Conversational Understanding Toolset for Automated Model Training
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The development of conversational understanding systems requires expertise in each step, including data labeling, which can be time-consuming and limits accessibility for developers who lack specialized knowledge.
Innovation Solution
A suite of tools within a conversational understanding service platform allows developers to select, extend, or create domains, using APIs, labeling, training, and validation tools to build and update models, enabling them to interact with the service and create applications without needing extensive expertise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If experts are used in each step of the CU building process including data labeling, then the quality and accuracy of the conversational understanding system is improved, but the development time and complexity increase significantly
Solution Approach 1:
The system enables developers to perform data labeling and model training themselves using automated tools and pre-configured templates, eliminating the need to rely on specialized experts for every step of the development process
Solution Approach 2:
The platform provides pre-configured domains, templates, and automated labeling tools that are prepared in advance, allowing developers to quickly start building CU systems without needing to create everything from scratch or require expert guidance for each component
2Measurement precision
If experts are used in each step of the CU building process, then the quality and accuracy of the conversational understanding system is improved, but the accessibility for developers without specialized knowledge deteriorates
Solution Approach 1:
The platform provides a universal set of tools and templates that can be used by developers with varying levels of expertise, allowing the same system to serve both beginners and advanced users without requiring specialized knowledge at each step
Solution Approach 2:
The automated labeling tools and pre-configured templates act as intermediaries between the developer and the complex expert-level processes, translating simple developer actions into sophisticated model training and data labeling operations
3Adaptability or versatility
If manual data labeling and model training processes are used, then the customization and flexibility of the CU system is improved, but the development complexity and resource requirements increase
Solution Approach 1:
The platform segments the complex CU development process into manageable, modular components such as domain selection, data labeling, model training, and validation, each handled by separate automated tools that can be independently configured and combined
Data Source
AI summary
Tools are provided to allow developers to enable applications for Conversational Understanding (CU) using assets from a CU service. The tools may be used to select functionality from existing domains, extend the coverage of one or more domains, as well as to create new domains in the CU service. A developer may provide example Natural Language (NL) sentences that are analyzed by the tools to assist the developer in labeling data that is used to update the models in the CU service. For example, the tools may assist a developer in identifying domains, determining intent actions, determining intent objects and determining slots from example NL sentences. After the developer tags all or a portion of the example NL sentences, the models in the CU service are automatically updated and validated. For example, validation tools may be used to determine an accuracy of the model against test data.