Extracting Databases from Dark Data with DeepDive

Zhang, Ce; Shin, Jaeho; Ré, Christopher; Cafarella, Michael; Niu, Feng

doi:10.1145/2882903.2904442

Cited by 36 publications

(16 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Second, we compare our greedy approximation algorithm for the joint inference of named entity disambiguation (NED) and co-reference resolution (CR) against an integer linear programming algorithm. Third, we compare QKBfly's capability of performing mentions of spouses extraction (i.e., the married to relation) against DeepDive [57]. Finally, as an extrinsic use case, we dynamically construct ad-hoc KBs for question answering.…”

Section: Methodsmentioning

confidence: 99%

“…The idea of on-the-fly KB construction was also brought up by [37] as a challenge, but not pursued any further. Declarative methods to IE and KBP, including DeepDive [42,48,57] and SystemT [11,12,45], work on carefully-predefined sets of predicates and rules. Thus, they cannot directly be applied in a spontaneous "on-thefly" manner.…”

Section: Methods Under Comparisonmentioning

confidence: 99%

“…The DeepDive project [42,48,57] has developed a highly versatile tool suite for information extraction (IE) and KB population (KBP), based on Markov Logic [14] and further techniques, including a variety of optimizations. Another ground-breaking project in this space is SystemT [11,12,45], which uses declarative rules for IE in a wide range of applications, including enterprise content analytics.…”

Section: Introductionmentioning

confidence: 99%

“…However, these approaches have high overhead and are not geared for dynamic knowledge acquisition. Declarative approaches to IE and KBP, such as DeepDive [42,48,57] and SystemT [11,12,45], require specifications of predicates and rules. Thus, they cannot be used in a spontaneous "on-the-fly" manner.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Query-driven on-the-fly knowledge base construction

Nguyen¹,

Abujabal²,

Tran

et al. 2017

Proc. VLDB Endow.

View full text Add to dashboard Cite

Today's openly available knowledge bases, such as DBpedia, Yago, Wikidata or Freebase, capture billions of facts about the world's entities. However, even the largest among these (i) are still limited in up-to-date coverage of what happens in the real world, and (ii) miss out on many relevant predicates that precisely capture the wide variety of relationships among entities. To overcome both of these limitations, we propose a novel approach to build on-the-fly knowledge bases in a query-driven manner. Our system, called QKBfly, supports analysts and journalists as well as question answering on emerging topics, by dynamically acquiring relevant facts as timely and comprehensively as possible. QKBfly is based on a semantic-graph representation of sentences, by which we perform three key IE tasks, namely named-entity disambiguation, co-reference resolution and relation extraction, in a light-weight and integrated manner. In contrast to Open IE, our output is canonicalized. In contrast to traditional IE, we capture more predicates, including ternary and higher-arity ones. Our experiments demonstrate that QKBfly can build high-quality, on-the-fly knowledge bases that can readily be deployed, e.g., for the task of ad-hoc question answering. PVLDB Reference Format:

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Methods Under Comparisonmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Query-driven on-the-fly knowledge base construction

Nguyen¹,

Abujabal²,

Tran

et al. 2017

Proc. VLDB Endow.

View full text Add to dashboard Cite

show abstract

“…Deepdive [7,31] is also an effort for populating knowledge bases from dark data [32]. The deep-dive approach utilizes database and machine learning techniques.…”

Section: Knowledge Base Constructionmentioning

confidence: 99%

A New Approach to Information Extraction in User-Centric E-Recruitment Systems

Awan

Khan

Latif³

et al. 2019

Applied Sciences

View full text Add to dashboard Cite

In modern society, people are heavily reliant on information available online through various channels, such as websites, social media, and web portals. Examples include searching for product prices, news, weather, and jobs. This paper focuses on an area of information extraction in e-recruitment, or job searching, which is increasingly used by a large population of users in across the world. Given the enormous volume of information related to job descriptions and users' profiles, it is complicated to appropriately match a user's profile with a job description, and vice versa. Existing information extraction techniques are unable to extract contextual entities. Thus, they fall short of extracting domain-specific information entities and consequently affect the matching of the user profile with the job description. The work presented in this paper aims to extract entities from job descriptions using a domain-specific dictionary. The extracted information entities are enriched with knowledge using Linked Open Data. Furthermore, job context information is expanded using a job description domain ontology based on the contextual and knowledge information. The proposed approach appropriately matches users' profiles/queries and job descriptions. The proposed approach is tested using various experiments on data from real life jobs' portals. The results show that the proposed approach enriches extracted data from job descriptions, and can help users to find more relevant jobs.

show abstract