Explore the concept of causal invariance in representation learning through this NYU ECE seminar featuring Leon Bottou from Facebook AI Research. Delve into topics such as the proxy problem, correlation environments, mixture coefficients, and adversarial domain adaptation. Gain insights into practical applications like hiring problems and observer self-supervised learning. Discover how causal invariance can enhance prediction accuracy across different environments and improve the robustness of AI models.
Overview
Syllabus
Introduction
Demo
Proxy Problem
Correlation
Environments
Mixture coefficient
trivialexistent cases
prediction
environment
adversarial domain adaptation
hiring problems
observer self
supervised problem
conclusion
Taught by
NYU Tandon School of Engineering
