0
7.0.18
Yokohama, Xanadu, Washington DC, Vancouver
Delivers advanced features for NLU Workbench including Multi-Model Batch Testing, Cross-model Conflict Review, Model Performance, and Expert Feedback Loop. This application is dependent on the Predictive Intelligence plugin and NLU Workbench plugin. This release provides a modern UI experience for a new NLU Home Page, Irrelevance Detection Page, and Expert Feedback Loop.
- Multi-model Batch Testing: Test large datasets with labeled utterances to see how your NLU models are performing without having to test each utterance individually. Use threshold recommendations to improve the prediction capabilities of your models.
- Cross-model Conflict Review: Identify conflicting intents within or across models so you can take corrective actions, resolve such conflicts, and improve your NLU performance.
- Model Performance: As NLU models for Virtual Agent (VA) are updated, published and deployed, the system tracks the efficacy of the intent predictions they make.
- Expert Feedback Loop: Provide feedback on correct intents for utterances selected from your VA logs. Your feedback helps the model continuously learn from your users' inputs to improve the model's performance. The new EFL page is built using the modern Servicenow UI components.
- An NLU Home Page with models for different applications organized in tabs.
- An Irrelevance Detection Page to manage utterances that are not relevant to NLU models.
- A new Expert Feedback Loop page to tune the ITSM Issue Auto Resolution model.
- Fixes related to batch testing and EFL
- Fixed - Defects - Security Bugs.
- Must have Predictive Intelligence and NLU Workbench plugins installed
- Requires Vancouver or later