III: Small: Comprehensive Methods to Learn to Augment Graph Data

机器学习算法从数据中学习。训练数据的质量和数量不仅关系到算法本身，也关系到机器学习项目的成功。数据扩充方法的目的是获得训练数据中不容易获得的信息，但这是提高学习系统性能所必需的。数据增强已成功地应用于图像分类：从特殊增强，如裁剪、翻转、旋转和学习增强方法。本文主要研究图机器学习(GML)中的数据扩充问题。图神经网络(GNN)等GML方法在研究许多类型的具有重要社会意义的图方面发挥着重要作用，包括社会网络、分子网络和知识图。虽然增强方法以前也被应用于GML，但现有的技术是以特殊的方式应用的，产生的结果不是最优的。该项目将研究构建计算图形的原则性方法，扩充原始图形数据，旨在提高图形学习方法的性能。该项目的技术目标分为三个方面。第一个推力开发了新的图形机器学习技术，通过对作为处理变量的边的影响进行反事实推理来增加图形数据。第二个推力开发了新的图形机器学习技术，通过预测数据中是否存在顺序、时间或动态模式来增强图形。第三个推力开发了新的图形机器学习技术，通过对非常不同的社区或集群中的节点进行伪标记和断开连接来增强图形。该项目将结合统计因果分析、序列建模和预测以及图挖掘算法等多个研究领域的理论和方法，提供增加图形数据的新方法。它将推动图形机器学习技术的发展，扩大深度学习的数据增强范围。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Machine learning algorithms learn from data. The quality and quantity of training data has as much to do with the success of machine learning projects as the algorithms themselves. Data augmentation methods aim at eliciting information that is not readily available in the training data, but that is needed to improve the performance of the learning systems. Data augmentation has been successfully used in image classification: from adhoc augmentation, such as cropping, flipping, rotation and 'learning ­to ­augment' methods. This work focuses on data augmentation for Graph Machine Learning (GML). GML methods such as graph neural networks (GNNs) play an important role in studying many types of graphs of significant societal importance, including social networks, molecular networks, and knowledge graphs. While augmentation methods have been applied to GML before, existing techniques are applied in an adhoc manner and yield sub-optimal results. This project will study principled methods for building computational graphs, augmenting raw graph data, aiming at improving the performance of graph learning methods.The technical aims of the project are divided into three thrusts. The first thrust develops novel graph machine learning techniques to augment the graph data by counterfactual inference on the effect of edges as treatment variables. The second thrust develops novel graph machine learning techniques to augment the graph by forecasting if sequential, temporal, or dynamic patterns exist in the data. The third thrust develops novel graph machine learning techniques to augment the graph by pseudo labeling and disconnecting the nodes in very different communities or clusters. This project will deliver novel methods to augment graph data integrating with theories and methods from a variety of research fields, such as statistical causal analysis, sequence modeling and prediction, and graph mining algorithms. It will advance the technologies of graph machine learning and expand the scope of data augmentation for deep learning.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.