Responsible and Ethical Machine Learning

Lead: Ayris Narock

This team educates and develops resources encouraging the use of Responsible Machine Learning.  Responsible ML includes the practice of building explainability and interpretability into a model and addressing the issue of data bias.  It allows the practitioner to connect the input data to the output and trace what specific elements have the most influence on the model’s decision.  Responsible ML encourages the development of models that are transparent, explainable, verifiable, robust, and repeatable. 

Objectives

• Capture, disseminate, and encourage best practices for explainable, interpretable, and trustworthy AI in the Heliophysics community.
• Assist CfHA Members and Projects in understanding and complying with current and future NASA requirements and community standards for Ethical AI.
• In concert with the Capacity Building Team, ensure that the CfHA strives for a diverse and equitable environment

Related Resources

• Ethical AI
• Addressing Data Bias
• Explainable and Interpretable ML
• ML Cautionary Tales