arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2102.06701cs.LGcond-mat.dis-nnstat.ML
收藏

解释神经缩放定律

Explaining Neural Scaling Laws

Google DeepMind(谷歌DeepMind);Johns Hopkins University(约翰斯·霍普金斯大学)

Yasaman Bahri, Ethan Dyer, Jared Kaplan, Jaehoon Lee, Utkarsh Sharma

AI总结:

该论文提出了一个理论框架,解释深度神经网络中群体损失与数据集大小和参数数量之间的幂律缩放关系,识别出四种缩放机制,并通过实验验证了这些机制的存在及其相互关系。

摘要:

训练后的深度神经网络的群体损失通常与训练数据集的大小或网络中的参数数量遵循精确的幂律缩放关系。我们提出了一个理论,解释这些缩放定律的起源并建立它们之间的联系。我们针对数据集和模型大小分别识别出方差受限和分辨率受限的缩放行为,共计四种缩放机制。方差受限的缩放仅源于存在良态的无限数据或无限宽度极限,而分辨率受限的机制可以通过假设模型有效地解析平滑数据流形来解释。在无限宽度极限下,这可以等价地从特定核的谱中获得,并且我们提出了证据表明大宽度与大数据集的分辨率受限缩放指数通过一种对偶性相关联。我们在大型随机特征和预训练模型的受控环境中展示了所有四种缩放机制,并在一系列标准架构和数据集上对预测进行了实证测试。我们还观察到在修改任务和架构长宽比的情况下,数据集与缩放指数之间的若干经验关系。我们的工作提供了一个分类不同缩放机制的分类法,强调了驱动损失改善的机制可能不同,并为缩放指数的微观起源及其相互关系提供了洞见。

英文摘要:

The population loss of trained deep neural networks often follows precise power-law scaling relations with either the size of the training dataset or the number of parameters in the network. We propose a theory that explains the origins of and connects these scaling laws. We identify variance-limited and resolution-limited scaling behavior for both dataset and model size, for a total of four scaling regimes. The variance-limited scaling follows simply from the existence of a well-behaved infinite data or infinite width limit, while the resolution-limited regime can be explained by positing that models are effectively resolving a smooth data manifold. In the large width limit, this can be equivalently obtained from the spectrum of certain kernels, and we present evidence that large width and large dataset resolution-limited scaling exponents are related by a duality. We exhibit all four scaling regimes in the controlled setting of large random feature and pretrained models and test the predictions empirically on a range of standard architectures and datasets. We also observe several empirical relationships between datasets and scaling exponents under modifications of task and architecture aspect ratio. Our work provides a taxonomy for classifying different scaling regimes, underscores that there can be different mechanisms driving improvements in loss, and lends insight into the microscopic origins of and relationships between scaling exponents.

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。

↑