Virtual Assistant Training Interface With Corrective Test Actions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing virtual assistant builders lack intuitive interfaces and comprehensive training methods, requiring specialized expertise and leading to inefficiencies and erroneous intent detection.
Innovation Solution
A virtual assistant server environment with a unified interface and training suggestion engine that generates executable corrective actions based on test suite results, enabling users to efficiently train and correct virtual assistants.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple web pages are used to train different intents or skills, then comprehensive training coverage is achieved, but interface complexity and ease of operation deteriorate
Solution Approach 1:
The patent merges multiple training pages into a single unified interface that presents all training options in one view. The training interface consolidates intent and skill training into one cohesive screen, eliminating the need to navigate through multiple separate web pages while maintaining comprehensive training coverage.
Solution Approach 2:
The training interface is designed as a universal platform that handles multiple training functions (intent training, skill training, utterance addition, pattern configuration) within a single interface. This multi-functional design allows stakeholders to perform various training tasks without switching between different specialized interfaces.
2Reliability
If test suites are used to evaluate training sufficiency, then training quality assessment is improved, but time consumption and productivity deteriorate
Solution Approach 1:
The system automatically generates test cases based on the training data and intent definitions before actual testing begins. This preliminary generation of test cases from training materials eliminates the need for manual test case creation and reduces overall testing time while maintaining comprehensive quality assessment.
Solution Approach 2:
The training system includes self-service testing capabilities where the system automatically executes tests against the trained model and generates reports without requiring extensive manual intervention. This automated self-testing reduces time consumption while providing reliable quality assessment of training effectiveness.
3Measurement precision
If training data is manually crafted with sufficient variety, then intent detection accuracy is improved, but ease of manufacture and productivity deteriorate
Solution Approach 1:
The system uses templates and patterns to generate training data automatically rather than requiring manual creation of each training example. Stakeholders can define patterns and the system generates varied training utterances by copying and transforming these patterns, maintaining accuracy while significantly improving productivity.
Solution Approach 2:
The system allows stakeholders to define training parameters and characteristics, then automatically generates training data by varying these parameters. This parameter-based generation approach creates diverse training examples efficiently without requiring manual crafting of each utterance, maintaining intent detection accuracy while improving data creation efficiency.
Data Source
AI summary
A virtual assistant server determines a subset of test data corresponding to changes between a first version of training data of a virtual assistant and a second version of the training data of the virtual assistant. Subsequently, the virtual assistant server creates a test suite with the subset of test data and runs the test suite on a second language model of the virtual assistant created using the second version of the training data. Based on the running the test suite, the virtual assistant server generates one or more executable corrective actions to be implemented at the user device and provides the one or more executable corrective actions to the user device to implement to train the virtual assistant.