arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2110.02424cs.LG
收藏

实践中的谱偏差:函数频率在推广中的作用

Spectral Bias in Practice: The Role of Function Frequency in Generalization

University of California, Berkeley(加州大学伯克利分校);Google Brain(谷歌大脑)

Sara Fridovich-Keil, Raphael Gontijo-Lopes, Rebecca Roelofs

摘要:

尽管深度学习模型有能力表示高度表达的函数,但它们似乎能找到简单的解决方案,而且泛化效果令人惊讶。频谱偏差--神经网络优先学习低频函数的倾向--是这种现象的一种可能的解释,但到目前为止,频谱偏差主要是在理论模型和简化的实验中观察到的。在这项工作中,我们提出了在CIFAR-10和ImageNet上测量现代图像分类网络中光谱偏差的方法。我们发现,这些网络确实表现出频谱偏差,并且提高CIFAR-10测试准确性的干预措施往往会产生具有较高频率的学习函数,但在每个类别的示例附近具有较低频率。这种趋势在训练时间、模型架构、训练样本数量、数据增强和自蒸馏方面的变化中保持不变。我们还探讨了函数频率和图像频率之间的联系,发现频谱偏差是敏感的低频普遍存在于自然图像。在ImageNet上,我们发现学习函数的频率也随着内部类的多样性而变化,在更多样化的类上频率更高。我们的工作能够测量并最终影响用于图像分类的神经网络的光谱行为,并且是理解为什么深度模型泛化良好的一步。

英文摘要:

Despite their ability to represent highly expressive functions, deep learning models seem to find simple solutions that generalize surprisingly well. Spectral bias -- the tendency of neural networks to prioritize learning low frequency functions -- is one possible explanation for this phenomenon, but so far spectral bias has primarily been observed in theoretical models and simplified experiments. In this work, we propose methodologies for measuring spectral bias in modern image classification networks on CIFAR-10 and ImageNet. We find that these networks indeed exhibit spectral bias, and that interventions that improve test accuracy on CIFAR-10 tend to produce learned functions that have higher frequencies overall but lower frequencies in the vicinity of examples from each class. This trend holds across variation in training time, model architecture, number of training examples, data augmentation, and self-distillation. We also explore the connections between function frequency and image frequency and find that spectral bias is sensitive to the low frequencies prevalent in natural images. On ImageNet, we find that learned function frequency also varies with internal class diversity, with higher frequencies on more diverse classes. Our work enables measuring and ultimately influencing the spectral behavior of neural networks used for image classification, and is a step towards understanding why deep models generalize well.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑