the most efficient generation path. By pre-assessing information requirements and selecting appropriate explanation strategies before full processing, the system minimizes unnecessary computational effort while ensuring all necessary information for data understanding is delivered.
Data Source
AI summary
Methods, systems, and computer-readable storage media for receiving a query from a digital assistant of an enterprise system, retrieving data that is responsive to the query from a data management system, inputting a first few-shot prompt to a LLM, and determining, in response to the first few-shot prompt, that a graphical representation of the data is to be generated, and in response: inputting a second few-shot prompt and a third few-shot prompt to the LLM, receiving code from the LLM responsive to the third few-shot prompt, and executing the code to render the graphical representation with the digital assistant.