Summary

You can also learn more by viewing docs, API docs and terms.

Features

• Open

  Build and develop trusted AI by using IDEs of your choice, and empower your business and support teams with data-driven insights into how AI affects business outcomes.

• Payload logging

  Capture payload data and deployment output to monitor ongoing health of AI in business applications through operations dashboards, alerts, and open data mart access for custom reporting.

• Bias detection and mitigation

  Automatically detect when AI systems are delivering unfair outcomes at runtime, based on fairness attributes determined by the business. Mitigate bias by through intelligent recommendations of new data for model retraining.

• Trust and Transparency in AI

  Build, run and manage your AI - with trust and transparency - to give your business the confidence to deploy AI into its most essential production workloads.

• Explainable AI recommendations

  Trace and audit AI predictions made in production applications. Confidence in model predictions and factors contributing to final outcomes, or different outcomes, are displayed in understandable business terms.

• AI health monitoring

  Ensure the ongoing health of AI in business applications - ensuring accuracy, fairness and performance - with operations dashboards, alerts, and open data mart access for custom reporting.

• NeuNetS - Neural Network Synthesis (Beta)

  NeuNetS takes training data (images or text) that you provide, and synthesizes a trained model that can then be downloaded and deployed in Watson Machine Learning, locally, or to a third-party host.

Media

Insights