Data Party THU

Sep 12, 2026 · Artificial Intelligence

Quantifying Neural Network Simplicity with Effective Degree: A Polynomial Approach

Tsinghua researchers propose Effective Degree (ED), a differentiable measure of neural network simplicity based on polynomial representations along data interpolation paths, which correlates strongly with generalization gap and can be used as a regularizer to improve performance across vision, language, and RL tasks.