AI Conversation Model Generation from API Test Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems for training and generating AI conversation agents are time-consuming and prone to human error, especially in highly technical scenarios like API support.
Innovation Solution
The process involves generating an AI conversation model directly from API test cases, building a corpus from API-related data, identifying intents and entities, and generating relationship models and dialog nodes to map to a conversational chat schema.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual training methods are used for AI conversation agents, then the conversation model can be trained with human expertise, but the process is time-consuming and prone to human error
Solution Approach 1:
The patent creates a conversational AI model by copying and transforming existing API test case data into a training corpus. Instead of manual creation, the system automatically generates intents, entities, and dialog nodes from API test case specifications, requests, and responses, thereby reducing training time while maintaining quality through structured data transformation
Solution Approach 2:
The patent performs preliminary actions by pre-processing API test case data into a structured training corpus format before actual model training. The system extracts intents, entities, and relationships from API test cases in advance, preparing the training data beforehand to accelerate the subsequent model training process
2Reliability
If manual training methods are used for AI conversation agents, then human expertise can be applied, but the process requires significant system resources
Solution Approach 1:
The system copies and reuses existing API test case data to create the training corpus, eliminating the need for separate manual data collection and annotation processes. This approach reduces system resources by leveraging already-available structured data from API testing
Solution Approach 2:
The patent makes the API test case data serve multiple functions: it is used both for API testing and for training the conversational AI model. This multi-functionality reduces system resources by eliminating redundant data preparation processes and leveraging the same data asset for dual purposes
3Productivity
If automated generation from API test cases is used, then time and resources are reduced, but the complexity of processing and mapping data increases
Solution Approach 1:
The patent segments the data processing into distinct components: extracting intents from API test case descriptions, extracting entities from requests and responses, identifying relationships between entities, and mapping these to dialog nodes. This segmentation manages complexity by breaking down the automated generation process into manageable, specialized steps
4Productivity
If automated generation from API test cases is used, then training speed increases, but the precision of extracting intents and entities must be maintained
Solution Approach 1:
The system incorporates feedback mechanisms in the automated extraction process by using the structured relationships between API test case components to validate and refine extracted intents and entities. The mapping process from test case data to dialog nodes includes verification steps that ensure extraction accuracy while maintaining training speed
Data Source
AI summary
An embodiment includes generating an intent representative of a verb identified in application programming interface (API) test data associated with an API test case. The embodiment also generates a first entity representative of a first noun identified in the API test data and a second entity representative of a second noun identified in the API test data. The embodiment identifies the intent as being associated with a relationship between the first and second entities in the API test case. The embodiment generates a relationship model associated with the API test case, where the relationship model includes the intent as the relationship between the first entity and the second entity. The embodiment generates a conversation model of an artificial intelligence (AI) conversation agent by mapping the intent and the first and second entities to dialog nodes of the conversation model.