US20260065540
2026-03-05
Physics
G06T11/26
The patent application introduces a system that leverages generative AI to manage multi-dimensional data visualizations through natural language requests. This system enables users to interact with data visualizations by issuing commands in natural language, which are then interpreted by a large language model (LLM) to trigger specific functionalities. The system is designed to update data representations in user interfaces, such as adjusting filters or changing visualization views, based on structured objects generated by the LLM.
The system operates by receiving a natural language request, which references specific actions to be performed on data items. This request is processed by the LLM, which generates a structured object that is validated and used to execute the desired actions. The structured object can manipulate the controls of a displayed data set, enabling changes like adding or removing data dimensions and modifying graphical characteristics of visual elements.
In practice, the system can be implemented through a variety of computing environments, including cloud services and machine-hosted services. It uses a combination of data processors and non-transitory computer-readable storage media containing executable instructions. The system can be integrated into existing applications, allowing them to utilize shared or private datasets for enhanced functionality.
The system is versatile and can be applied in various contexts where data visualization is crucial. It can be used in applications that manage data through spreadsheets, databases, or other visualization tools. The LLM's ability to understand and process natural language requests makes it easier for users to interact with complex data sets without needing specialized technical knowledge.
By using natural language to control data visualizations, the system simplifies the process of data manipulation and analysis. It enables more intuitive user interactions with data, reducing the need for manual data manipulation. This approach can improve efficiency and accessibility in data-driven environments, making it easier for users to derive insights from multi-dimensional data.