Scalable Conditional Independence Testing Using Spectral Representations

This research explores improvements in the efficiency and scalability of conditional independence testing, a crucial aspect of causal inference and machine learning. The use of spectral representations offers a novel approach to address computational bottlenecks in this important field.

• •Focuses on improving the computational efficiency of conditional independence testing.
• •Employs spectral representations as a key technique.
• •Addresses the scalability challenges in causal inference methods.