Scalable knowledge harvesting with high precision and high recall

Nakashole, Ndapandula; Theobald, Martin; Weikum, Gerhard

doi:10.1145/1935826.1935869

Cited by 150 publications

(129 citation statements)

References 36 publications

Supporting

Mentioning

117

Contrasting

Order By: Relevance

“…Semi-supervised bootstrapping approaches such as NELL [5], PROSPERA [15] and BOA [9] start with a set of seed natural language patterns, then employ an iterative approach to both extract information for those patterns and learn new patterns. For NELL and PROPERA, the patterns and underlying schema are created manually, whereas they are created automically for BOA by using knowlege contained in DBpedia.…”

Section: Related Workmentioning

confidence: 99%

Relation Extraction from the Web Using Distant Supervision

Augenstein

Maynard

Ciravegna

2014

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. Extracting information from Web pages requires the ability to work at Web scale in terms of the number of documents, the number of domains and domain complexity. Recent approaches have used existing knowledge bases to learn to extract information with promising results. In this paper we propose the use of distant supervision for relation extraction from the Web. Distant supervision is a method which uses background information from the Linking Open Data cloud to automatically label sentences with relations to create training data for relation classifiers. Although the method is promising, existing approaches are still not suitable for Web extraction as they suffer from three main issues: data sparsity, noise and lexical ambiguity. Our approach reduces the impact of data sparsity by making entity recognition tools more robust across domains, as well as extracting relations across sentence boundaries. We reduce the noise caused by lexical ambiguity by employing statistical methods to strategically select training data. Our experiments show that using a more robust entity recognition approach and expanding the scope of relation extraction results in about 8 times the number of extractions, and that strategically selecting training data can result in an error reduction of about 30%.

show abstract

Section: Related Workmentioning

confidence: 99%

Relation Extraction from the Web Using Distant Supervision

Augenstein

Maynard

Ciravegna

2014

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…In each iteration, NELL uses the available instance knowledge to retrieve new instances of existing categories and relations between known instances by using pattern harvesting. The approach followed by PROSPERA [17] is similar to that of NELL but relies on the iterative harvesting of n-grams-itemset patterns. These patterns allow to generalize NL patterns found in text without introducing more noise into the patterns during the generalization process.…”

Section: Related Workmentioning

confidence: 99%

“…When using BOA iteratively, the output of each RDF generation would provide parts of the input for the subsequent extraction process. In previous work, semantic drift has been shown to be one of the key problems of such iterative approaches [4,17]. In order to maintain a high precision and to avoid semantic drift within the BOA framework, we solely select the top-n percent of all scored patterns for generating RDF.…”

Section: Rdf Generationmentioning

confidence: 99%

Extracting Multilingual Natural-Language Patterns for RDF Predicates

Gerber

Ngomo

2012

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. Most knowledge sources on the Data Web were extracted from structured or semi-structured data. Thus, they encompass solely a small fraction of the information available on the document-oriented Web. In this paper, we present BOA, a bootstrapping strategy for extracting RDF from text. The idea behind BOA is to extract naturallanguage patterns that represent predicates found on the Data Web from unstructured data by using background knowledge from the Data Web. These patterns are used to extract instance knowledge from naturallanguage text. This knowledge is finally fed back into the Data Web, therewith closing the loop. The approach followed by BOA is quasi independent of the language in which the corpus is written. We demonstrate our approach by applying it to four different corpora and two different languages. We evaluate BOA on these data sets using DBpedia as background knowledge. Our results show that we can extract several thousand new facts in one iteration with very high accuracy. Moreover, we provide the first multilingual repository of natural-language representations of predicates found on the Data Web.

show abstract

“…The process of populating a structured relational database from unstructured sources has received renewed interest in the database community through high-profile start-up companies (e.g., Tamr and Trifacta), established companies like IBM's Watson [7,16], and a variety of research efforts [11,25,28,36,40]. At the same time, communities such as natural language processing and machine learning are attacking similar problems under the name knowledge base construction (KBC) [5,14,23].…”

Section: Introductionmentioning

confidence: 99%

“…1 DeepDive's language and execution model are similar to other KBC systems: DeepDive uses a high-level declarative language [11,28,30]. From a database perspective, DeepDive's language is based on SQL.…”

Section: Introductionmentioning

confidence: 99%

Incremental knowledge base construction using DeepDive

Shin

Wang

et al. 2016

The VLDB Journal

View full text Add to dashboard Cite

Populating a database with unstructured information is a long-standing problem in industry and research that encompasses problems of extraction, cleaning, and integration. Recent names used for this problem include dealing with dark data and knowledge base construction (KBC). In this work, we describe DeepDive, a system that combines database and machine learning ideas to help develop KBC systems, and we present techniques to make the KBC process more efficient. We observe that the KBC process is iterative, and we develop techniques to incrementally produce inference results for KBC systems. We propose two methods for incremental inference, based respectively on sampling and variational techniques. We also study the tradeoff space of these methods and develop a simple rule-based optimizer. DeepDive includes all of these contributions, and we evaluate Deep-Dive on five KBC systems, showing that it can speed up KBC inference tasks by up to two orders of magnitude with negligible impact on quality.

show abstract

Scalable knowledge harvesting with high precision and high recall

Cited by 150 publications

References 36 publications

Relation Extraction from the Web Using Distant Supervision

Relation Extraction from the Web Using Distant Supervision

Extracting Multilingual Natural-Language Patterns for RDF Predicates

Incremental knowledge base construction using DeepDive

Contact Info

Product

Resources

About