Active knowledge

Preda, Nicoleta; Kasneci, Gjergji; Suchanek, Fabian M.; Neumann, Thomas; Yuan, Wenjun; Weikum, Gerhard

doi:10.1145/1807167.1807212

Cited by 26 publications

(1 citation statement)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Examples include movie sales, stock prices, or chart positions of songs. Preda et al [62] refer to it as active knowledge, and propose to gather and integrate data from web services on the fly, transparently to the user, as if it was already contained in the knowledge base. In a recent study, Galárraga et al [29] make the first steps toward predicting the completeness of certain properties of objects in a KB.…”

Section: Further Readingmentioning

confidence: 99%

Populating Knowledge Bases

Balog

2018

Entity-Oriented Search

View full text Add to dashboard Cite

Populating Knowledge Bases A knowledge base (KB) contains information about entities and their properties (types, attributes, and relationships). In the case of large KBs, the number of entities is in the millions and the number of facts is in the billions. Commonly, this information is represented in the form of (sets of) subject-predicate-object (SPO) triples, according to the RDF data model (cf. Sect. 2.3). KBs are utilized in a broad variety of information access tasks, including entity retrieval (Chaps. 3 and 4), entity linking (Chap. 5), and semantic search (Chaps. 7-9). Two main challenges associated with knowledge bases are that (1) they are inherently incomplete (and will always remain so, despite any effort), and (2) they need constant updating over time as new facts and discoveries may turn the content outdated, inaccurate, or incomplete. Knowledge base population (KBP) refers to the task of discovering new facts about entities from a large text corpus, and augmenting a KB with these facts. KBP is a broad problem area, with solutions ranging from fully automated systems to setups with a human content editor in the loop, who is in charge of any changes made to the KB. Our interest in this chapter will be on the latter type of systems, which "merely" provide assistance with the labor-intensive manual process. Specifically, we will focus on a streaming setting, with the goal to discover and extract new information about entities as it becomes available. This information can then be used to augment an existing KB. This flavor of KBP has been termed knowledge base acceleration (KBA) [27]. KBA systems "seek to help humans expand knowledge bases [.. . ] by automatically recommending edits based on incoming content streams" [7]. There is a practical real-world motivation behind this particular problem formulation. Many large knowledge repositories are maintained by a small workforce of content editors. Due to the scarcity of human resources, "most entity profiles lag far behind current events" [26]. For example, Frank et al. [27] show that the median time elapsed between the publication dates of news articles that are cited in Wikipedia articles of living people and the dates of the corresponding edits to

show abstract

Section: Further Readingmentioning

confidence: 99%

Populating Knowledge Bases

Balog

2018

Entity-Oriented Search

View full text Add to dashboard Cite

show abstract

Calculating Error Bars on Inferences from Web Data

Nuamah

Bundy

2018

Advances in Intelligent Systems and Computing

View full text Add to dashboard Cite

In this work, we explore uncertainty in automated question answering over real-valued data from knowledge bases on the Internet. We argue that the coefficient of variation (cov) is an intuitive and general form in which to express this uncertainty, with the added advantage that it can be calculated exactly and efficiently. The large amounts of data on the Internet presents a good opportunity to answer queries that go beyond simply looking up facts and returning them. However, such data is often vague and noisy. For discrete results, e.g., stating that a particular city is the capital of a particular country, probabilities are a natural way to assign uncertainty to answers. For continuous variables or quantities that are typically treated as continuous (such as populations of countries), probabilities are uninformative, being infinitesimal For instance, the probability that the population of India is exactly equal to last census count is effectively zero. Our aim is to capture uncertainty in these estimates in an intuitive, uniform, and computationally efficient way. We present initial efforts at automating the inference process over real-valued web data while accounting for some of the typical sources of uncertainty: noisy data and errors from inference operations. Having considered several problem domains and query types, we find that approximating all continuous random variables with Gaussian distributions, and communicating uncertainties to users as coefficients of variation. Our experiments show that the estimates of uncertainty derived by our method are well-calibrated and correlate with the actual deviations from the true answer. An immediate benefit of our approach is that our inference framework can attach credible intervals 1 to real-valued answers that it infers. This conveys to a user the plausible magnitudes of the error in the answer, a meaningful measure of uncertainty compared to ranking scores provided in other question answering systems.

show abstract