A knowledge base (KB) is fact-oriented and an ontology is much more schema-oriented. In the Google KGraph, you have certainly a schema (like the one described by DBPedia, Freebase, Yago, etc.) and a set of facts "Paris isA City, Paris hasInhabitants 2M, Paris isCapitalOf France, etc.) So, you can (easily) answer to questions like "Whiat is the capital of France?", "How many inhabitants does Paris have?" or you can provide a short description of Paris (or any given entity in general). But a domain ontology tells people which are the main concepts of a domain, how are these concepts related and which attributes do they have. Here the focus is on the description, with the highest possible expressiveness (disjointness, values restrictions, cardinality restrictions) and the useful annotations (synonyms, definition of terms, examples, comments on design choices, etc.), of the entities of a given domain. Data (or facts) are not the main concern when designing a domain ontology VS KB.
Hi Nazmul Hussain,
Relating to wikipedia, google knowledge graph is a knowledge base that can be represented by using ontology. In my opinion, google knowledge graph is a multi-domain ontology. It contains famous people, places, animals, events, history, and topics.
You can read the attached hyperlink to understand how google knowledge graph works for semantic search engine.
best regards,
Harry Achsan
https://partners.yext.com/community/weekly/semantic-search-knowledge-graph/
A Knowledge Graph is a knowledge base used by a company to enhance its search engine’s search results with semantic information gathered from a wide variety of sources based upon (unknown to us) “models”. Most importantly these include a “model” of the user, i.e. what the company assumes to be known about the user, say from input keywords and search history.
A Domain Ontology characterizes a domain in a neutral way, in the sense the no “models” are linked to it. Thus a Domain Ontology can be potentially used to answer the most challenging questions a user may pose: How to reach a piece of knowledge interesting for the user, which the user does not know about its existence?
A Knowledge Graph does a bit of organization of materials, saving some search steps, but cannot answer challenging questions per se. A Domain Ontology – demanding user judgment about the ontology quality and ontology fine tuning – has the potential to be used in tools to answer challenging questions.
Bibliography
I. Exman, "Interestingness – A Unifying Paradigm: Bipolar Function Composition", Proc. KDIR 2009, Int. Conf. Knowledge Discovery and Information Retrieval, A. Fred (ed.)- pp. 196-201, Funchal, Madeira, Portugal, October 2009. DOI:10.5220/0002308401960201. Also as: arXiv:1404.0091 [cs.IR]
http://arxiv.org/abs/1404.0091
Precisely, we can say that Google KnowledgeGraph deals with the various types of the REAL WORLD NAMED ENTITIES and the knowledge about them. For instance, Eiffel Tower (it is an artifact and got a name) -locatedIn-Paris (it is a city and got a name). It does not deal with any particular domain. The entire focus is ONLY on the entities.
On the other hand, a domain ontology is a collection of entity types and their properties. For instance, "Institutional ontology", a domain ontology can consist of a set of entity types (e.g., Person, Place, Publications, etc.) and their properties.
You may find an interesting article about the same here- https://hackernoon.com/wtf-is-a-knowledge-graph-a16603a1a25f
Followed by a brief research paper debating the definition of knowledge graphs - http://ceur-ws.org/Vol-1695/paper4.pdf
Hope it helps!
- Badges
- Science topic
- Similar topics
- Engineering
- Materials Engineering
- Coating