Transactional Knowledge Graph Generation To Model Adversarial Activities
Sumit Purohit, Patrick Mackey, William Smith, Madelyn Dunning, Miquette Orren, Trevor M Langlie-Miletich, Rahul Deshmukh, Ankur Bohra, Tonya J Martin, Dan J Aimone, George Chin · 2021 IEEE International Conference on Big Data (Big Data) · 2021
A Knowledge Graph (KG) is a formal and structured representation of entities, relationships, and their semantic descriptions. Traditionally, KGs are used to describe metadata about entities and provide additional context to a target application. Many real-world domains also involve temporal interactions between entities, in addition to the metadata. Modeling these attributed transactions is a critical requirement when using KGs in complex real-world applications, such as modeling adversarial activities. Adversarial activity modeling requires methodology and tools to produce realistic large-scale background graphs that include embedded Weapons of Mass Destruction (WMD) activity patterns [1]. We present a novel framework for constructing transactional knowledge graphs from a diverse set of sources. We present cloud-scale architecture of the framework, core components, and a real-world use case to demonstrate the generation of a background knowledge graph and WMD activity templates to evaluate network alignment and subgraph matching algorithms.