Robust Graph-Based Semi-Supervised Learning via $p$-Conductances
Department of Mathematics, UC San Diego(加州大学圣迭戈分校数学系);Halicioğlu Data Science Institute, UC San Diego(加州大学圣迭戈分校哈利乔格鲁数据科学研究所);Department of Mathematics, University of Missouri(密苏里大学数学系)
Sawyer Jack Robertson, Chester Holtz, Zhengchao Wan, Gal Mishne, Alexander Cloninger
We study the problem of semi-supervised learning on graphs in the regime where data labels are scarce or possibly corrupted. We propose an approach called $p$-conductance learning that generalizes the $p$-Laplace and Poisson learning methods by introducing an objective reminiscent of $p$-Laplacian regularization and an affine relaxation of the label constraints. This leads to a family of probability measure mincut programs that balance sparse edge removal with accurate distribution separation. Our theoretical analysis connects these programs to well-known variational and probabilistic problems on graphs (including randomized cuts, effective resistance, and Wasserstein distance) and provides motivation for robustness when labels are diffused via the heat kernel. Computationally, we develop a semismooth Newton-conjugate gradient algorithm and extend it to incorporate class-size estimates when converting the continuous solutions into label assignments. Empirical results on computer vision and citation datasets demonstrate that our approach achieves state-of-the-art accuracy in low label-rate, corrupted-label, and partial-label regimes.
Ladder Networks for Semi-Supervised Hyperspectral Image Classification
Purdue University(普渡大学);ETH Zürich(苏黎世联邦理工学院)
我们使用梯形网络[Rasmus et al.(2015)]在半监督设置中执行高光谱图像分类。梯形网络通过联合优化监督和非监督成本来将自己与其他半监督方法区分开来。在许多情况下,这已被证明比其他半监督技术更成功,例如使用未标记数据进行预训练。我们还表明,卷积梯形网络的性能优于目前大多数用于高光谱图像分类的技术,并在帕维亚大学数据集上实现了新的最先进的性能,每个类只有5个标记的数据点。