Correcting Biased Centered Kernel Alignment Measures in Biological and Artificial Neural Networks

arxiv:2405.01012

Original source · Paper JSON

Representation alignment · 2024-05-02

Cites (0) · Cited by (2)

Complete cleaned Markdown · Source Markdown including bibliography

Read in parts

1. Part 1
2. Part 2
3. Part 3

Contents

• Correcting Biased Centered Kernel Alignment Measures in Biological and Artificial Neural Networks
• Abstract
• 1 Introduction
• 2 CKA Overview
• 2.1 Mathematical Formulation & Properties
• 2.2 Criticisms
• 2.3 Applications to Neuroscience
• 3 Methods
• 3.1 Brain Representations
• 3.2 Model Representations
• 4 CKA Sensitivity to Random Data
• 5 CKA Sensitivity to Stimuli-Driven Neural Responses
• 6 CKA Sensitivity to Shuffled Data
• 7 Conclusion
• Acknowledgments
• Appendix A Appendix
• A.1 Generating Random Data
• A.2 MEG Preprocessing & Sensor Selection
• A.3 Code Implementation
• A.4 Per-Participant Functional MRI ROI Sizes
• A.5 Unbiased HSIC Kernel-Centering Formula
• A.6 Results for CKA Sensitivity to Shuffled Data (CORnet-S)
• A.7 Results for CKA Sensitivity to Partially Shuffled Data (ResNet18)
• Instructions for reporting errors