Generally speaking, Semantics is the study of language and its meaning. As a word, Semantics was first used by Michel Bréal, a French philologist in 1883, and can be used to describe how words can have different meanings for different people, due to their experiential and emotional backgrounds. A language can be a natural language, such as French, Dutch, or Hindi, or it can be an artificial language, such as a programming language for computers. Theoretical computer scientists study and develop artificial languages, while linguists study natural languages. In the last two decades, the dream of adding natural languages to computers, and having them speak as casually as humans, has developed significantly.
The use of natural languages by computers creates access to many new forms of data. In a 2015 industry report, Expert Systems had this to say:
“Semantics will be increasingly relied upon to help companies identify risks and pluck meaningful details, from even the weakest signals, from all of their information streams. As a result, operational risks are reduced, and you can anticipate market developments in near real time.”
In 1967, Robert W. Floyd wrote a paper describing the use of language semantics in computers and has been given credit for starting the field of programming language semantics. His work included the analysis and design of algorithms used to find the most efficient paths in a network, the calculation of quantiles, the parsing (or decomposing) of programming languages, and the sorting of information. Floyd described programming languages as being made up of two parts, Semantics (meaning) and Syntax (form). To be read, a computer algorithm must combine Semantics and Syntax, encoding them precisely so they can be processed automatically by the computer. Professor Donald Knuth said this about Floyd, “in the old days, programmers would just twiddle with programs till they seemed to work. His approach of marrying math with computer science was a revelation to the field.”
The concept of a global information system became technologically possible in the late 1980s. By 1985, the Internet was gaining popularity in Europe. In 1988, the very first direct IP connection between North America and Europe took place. This was quickly followed by discussions of a web-like communications and information system.
The World Wide Web and Social Media
When people originally began working on the World Wide Web, the “type” of companies with an interest in the technology determined the business direction it would take. As people experimented with the World Wide Web, the focus shifted to social interactions, and social media platforms like Facebook, LinkedIn, Google+, Instagram, Vine, Pinterest, Twitter, and Tumblr, all of which require human interaction. Because natural language has a structure humans can interpret, but machines cannot, humans had to be present to “read” a natural language’s meanings, and interact with the system.
More recently, researchers have begun merging programming languages with linguistics, allowing researchers to combine Semantics and Big Data as they strive to take Artificial Intelligence to the next level. Semantics is much more of a cognitive process than files and computer memories can claim. It is the process of designing and using a language for communicating and expressing knowledge. It can also provide a foundation for the process of thinking.
The Semantic Web versus The World Wide Web
‘The Semantic Web‘ feature article was published in May of 2001, and authored by James Hendler, Ora Lassila, and Tim Berners-Lee (Tim Berners-Lee went on to become the director of the World Wide Web Consortium, or W3C). Their paper described a new way to use and search the Internet, an added dimension full of new possibilities. While a human can read the text of an HTML web page, a computer/search engine cannot (unless tags it can read are deliberately inserted). This is because HTML is designed to store visual information, and is not written in a programming language.
The Semantic Web, as an extension of the World Wide Web, has focused on technology. The World Wide Web needs a human presence, the Semantic Web does not. It uses the “hidden,” encoded data and more recently, natural languages, to search for, compile, and organize information from the web. The Semantic Web needs a human presence only to initiate the request.
Linked Data has been a very useful aspect of the Semantic Web. It can be used to publish and share information all across the Internet. The phrase “linked open data” has been used since at least 2007, when the mailing list for “Linking Open Data” was first created. The Linking Open Data community’s goal was to extend the web with a data commons, providing information, commonly in the form of graphs, as free information.
The Internet provides a nearly infinite amount of information. Ranging from spreadsheets to images and from videos to the websites bringing it all together, links connect one site to another, and allow us to discover a constantly growing stream of information. The World Wide Web is described as a web of linked “documents,” while Linked Data describes a web of linked “data.” Linked Data gives computers a way to bring data and information together in many complex ways. This situation was made possible through the use of standardized vocabularies, and the major search engines using them. Bing, Google, and Yahoo have started using microdata formats placed within HTML documents to communicate information.
Consider these sentences in a spoken format, “fruit flies like pears” and “time flies like a butterfly.” While the sentence structure of each example is quite similar, their meanings are very different, with the words “flies” and “like” having different definitions – definitions that are determined by context. The example shows how even a remarkably simple sentence requires a significant amount of linguistic understanding.
While computers are excellent at using the simple language of mathematics, human languages are remarkably confusing in their complexity and periodic exceptions to the rules. A chess-playing program can play against, and defeat, most people in a game of chess. The same cannot be said true for trivia-playing programs. A normal child could beat such a program, because it lacks a sufficiently broad understanding of meaning, context, and subtleties of the language. This problem applies to a significant number of services and applications. Without understanding context, a search engine cannot respond with efficient results for words with multiple meanings.
Barry Zane, Vice President of Engineering for Cambridge Semantics said:
“Semantic-based technologies are the key to making data easily understandable to both humans and computers, enabling data harmonization using common business meanings.”
With the World Wide Web as a foundation, and the evolution of Semantics to include natural languages, Virtual Assistants are now becoming a reality. Apple’s Siri provides a good example of a Virtual Assistant. Siri doesn’t just retrieve information; it also helps people complete their online work more quickly and easily. Siri can interpret the spoken word, up to a point, and can also perform a variety services for the user. Initially, the tasks Siri could perform focused on the mobile Internet user. It would book restaurant reservations, check the status of a flight, or coordinate various Internet activities. Siri has now transitioned to other platforms and devices, including vehicles.
Virtual Assistants and services are beginning to exchange useful information all over the Semantic Web. Virtual Assistants, such as Google Now and Siri have initiated a wide range of cottage industry start-ups, especially those providing automated services. We are witnessing the emergence of new Semantics’ services and technologies. The merging of trends in technology and the business world are creating a new cycle of innovation affecting the way individuals and businesses perform their work, and even how Big Data is collected into useful information.
The flexibility of Virtual Assistants working across the web, and available on different devices, is an important part of the Semantic Web’s purpose. One aspect of the Semantic Web is its ability to communicate with other computers and operate without human presence. A human needs to initiate the work, but then they can do something else with their time. The use of Semantics is providing a Virtual Assistant capable of working independently and processing Big Data.