Model risk management and model governance in Watson OpenScale
Use Watson OpenScale to manage risk from machine learning models and to remain in compliance with governance standards.
Planning for model risk management and governance
Plan your risk management and governance strategy by considering what machine learning models you are deploying in your organization and considering your governance requirements.
Governance with Watson OpenScale
Use Watson OpenScale as a tool. Enable the model risk management features to create pre-production and production repositories and compare models. Monitor your deployed models for the following considerations:
- Drift: Any change in input data also known as Drift can cause the model to make inaccurate decisions, impacting business KPIs
- Bias: Training data can be cleaned to be free from bias but runtime data might induce biased behavior of model
- Explainability: Traditional statistical models are simpler to interpret and explain
- Missing Validation or Test Data: Model training data sets might not capture the range of data or combinations that are encountered in runtime. Validation and monitoring of AI models is necessary in addition to govern and manage risk.
Learn more
Parent topic: Evaluating AI models with Watson OpenScale