Explore a cutting-edge approach to machine learning explainability in this 31-minute presentation from Anyscale. Dive into FastCFE, an innovative algorithm that leverages reinforcement learning to generate real-time counterfactual explanations. Learn how counterfactuals provide actionable steps for moving data points across decision boundaries, with applications in loan decisions and healthcare diagnoses. Gain insights into the fundamentals of counterfactuals and reinforcement learning, and witness a practical demonstration using OpenAI Gym and Ray+RLlib frameworks. Examine benchmarks and results that showcase the effectiveness of this approach in providing valuable explanations for stakeholders seeking to achieve different outcomes.
FastCFE - A Counterfactual Explainer Using Reinforcement Learning
Overview
Syllabus
FastCFE- A counterfactual explainer using reinforcement learning
Taught by
Anyscale
Related Courses
-
Understanding Algorithms for Reinforcement Learning
-
Reinforcement Learning Course - Full Machine Learning Tutorial
-
Python Reinforcement Learning using OpenAI Gymnasium – Full Course
-
Fundamentals of Reinforcement Learning
4.9 -
Ray/RLLib-Empowered Reinforcement Learning Based Recommender Systems in NetEase Game
-
Semiconductor Fab Production Scheduling Using Deep Reinforcement Learning
