Relevance Search over Schema-Rich Knowledge Graphs

Gu, Yu; Zhou, Tianshuo; Cheng, Gong; Pan, Jeff Z.; Qu, Yuzhong

doi:10.1145/3289600.3290970

Cited by 16 publications

(32 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Following Definition 2.4, the rel function is conditioned on a set of user-provided examples S, so we rewrite rel(q, v) as rel(q, v |S). Previous research computes meta-path based relevance [3,6,9,15,18,20,22,24]. Along this line, we decompose rel into a linear combination of weighted relevance over a set of meta-paths denoted by Ω mp = {P 1 , .…”

Section: Generative Relevance Modelmentioning

confidence: 99%

“…Our creation of D1-D10 followed a common practice in related work [6,22]. Compared with D11-D14 and Y1-Y4 created in [6], our D1-D10 are more generalized as they allow the source entity in a user-provided example to be different from the query entity, and they are more challenging as D6-D10 involve properties. In general, their desired semantics of relevance are more complex than all the known queries used in the literature.…”

Section: Datasetsmentioning

confidence: 99%

“…For each query entity, based on the predefined semantics of Table 2 illustrates each group with one query instance under |S | = 2 and two of its gold-standard answer entities. The creation of D11-D14 and Y1-Y4 in [6] adopted a similar procedure, and we refer the reader to [6] for details. D11-D14 are based on DBpedia, and Y1-Y4 are based on YAGO.…”

Section: Datasetsmentioning

confidence: 99%

“…Problem. To bridge the gap between non-expert users and structured KGs, one practical solution is to request a small number of examples from the user, and then to learn user-defined relevance from user-provided examples [1,3,5,6,9,15,22,24]. The user can specify an example in the form of an ordered query-answer entity pair to illustrate the desired semantics of relevance [9,15,22].…”

Section: Introductionmentioning

confidence: 99%

“…Challenges. It has been common to assume that the user-defined relevance can be represented by one or more meta-paths [3,6,9,15,18,20,22,24]. A meta-path is a sequence of relation types, i.e., a path at the schema level.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

GREASE: A Generative Model for Relevance Search over Knowledge Graphs

Zhou

Cheng

Wang

et al. 2020

Proceedings of the 13th International Conference on Web Search and Data Mining

Self Cite

View full text Add to dashboard Cite

Relevance search is to find top-ranked entities in a knowledge graph (KG) that are relevant to a query entity. Relevance is ambiguous, particularly over a schema-rich KG like DBpedia which supports a wide range of different semantics of relevance based on numerous types of relations and attributes. As users may lack the expertise to formalize the desired semantics, supervised methods have emerged to learn the hidden user-defined relevance from user-provided examples. Along this line, in this paper we propose a novel generative model over KGs for relevance search, named GREASE. The model applies to meta-path based relevance where a meta-path characterizes a particular type of semantics of relating the query entity to answer entities. It is also extended to support properties that constrain answer entities. Extensive experiments on two large-scale KGs demonstrate that GREASE has advanced the state of the art in effectiveness, expressiveness, and efficiency.

show abstract