Controlling Neural Collapse Enhances Out-of-Distribution Detection and Transfer Learning
This addresses a key problem in machine learning for improving model robustness and reliability, though it is incremental as it builds on existing NC theory.
The paper tackles the trade-off between out-of-distribution (OOD) detection and generalization in deep neural networks by showing that neural collapse (NC) inversely affects these tasks, and proposes a method to control NC at different layers, achieving strong performance in experiments across datasets and architectures.
Out-of-distribution (OOD) detection and OOD generalization are widely studied in Deep Neural Networks (DNNs), yet their relationship remains poorly understood. We empirically show that the degree of Neural Collapse (NC) in a network layer is inversely related with these objectives: stronger NC improves OOD detection but degrades generalization, while weaker NC enhances generalization at the cost of detection. This trade-off suggests that a single feature space cannot simultaneously achieve both tasks. To address this, we develop a theoretical framework linking NC to OOD detection and generalization. We show that entropy regularization mitigates NC to improve generalization, while a fixed Simplex Equiangular Tight Frame (ETF) projector enforces NC for better detection. Based on these insights, we propose a method to control NC at different DNN layers. In experiments, our method excels at both tasks across OOD datasets and DNN architectures. Code for our experiments is available at: https://yousuf907.github.io/ncoodg