Knowledge Graph: Applications with ML and AI and Open-Source Database Links in 2022

By: A. Khan, F. Colace

Introduction

With every passing year, machine learning attracts researchers towards a new paradigm of learning. Nowadays dramatic advancements can be seen in the learning process of machines. A new terminology is coined by Google in 2012 “Knowledge Graph”. This knowledge graph has its own significance in the field of machine learning due to which, performing capabilities of machine learning techniques are getting better day by day with a high accuracy rate [1]. This article will provide the basics of Knowledge Graphs, its real-time applications with the integration of Machine Learning [8-10] and Artificial Intelligence. Moreover, direct links are also provided to the various free databases for Knowledge Graphs for model training purposes [2].

Till now a Knowledge Graph didn’t get any proper definition, many researchers present their own definitions of Knowledge Graphs. According to Amber Lee Dennis, “An interconnected set of information, able to meaningfully bridge enterprise data silos and provide a holistic view of the organization through relationships”. Forbes describes KG’s as a “database which stores information in a graphical format and, importantly, can be used to generate a graphical representation of the relationships between any of its data points”. In layman’s definition, Knowledge graphs can be defined as the graphs that store data in the form of graphs and for relating data points of different data sets it makes graphical representations [3].

In this era of information and technology, we are generating and collecting a huge amount of data. To collect, store and organize this abundant information a lot of research is going on. In 1970’s, relation database model was developed to organize raw data in a systematic manner in form of tables in rows and columns [4]. The relational database model derives relationships between different types of data sets that are available inside the different tables in form of rows and columns. With the advancements in technology and increment in raw data, Knowledge Graphs came into existence [5]. These graphs are implemented to boost the functionalities of the rational database models. Apparently, the Knowledge Graphs gained popularity due to their flexibility and ability to deal with complex data sets.

How Knowledge Graph Works?

The fundamental structure of knowledge graphs is based on the datasets from various sources that differ in various aspects like Schemas, Identities, and Contexts. The framework is designed by the schemas for KG’s, the classification of nodes are done by identities and the context generates the meaning of the relation between nodes [6]. With the help of these ingredients some major companies like Google identify the word “Apple” as a fruit or Apple Company or name of any person, Amazon, and Netflix provides a list of products and shows according to the region of interest of their customers.

Knowledge Graphs consist of three main components Nodes, Edges [7], and Labels. A node can be anything, for instance, an object, place, person, or any physical entity. An edge can be defined as relationship between set of nodes. For example, the relationship between organizations with their customers and employees. The label can reflects the meaning of relationship like a friendship relation between two or more persons. The basic representation of relation in Knowledge Graph can be seen in figure 1 below;

The implementation of the Knowledge Graphs varies from application to application [11] [12]. In formal representation, N refers to the set of Nodes, L refers to set of Labels and the cross product (N×L×N) of these sets will provide knowledge graph. “Triple” is the term that defines member of these sets.

Applications of Knowledge Graph:

1. Healthcare: By collecting and classifying links within medical research, knowledge graphs also aid the health industry. This data aids doctors in verifying diagnoses and determining treatment programs that are tailored to the individual’s requirements.
2. Entertainment: With the integration of Artificial Intelligence, knowledge graphs are implemented in the field of entertainment as well. Most of the social media and OTA platforms utilize KG’s in a very efficient way. It works on the bases of the search made by their customers. If somebody like to watch thrill and action movies or shows, he/she will get the related suggestions in future.
3. Banking: Knowledge Graphs plays a vital role in the field of finance and banking. This technology is implemented for doing online Know Your Customer (KYC) and financial fraud detection. This allow major banks to track the flow of money of their clients and unauthorized transactions.
4. Retail: The use of knowledge graph is also implemented in the field of retail also. With the help of this technology, the supplier get a crystal clear image of the demand of product is high and low. Moreover, suppliers get to know about the behavior of their customers (likes and dislikes) so that they can change their strategy to earn more profits.
5. Education: This sector is mostly effected by this technology. Nowadays, most of the big institutes are implementing knowledge graphs in their curriculum to improve the educational system.

Some useful links for the open source database used for the training purposes:

1. DBpedia :DBpedia is a community-driven initiative to derive organized material from Wikimedia projects’ resources. This organised data is similar to an open knowledge graph (OKG), which is freely accessible on the Internet.
2. Google Knowledge Graphs : The Google Knowledge Graph is a knowledge base that Google and its services utilize to improve their searching engine’s that give results by combining data from many sources.
3. Geonames : A global geographic database that “contains over 10 million geographical names including over 9 million distinct characteristics, including 2.8 million inhabited areas,” according to the website.
4. WordNet : The widely used English conceptual dictionary. “Words are organized into sets of cognitive synonyms, each of which expresses a separate notion.” Synsets are connected through conceptual-semantic and linguistic relationships.” It has 117 thousand synsets in it.
5. FactForge : FactForge is a repository of public information on individuals, organisations, and places.
6. WorldFacts : Database including information regarding nations, languages, currencies, and other similar topics. It was created by the DBPedia organization and contains data from LEXVO, the CIA World Fact book, and other sources.
7. GLEI : It stands for “Global Legal Entity Identifier”. In worldwide online source of open, unified, and high-quality judicial person reference data GLEIF makes it possible for consumers and businesses to make better informed, cost-effective, and dependable judgments about who they do business with.
8. Amazon Neptune : Neptune is built on a purpose-made, high-productive graph database engine that is geared for saving billions of relationships and swiftly accessing them and enables us to create and execute applications that interact with large, interconnected datasets
9. Cambridge Semantics : This database enables their users to speed up data integration analysis. In addition to this, it contains 40 predefined functions for business analytics.
10. Datastax : It delivers an Apache Cassandra-based distributed hybrid cloud database. Moreover it makes simple for businesses to take advantage of hybrid cloud environment.
11. Dgraph : In this software users may develop a schema, deploy it, and obtain quick database and API access without having to write any code. Dgraph allows you to pick between GraphQL and DQL, allowing even people without any prior knowledge with graph databases to get started.
12. IBM : This is an open source data base in which property graph, the product allows us to store, query, and view datasets, relationships, and attributes.
13. Mark Logic : MarkLogic has created a reputation for itself by emphasizing the unification of data silos. It’s ideal for applications, demanding large-scale heterogeneous interoperability or content distribution. ACID transactions, scalability and flexibility, along with guaranteed privacy, are all features of the database.
14. Microsioft : Regarding contemporary application development, Azure Cosmos DB offers a seamless NoSQL database service. Users may operate applications with irregular or infrequent traffic and modest performance requirements
15. Neo4j : Neo4j is a graph database that aids businesses in making sense of their data by displaying the relationships between people, processes, and systems. Neo4j saves linked data by default, making data easier to comprehend.
16. Oracle : Graph data management is automated, and modelling, evaluation, and visualization are simplified throughout the lifetime. Oracle supports both property and RDF knowledge graphs.
17. OrientDB : It is a java based NoSQL DBM system. This is an open source software that comes with multi model data base that consist of graphs, documents and relationship among entities.
18. Redis : It uses contemporary in-memory technologies like NVMe and Persistent Memory to offer installation across cloud and on-premise data centers. This software comes with real-time search engine and data modelling approaches such as streams, graphs, documents, and machine learning.
19. TigerGraph : Tiger Graph is a platform for enterprisers that provide them graph database for business use. This software provides real time analysis and huge data volumes for the applications like Internet of Things, Artificial Intelligence and machine learning. It also immune to fraud prevention and supply chain.
20. ConceptNet : ConceptNet is a free linguistic network that aims to assist computers in comprehending the meanings of words that human use. ConceptNet grew out of the MIT Media Lab’s crowdsourcing project Open Mind Common Sense, which began in 1999. Since then, it has expanded to include information from additional crowdsourcing sources, specialist materials, and purpose-driven games.
21. ImageNet : ImageNet is a picture database arranged as per the WordNet taxonomy, with hundreds of thousands of photos depicting every node of the network. The project has made significant contributions to computer vision and deep learning studies. Scholars can utilize the information for non-commercial purposes for free.

References:

1. Wang, Q., Mao, Z., Wang, B., & Guo, L. (2017). Knowledge graph embedding: A survey of approaches and applications. IEEE Transactions on Knowledge and Data Engineering, 29(12), 2724-2743.
2. Dettmers, T., Minervini, P., Stenetorp, P., & Riedel, S. (2018, April). Convolutional 2d knowledge graph embeddings. In Thirty-second AAAI conference on artificial intelligence.
3. Auer, S., Kovtun, V., Prinz, M., Kasprzik, A., Stocker, M., & Vidal, M. E. (2018, June). Towards a knowledge graph for science. In Proceedings of the 8th International Conference on Web Intelligence, Mining and Semantics (pp. 1-6).
4. Zou, X. (2020, March). A survey on application of knowledge graph. In Journal of Physics: Conference Series (Vol. 1487, No. 1, p. 012016). IOP Publishing.
5. Paulheim, H. (2017). Knowledge graph refinement: A survey of approaches and evaluation methods. Semantic web, 8(3), 489-508.
6. Guo, Q., Zhuang, F., Qin, C., Zhu, H., Xie, X., Xiong, H., & He, Q. (2020). A survey on knowledge graph-based recommender systems. IEEE Transactions on Knowledge and Data Engineering.
7. A Dahiya, B. Gupta (2021), Edge Intelligence: A New Emerging Era, Insights2Techinfo, pp.1
8. K. Yadav, M. Quamara, B. Gupta (2021), 2021 Hot Topics in Machine Learning Research, Insights2Techinfo, pp.1
9. A. Khan, K. T. Chui, D. Peraković (2021) Future Scope of AI and Machine Learning in 2022, Insights2Techinfo, pp.1
10. Muskaan Chopra, Sunil K. Singh (2021), MLOps: A New Era of DevOps, Powered by Machine Learning, Insights2Techinfo, pp. 1
11. Do, P., Phan, T. H., et al. (2021). Developing a Vietnamese tourism question answering system using knowledge graph and deep learning. Transactions on Asian and Low-Resource Language Information Processing, 20(5), 1-18.
12. Do, P., Phan, T., Le, H., et al. (2020). Building a knowledge graph by using cross-lingual transfer method and distributed MinIE algorithm on apache spark. Neural Computing and Applications, 1-17.

Cite this article as:

A. Khan, F. Colace (2021) Knowledge Graph: Applications with ML and AI and Open-Source Database Links in 2022, Insights2Techinfo

152161cookie-checkKnowledge Graph: Applications with ML and AI and Open-Source Database Links in 2022yes