The Dynamics of Quasiregular Neural Learning

2026-09-22

AI Summary

研究通过控制的准规则回归问题探讨了神经网络学习中规律性和例外性之间的竞争，发现罕见例外会导致更强的过度规则化。

Abstract

Many learning problems combine a dominant regularity with systematic exceptions. Motivated by U-shaped learning in language acquisition, we study this interaction in controlled quasiregular regression problems where regular and exceptional solutions are explicitly known. Neural networks can partially acquire exceptions, subsequently regress toward the dominant regularity, and finally recover. This overregularization becomes substantially stronger when exceptions are rare, despite their early acquisition, but does not emerge equally across all regularities considered. Our results isolate a simple form of competition between regularities and exceptions during neural learning.

Problem

• quasiregular regression
• neural learning
• overregularization
• U-shaped learning

Innovation

• quasiregular regression
• overregularization
• U-shaped learning