SEMLA MLOps Day at Polytechnique Montreal

Fairly AI co-founder and CEO David Van Bruwaene will be presenting at the Software Engineering for Machine Learning Applications (SEMLA, https://semla.quebec/en/) MLOps event in person on April 27, 2023.

‍

SEMLA is a Quebec research group that rallies software engineering and applied machine learning researchers from Quebec and Canadian Universities specialized in the engineering of Intelligent Software systems.

‍

Title

Policy-driven monitoring of AI systems

Abstract

The use of AI governance systems is essential in ensuring ethical and responsible use of AI technology. A comprehensive AI governance system should include features such as policy management, compliance monitoring, risk management, decision-making processes, metrics and reporting, and auditing and monitoring. The FAIRLY platform is a notable example of an AI governance system that promotes the safe and responsible use of AI. The “Critical Care Policy” is an example of a policy applied to AI systems in healthcare that ensures the AI system’s performance and ethical use by setting controls such as accuracy, precision, recall, human decision making, and racial group distribution. These policies are essential in ensuring that AI technology is used within ethical, legal, and performance boundaries. The evaluation process of a policy involves determining the risk status and compliance status of the AI system based on various inputs such as test results, metadata, human inputs/approvals, operational parameters, and contextual information. By using an AI governance system, organizations can streamline compliance and achieve minimum performance guarantees while fulfilling regulatory and internal obligations with ease.

‍