Natural-Language Intent Inference for Data Visualization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing visual analytics tools struggle to accurately interpret user intent, particularly when it is ambiguous or underspecified, leading to inefficient and device-specific implementations that are difficult to fine-tune.
Innovation Solution
A Visual Analytics Intent Language (VAIL) is introduced to translate high-level user intents into lower-level system actions, incorporating intent specifications, data semantics, and rules for inferring ambiguous or missing intents, enabling generalized use across various platforms and devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If visual analytics tools incorporate natural language interactions and machine learning-based analyses, then the sophistication of analytical tasks is improved, but the complexity of interpreting user intent and translating it into system actions increases
Solution Approach 1:
The patent introduces an intermediary layer (visual analytics intent language and inference system) between the user's natural language input and the system's analytical operations. This intermediary translates high-level user intents into lower-level system actions, resolving the contradiction by automating complex intent translation without requiring users to understand the underlying complexity.
Solution Approach 2:
The patent segments the intent translation process into distinct components: intent specification, data semantics, and inference rules. By dividing the complex translation process into manageable segments, the system can handle sophisticated analytical tasks while keeping each component's complexity controlled and manageable.
2Measurement precision
If visual analytics tools require users to select specific chart types and data attributes, then the precision of data analysis is improved, but the ease of operation deteriorates
Solution Approach 1:
The patent inverts the traditional approach by allowing users to specify what they want to analyze (high-level intent) rather than requiring them to specify how to analyze it (low-level operations). The system then automatically determines the appropriate chart types and data attributes, improving ease of operation while maintaining analysis precision through the inference system.
Solution Approach 2:
The system performs self-service by automatically selecting appropriate data attributes and visualization types based on the user's intent specification. This eliminates the need for users to manually configure complex parameters, making the tool easier to operate while maintaining precise data analysis through the system's automated decision-making.
3Measurement precision
If visual analytics tools implement bespoke business logic for each application, then the accuracy of intent interpretation is improved, but the adaptability across different platforms deteriorates
Solution Approach 1:
The patent creates a universal visual analytics intent language and inference system that can be applied across different platforms and applications. Instead of implementing bespoke business logic for each application, the system uses a common framework with platform-specific adapters, maintaining accurate intent interpretation while enabling broad adaptability across diverse platforms.
4Productivity
If visual analytics tools provide detailed control over analytical operations, then the productivity of data analysis is improved, but the cognitive burden on users increases
Solution Approach 1:
The patent extracts the complex cognitive tasks of selecting chart types, data attributes, and analysis parameters from the user and transfers them to the system's inference engine. Users only need to specify their analytical intent at a high level, while the system handles the detailed operational decisions, maintaining productivity by preserving detailed control capabilities while reducing cognitive burden.
Data Source
AI summary
A computing device receives a natural language command directed to a data source. The device applies one or more respective rules to determine whether the natural language command comprises an ambiguous and/or underspecified request. In accordance with a determination that the natural language command comprises an ambiguous and/or underspecified request, the device infers information to resolve the ambiguous and/or underspecified request according to (i) the one or more respective rules, (ii) metadata for the data source, and (iii) metadata for one or more data fields specified in the natural language command. The device updates a respective intent specification based on the inferred information, retrieving data sets from the data source according to the respective intent specification, and generates and displays a data visualization based on the retrieved data sets.