We present RUDIK, a system for the discovery of declarative rules over knowledge-bases (KBs). RUDIK discovers rules that express positive relationships between entities, such as "if two persons have the same parent, they are siblings", and negative rules, i.e., patterns that identify contradictions in the data, such as "if two persons are married, one cannot be the child of the other". While the former class infers new facts in the KB, the latter class is crucial for other tasks, such as detecting erroneous triples in data cleaning, or the creation of negative examples to bootstrap learning algorithms. The system is designed to: (i) enlarge the expressive power of the rule language to obtain complex rules and wide coverage of the facts in the KB, (ii) discover approximate rules (soft constraints) to be robust to errors and incompleteness in the KB, (iii) use disk-based algorithms, effectively enabling rule mining in commodity machines. In contrast with traditional ranking of all rules based on a measure of support, we propose an approach to identify the subset of useful rules to be exposed to the user. We model the mining process as an incremental graph exploration problem and prove that our search strategy has guarantees on the optimality of the results. We have conducted extensive experiments using real-world KBs to show that RUDIK outperforms previous proposals in terms of efficiency and that it discovers more effective rules for the application at hand.

RuDiK is a system for the discovery of declarative rules over knowledge-bases (KBs). RuDiK discovers both positive rules, which identify relationships between entities, e.g., "if two persons have the same parent, they are siblings", and negative rules, which identify data contradictions, e.g., "if two persons are married, one cannot be the child of the other". Rules help domain experts to curate data in large KBs. Positive rules suggest new facts to mitigate incompleteness and negative rules detect erroneous facts. Also, negative rules are useful to generate negative examples for learning algorithms. RuDiK goes beyond existing solutions since it discovers rules with a more expressive rule language w.r.t. previous approaches, which leads to wide coverage of the facts in the KB, and its mining is robust to existing errors and incompleteness in the KB. The system has been deployed for multiple KBs, including Yago, DBpedia, Freebase and Wiki-Data, and identifies new facts and real errors with 85% to 97% accuracy, respectively. This demonstration shows how RuDiK can be used to interact with domain experts. Once the audience pick a KB and a predicate, they will add new facts, remove errors, and train a machine learning system with automatically generated examples.

Automated web scraping is a popular means for acquiring data from the web. Scrapers (or wrappers) are derived from either manually or automatically annotated examples, often resulting in under/over segmented data, together with missing or spurious content. Automatic repair and maintenance of the extracted data is thus a necessary complement to automatic wrapper generation. Moreover, the extracted data is often the result of a long-term data acquisition effort and thus jointly repairing wrappers together with the generated data reduces future needs for data cleaning. We study the problem of computing joint repairs for XPath-based wrappers and their extracted data. We show that the problem is NP-complete in general but becomes tractable under a few natural assumptions. Even tractable solutions to the problem are still impractical on very large datasets, but we propose an optimal approximation that proves effective across a wide variety of domains and sources. Our approach relies on encoded domain knowledge, but require no per-source supervision. An evaluation spanning more than 100k web pages from 100 different sites of a wide variety of application domains, shows that joint repairs are able to increase the quality of wrappers between 15% and 60% independently of the wrapper generation system, eliminating all errors in more than 50% of the cases.

Web scraping (or wrapping) is a popular means for acquiring data from the web. Recent advancements have made scalable wrapper-generation possible and enabled data acquisition processes involving thousands of sources. This makes wrapper analysis and maintenance both needed and challenging as no scalable tools exists that support these tasks. We demonstrate WADaR, a scalable and highly automated tool for joint wrapper and data repair. WADaR uses off-the-shelf entity recognisers to locate target entities in wrapper-generated data. Markov chains are used to determine structural repairs, that are then encoded into suitable repairs for both the data and corresponding wrappers. We show that WADaR is able to increase the quality of wrapper-generated relations between 15% and 60%, and to fully repair the corresponding wrapper without any knowledge of the original website in more than 50% of the cases.

Named entity extractors can be used to enrich both text and Web documents with semantic annotations. While originally focused on a few standard entity types, the ecosystem of annotators is becoming increasingly diverse, with recognition capabilities ranging from generic to specialised entity types. Both the overlap and the diversity in annotator vocabularies motivate the need for managing and integrating semantic annotations: allowing users to see the results of multiple annotations and to merge them into a unified solution. We demonstrate ROSEANN, a system for the management of semantic annotations. ROSEANN provides users with a unified view over the opinion of multiple independent annotators both on text and Web documents. It allows users to understand and reconcile conflicts between annotations via ontology-aware aggregation. ROSEANN incorporates both supervised aggregation, appropriate when representative training data is available, and an unsupervised method based on the notion of weighted-repair.

A growing number of resources are available for enriching documents with semantic annotations. While originally focused on a few standard classes of annotations, the ecosystem of annotators is now becoming increasingly diverse. Although annotators often have very different vocabularies, with both high-level and specialist concepts, they also have many semantic interconnections. We will show that both the overlap and the diversity in annotator vocabularies motivate the need for semantic annotation integration: middleware that produces a unified annotation on top of diverse semantic annotators. On the one hand, the diversity of vocabulary allows applications to benefit from the much richer vocabulary available in an integrated vocabulary. On the other hand, we present evidence that the most widely-used annotators on the web suffer from serious accuracy deficiencies: the overlap in vocabularies from individual annotators allows an integrated annotator to boost accuracy by exploiting inter-annotator agreement and disagreement. The integration of semantic annotations leads to new challenges, both compared to usual data integration scenarios and to standard aggregation of machine learning tools. We overview an approach to these challenges that performs ontology-aware aggregation. We introduce an approach that requires no training data, making use of ideas from database repair. We experimentally compare this with a supervised approach, which adapts maximal entropy Markov models to the setting of ontology-based annotations. We further experimentally compare both these approaches with respect to ontology-unaware supervised approaches, and to individual annotators.