Antoine Amarilli

Predicting Completeness in Knowledge Bases

Knowledge bases such as Wikidata, DBpedia, or YAGO contain millions of entities and facts. In some knowledge bases, the correctness of these facts has been evaluated. However, much less is known about their completeness, i.e., the proportion of real facts that the knowledge bases cover. In this work, we investigate different signals to identify the areas where a knowledge base is complete. We show that we can combine these signals in a rule mining approach, which allows us to predict where facts may be missing. We also show that completeness predictions can help other applications such as fact prediction.

Finite Open-World Query Answering with Number Restrictions

Open-world query answering is the problem of deciding, given a set of facts, conjunction of constraints, and query, whether the facts and constraints imply the query. This amounts to reasoning over all instances that include the facts and satisfy the constraints. We study finite open-world query answering (FQA), which assumes that the underlying world is finite and thus only considers the finite completions of the instance. The major known decidable cases of FQA derive from the following: the guarded fragment of first-order logic, which can express referential constraints (data in one place points to data in another) but cannot express number restrictions such as functional dependencies, and the guarded fragment with number restrictions but on a signature of arity only two. In this paper, we give the first decidability results for FQA that combine both referential constraints and number restrictions for arbitrary signatures: we show that, for unary inclusion dependencies and functional dependencies, the finiteness assumption of FQA can be lifted up to taking the finite implication closure of the dependencies[5]. Our result relies on new techniques to construct finite universal models of such constraints, for any bound on the maximal query size.

IBEX: Harvesting Entities from the Web Using Unique Identifiers

In this paper we study the prevalence of unique entity identifiers on the Web. These are, e.g., ISBNs (for books), GTINs (for commercial products), DOIs (for documents), email addresses, and others. We show how these identifiers can be harvested systematically from Web pages, and how they can be associated with humanreadable names for the entities at large scale. Starting with a simple extraction of identifiers and names from Web pages, we show how we can use the properties of unique identifiers to filter out noise and clean up the extraction result on the entire corpus. The end result is a database of millions of uniquely identified entities of different types, with an accuracy of 73--96% and a very high coverage compared to existing knowledge bases. We use this database to compute novel statistics on the presence of products, people, and other entities on the Web.

Recent Topics of Research around the YAGO Knowledge Base

A knowledge base (KB) is a formal collection of knowledge about the world. In this paper, we explain how the YAGO KB is constructed. We also summarize our contributions to different aspects of KB management in general. One of these aspects is rule mining, i.e., the identification of patterns such as spouse(x,y) ∧ livesIn(x,z) ⇒livesIn(y,z). Another aspect is the incompleteness of KBs. We propose to integrate data from Web Services into the KB in order to fill the gaps. Further, we show how the overlap between existing KBs can be used to align them, both in terms of instances and in terms of the schema. Finally, we show how KBs can be protected by watermarking.