Linked Data Entity Summarization

Andreas Thalhammer
In recent years, the availability of structured data on the Web has grown and the Web has become more and more entity-focused. An entity can be a person, a book, a city, etc. In fact, all of these entities are connected in a large knowledge graph. In consequence, a lot of data is often available for single entities. However, in its complete form, the data is not always useful for humans unless it is presented in a concise manner. 2. We introduce a common API for publishing and consuming entity
more » ... ummaries. 3. We propose an entity-centric data fusion approach that enables an alignment of facts about entities from multiple open Web sources in a schema-agnostic way. We evaluated the contributions individually in accordance to state-of-the-art evaluation setups, implemented prototypes, and made different research datasets publicly available. The outcomes of our experiments lead us to conclude that 1) minimal background knowledge can be leveraged for producing state-of-the-art entity summaries (as exemplified with Web link structure and usage data); 2) entity summaries share many characteristics that make a common entity summarization API feasible (demonstrated with the introduction of an API, a proof-of-concept implementation, and an empirical analysis); 3) for Web-scale entity data fusion, two factors can enable robustness against the use of different vocabularies and modeling granularities: entity-centricity and the use of path features. v Similar to entities, it is very hard to summarize all the support that I received in almost six years. This is an attempt: I'm deeply indebted to Rudi Studer, who gave me the opportunity to work in this great environment and who interprets his role as Doktorvater (German for "doctoral adviser") literally: his guidance has always been built on values, freedom, and trust.
doi:10.5445/ir/1000065395 fatcat:ykjfzsymaza4zdiyytb4ttbimi