信息工程学院

讲座时间：2022年4月15日14时30分

讲座地点：线下工B315，线上腾讯会议

讲座对象：信息工程学院全体师生

讲座摘要：

Anomalies are rare observations that deviate significantly from the others in the sample. Over the past few decades, research on anomaly mining has received increasing interest due to the implications of these occurrences in various disciplines - for instance, security, finance, and medicine. Anomaly detection has shown its power in preventing detrimental events, such as financial fraud, network intrusions, and social spam. The detection task is typically solved by identifying outlying data points in the feature space, which overlooks the relational information in real-world data. At the same time, graphs have been prevalently used to represent the structural/relational information, which raises the graph anomaly detection problem - identifying anomalous graph objects. The objects are nodes, edges and sub-graphs in a single graph or a set/database. Conventional anomaly detection techniques cannot tackle the complexity of graph data well, such as irregular structures, relational dependencies, node/edge types/attributes/directions/multiplicities/weights, large scale, etc. The advent of deep learning in breaking these limitations. This seminar will include a brief introduction to deep graph anomaly detection, an example of methodology development and an example of implementation.