Area B: Natural Language Processing for multi-document summarization

The central research topic of Area B is the development of methods for the creation of informative texts summarizing multiple heterogeneous documents. This includes methods for structuring document collections, as well as extracting and condensing information from heterogeneous documents, and making it uniform regarding the language style. All of these tasks are investigated in close collaboration with Area C: Representation and analysis of text-based structures. Also, they complement Area A: Graph-based discourse processing, which investigates the identification of entities and events across multiple documents based on discourse semantics.

The PhD projects in Area B explore the interaction between natural language processing and linguistically motivated discourse graphs on the one hand, and statistical and punctual phenomena in the induced graphs (networks) and other induced structures ?on the other hand. The three guiding themes of Area B approach their common goal from three different, complementary perspectives and address adaptive information processing, i.e. multi-document summarization in the first phase of AIPHES, under a common umbrella.