Book Image

Responsible AI in the Enterprise

By : Adnan Masood, Heather Dawe
5 (1)
Book Image

Responsible AI in the Enterprise

5 (1)
By: Adnan Masood, Heather Dawe

Overview of this book

Responsible AI in the Enterprise is a comprehensive guide to implementing ethical, transparent, and compliant AI systems in an organization. With a focus on understanding key concepts of machine learning models, this book equips you with techniques and algorithms to tackle complex issues such as bias, fairness, and model governance. Throughout the book, you’ll gain an understanding of FairLearn and InterpretML, along with Google What-If Tool, ML Fairness Gym, IBM AI 360 Fairness tool, and Aequitas. You’ll uncover various aspects of responsible AI, including model interpretability, monitoring and management of model drift, and compliance recommendations. You’ll gain practical insights into using AI governance tools to ensure fairness, bias mitigation, explainability, privacy compliance, and privacy in an enterprise setting. Additionally, you’ll explore interpretability toolkits and fairness measures offered by major cloud AI providers like IBM, Amazon, Google, and Microsoft, while discovering how to use FairLearn for fairness assessment and bias mitigation. You’ll also learn to build explainable models using global and local feature summary, local surrogate model, Shapley values, anchors, and counterfactual explanations. By the end of this book, you’ll be well-equipped with tools and techniques to create transparent and accountable machine learning models.
Table of Contents (16 chapters)
1
Part 1: Bigot in the Machine – A Primer
4
Part 2: Enterprise Risk Observability Model Governance
9
Part 3: Explainable AI in Action

Monitoring and mitigating drift in ML models

All three types of drift (model, data, and concept) are important to measure, as they impact the performance and accuracy of ML models. Monitoring and mitigating each type of drift is essential to maintain model performance over time:

  • As discussed, model drift occurs when a model’s performance degrades as it becomes outdated due to changes in the underlying data distribution or concept.

Mitigation: Regularly retrain the model with fresh, representative data to maintain its performance. For example, retrain a sales prediction model with new sales data to capture recent trends and changes in customer behavior.

  • Data drift occurs when the input data distribution changes over time, making the model’s training data less representative of the current data.

Mitigation: Continuously monitor the distribution of input features and compare them to the training data. If significant deviations are detected,...