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 2406.04047stat.MLcs.LG
收藏

神经网络的切片互信息泛化界

Slicing Mutual Information Generalization Bounds for Neural Networks

MIT(麻省理工学院);MIT-IBM Watson AI Lab(麻省理工学院-IBM沃森人工智能实验室);IBM Research(IBM研究院)

Kimia Nadjahi, Kristjan Greenewald, Rickard Brüel Gabrielsson, Justin Solomon

AI总结:

本文针对神经网络提出基于切片参数空间的互信息泛化界,利用分解互信息和k-切片互信息获得更紧且可计算的界,并结合率失真理论提出正则化方案,首次实现神经网络非空信息论泛化界计算。

摘要:

机器学习(ML)算法对未见数据良好泛化的能力已通过信息论的视角加以研究,其方式是用输入-输出互信息(MI),即训练数据与所学假设之间的互信息,来界定泛化误差。然而,由于在高维中评估互信息十分困难,这些界对现代机器学习应用(例如深度学习)的实用性有限。受近期关于神经网络可压缩性发现的启发,我们考虑通过切片参数空间来运作的算法,即在随机的低维子空间上进行训练。我们为这类算法引入了新的、更紧的信息论泛化界,证明切片能改善泛化。与标准互信息界相比,我们的界提供了显著的计算和统计优势,因为它们依赖于可扩展的替代依赖性度量,即分解互信息和$k$-切片互信息。随后,我们借助率失真理论将分析扩展到参数不必精确位于随机子空间上的算法。这一策略产生的泛化界包含一个失真项,用于度量切片下模型的可压缩性,从而在不牺牲性能或不需要模型压缩的情况下收紧现有界。在此基础上,我们提出了一种正则化方案,使实践者能够通过可压缩性来控制泛化。最后,我们通过实验验证了我们的结果,并实现了神经网络非空信息论泛化界的计算,而这一任务此前是无法完成的。

英文摘要:

The ability of machine learning (ML) algorithms to generalize well to unseen data has been studied through the lens of information theory, by bounding the generalization error with the input-output mutual information (MI), i.e., the MI between the training data and the learned hypothesis. Yet, these bounds have limited practicality for modern ML applications (e.g., deep learning), due to the difficulty of evaluating MI in high dimensions. Motivated by recent findings on the compressibility of neural networks, we consider algorithms that operate by slicing the parameter space, i.e., trained on random lower-dimensional subspaces. We introduce new, tighter information-theoretic generalization bounds tailored for such algorithms, demonstrating that slicing improves generalization. Our bounds offer significant computational and statistical advantages over standard MI bounds, as they rely on scalable alternative measures of dependence, i.e., disintegrated mutual information and $k$-sliced mutual information. Then, we extend our analysis to algorithms whose parameters do not need to exactly lie on random subspaces, by leveraging rate-distortion theory. This strategy yields generalization bounds that incorporate a distortion term measuring model compressibility under slicing, thereby tightening existing bounds without compromising performance or requiring model compression. Building on this, we propose a regularization scheme enabling practitioners to control generalization through compressibility. Finally, we empirically validate our results and achieve the computation of non-vacuous information-theoretic generalization bounds for neural networks, a task that was previously out of reach.

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。

↑