Active learning (AL) aims to improve model performance within a fixed labeling budget by choosing the most informative data points to label. Existing AL focuses on the single-domain setting, where all data come from the same domain (e.g., the same dataset). However, many real-world tasks often involve multiple domains. For example, in visual recognition, it is often desirable to train an image classifier that works across different environments (e.g., different backgrounds), where images from each environment constitute one domain. Such a multi-domain AL setting is challenging for prior methods because they (1) ignore the similarity among different domains when assigning labeling budget and (2) fail to handle distribution shift of data across different domains. In this paper, we propose the first general method, dubbed composite active learning (CAL), for multi-domain AL. Our approach explicitly considers the domain-level and instance-level information in the problem; CAL first assigns domain-level budgets according to domain-level importance, which is estimated by optimizing an upper error bound that we develop; with the domain-level budgets, CAL then leverages a certain instance-level query strategy to select samples to label from each domain. Our theoretical analysis shows that our method achieves a better error bound compared to current AL methods. Our empirical results demonstrate that our approach significantly outperforms the state-of-the-art AL methods on both synthetic and real-world multi-domain datasets. Code is available at https://github.com/Wang-ML-Lab/multi-domain-active-learning.
Cyclic Block Coordinate Descent With Variance Reduction for Composite Nonconvex Optimization
University of Wisconsin-Madison（威斯康星大学麦迪逊分校）
非凸优化是解决许多机器学习问题的核心，在这些问题中通常会遇到块结构。在这项工作中，我们提出了非渐近梯度范数保证的非凸最优化问题的循环块坐标方法。我们的收敛性分析是基于梯度利普希茨条件的Mahalanobis规范，启发循环块坐标方法的最新进展。在确定性设置中，我们的收敛保证与（全梯度）梯度下降的保证相匹配，但梯度Lipschitz常数被定义为w.r.t.~ a Mahalanobis规范.在随机环境下，我们使用递归方差缩减来降低每次迭代的代价，并与当前最优随机全梯度法的运算复杂度相匹配，同时对有限和和无限和情况进行了统一的分析。我们证明了一个更快的线性收敛的结果时，Polyak-Pjojasiewicz（Pjs）条件成立。据我们所知，这项工作是第一个提供非渐近收敛保证-方差减少或不-循环块坐标方法在一般复合（光滑+非光滑）非凸设置。我们的实验结果证明了所提出的循环方案在训练深度神经网络方面的有效性。
Cell Library Characterization for Composite Current Source Models Based on Gaussian Process Regression and Active Learning
National ASIC System Engineering Technology Research Center, Southeast University, Nanjing, China（南京大学国家ASIC系统工程技术研发中心）；National Center of Technology Innovation for EDA, Nanjing, China（南京EDA技术创新国家中心）；University of California, Los Angeles, United States（加州大学洛杉矶分校）；Ningbo Institute of Digital Twin, Eastern Institute of Technology, Ningbo, China（宁波数字孪生研究所）；The University of Sheffield, Sheffield, United Kingdom（谢菲尔德大学）
