深度学习中可验证准确性、鲁棒性与泛化性的边界
The Boundaries of Verifiable Accuracy, Robustness, and Generalisation in Deep Learning
AI总结:
本文从理论上证明,在分布无关框架和经验风险最小化设置下,即使理想解存在,计算并验证既稳定又准确的神经网络对一大类任务而言也极其困难。
摘要:
在这项工作中,我们评估了在分类任务中确定神经网络保证稳定性和准确性的理论局限性。我们考虑经典的分布无关框架以及最小化经验风险并可能施加某些权重正则化的算法。我们证明,存在一大类任务,在上述设置下计算和验证理想稳定且准确的神经网络极其困难,甚至可能无法实现,即使此类理想解存在于给定的神经架构类别中。
英文摘要:
In this work, we assess the theoretical limitations of determining guaranteed stability and accuracy of neural networks in classification tasks. We consider classical distribution-agnostic framework and algorithms minimising empirical risks and potentially subjected to some weights regularisation. We show that there is a large family of tasks for which computing and verifying ideal stable and accurate neural networks in the above settings is extremely challenging, if at all possible, even when such ideal solutions exist within the given class of neural architectures.
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。