论文标题
神经崩溃，不受约束
Neural collapse with unconstrained features
论文作者
论文摘要
神经崩溃是深度学习中的一种新兴现象，该现象最近是由帕皮恩（Papyan），汉（Han）和多诺霍（Donoho）发现的。我们提出了一个简单的“无约束特征模型”，其中神经崩溃也从经验上出现。通过研究该模型，我们就经验风险的景观方面为神经崩溃的出现提供了一些解释。
Neural collapse is an emergent phenomenon in deep learning that was recently discovered by Papyan, Han and Donoho. We propose a simple "unconstrained features model" in which neural collapse also emerges empirically. By studying this model, we provide some explanation for the emergence of neural collapse in terms of the landscape of empirical risk.
