Semantic Web and Semantic Technology Trends in 2019

By on

Semantic Web and Semantic Technology TrendsWhat to expect of Semantic Web and other Semantic Technologies in 2019? Quite a bit. DATAVERSITY® engaged with leaders in the space to get their thoughts on how Semantic Technologies will have an impact on multiple areas.

As DATAVERSITY discussed last year, Knowledge Graphs have gotten a lot of attention as a backbone for Machine Learning, Deep Learning, and AI business use cases. Expect that to continue. There has been a growing appreciation of Knowledge Graphs — ontologies — to provide “an umbrella overlay for cross-walks across siloed information resources,” says Michael Bergman, senior principal at Semantic Technology consulting firm Cognonto Corporation.

“Semantic Technology approaches provide a powerful basis for Machine Learning and AI. In supervised learning, 60 to 80 percent of the effort resides in cleaning up the input data and then labeling it as the basis to train the learners. We can get these clean labels in proper training sets for ‘free’ with the use of Semtech and Knowledge Graphs.”

Among the semantic-driven AI ventures next year will be those that relate to the healthcare space, says Dr. Jans Aasman, CEO of Semantic Web technology company Franz, Inc:

“In the last two years some of the technologies were starting to get used in production,” he says. “In 2019 we will see a ramp-up of the number of AI applications that will help save lives by providing early warning signs for impending diseases. Some diseases will be predicted years in advance by using genetic patient data to understand future biological issues, like the likelihood of cancerous mutations — and start preventive therapies before the disease takes hold.”

Generally speaking, semantic ontologies or Knowledge Graphs can be an asset for AI-inspired, NLP-based conversational interactions as utilized in bots, virtual assistants, customer care, and other apps, says Tony Sarris, Founder and Principal of consulting practice N2Semantics. Such NLP interactions will remain high on enterprise agendas, but they have to make those applications smarter.

For one thing, these conversational systems:

“Need to improve their understanding of what the users really want to know or what they are trying to do,” he says. “That only comes through a better understanding of context and the deeper meaning behind the words the users are speaking or typing. Using ontologies or Knowledge Graphs as part of the application systems offers a way to help in that regard.”

Mingxi Wu, VP of Engineering at Graph Database and Graph Analytics Platform vendor TigerGraph, speaks in a similar vein.

“More and more Question Answer (QA) systems, such as domain chat bots, will rely on a semantic knowledge base to do reasoning,” he says. “The driving force would be the demand of a more intelligent QA system to go beyond a one-hop question on the Knowledge Graph and the maturity of real-time Graph Technology.”

Other experts chime in on this. “Using semantics to drive chatbots is emerging,” says Dave McComb, President of information consultancy Semantic Arts. “Getting from simple question answering to actual conversation is going to involve a number of interlocking Knowledge Graphs.”

eBay ShopBot, a voice-activated (Alexa, Google Home) system for shopping from eBay’s huge catalog by conversing with the agent already impresses Jim Webber, Chief Scientist at Graph Database vendor Neo4j:

“This is no small feat, encompassing Natural Language Processing and a huge underlying Knowledge Graph that can distinguish one customer’s needs for a sports gear bag from another customer’s needs for a laptop bag,” he says. “The level of intelligence that the system projects is really phenomenal.” Where brands like eBay go, “others will follow.”

He also envisions the use of Knowledge Graphs that focus on fluently processing natural language for a specific domain. As an example, he says that a user who wants to place a bet on a soccer match via a gaming app would idiomatically say “I want to put a lady on the Gunners for a whitewash on Saturday afternoon.” A general NLP system might be able to transcribe that, but it takes a specialist underlying graph to decode it to:

“I would like to place a bet of £5 on Arsenal to win without conceding a goal in the forthcoming game on Saturday,” he says. “What’s exciting about this is that eBay has demonstrated the feasibility of a domain-specific Knowledge (Semantic) Graph, and now that pattern can be applied in a myriad of domains like gambling, food delivery, movie streaming, and healthcare.”

If that’s not enough, how about digital immortality via AI Knowledge Graphs, where an interactive voice system will bring public figures in contact with anyone in the real world? “We’ll see the first examples of Digital Immortality in 2019 in the form of AI Digital Personas for public figures,” says Aasman, whose company is a partner in the Noam Chomsky Knowledge Graph:

“The combination of Artificial Intelligence and Semantic Knowledge Graphs will be used to transform the works of scientists, technologists, politicians, and scholars like Noam Chomsky into an interactive response system that uses the person’s actual voice to answer questions,” he comments.

“AI Digital Personas will dynamically link information from various sources — such as books, research papers, notes and media interviews — and turn the disparate information into a knowledge system that people can interact with digitally.” These AI Digital Personas could also be used while the person is still alive to broaden the accessibility of their expertise.

Knowledge Graphs and Data Governance

There’s another use case for Semantic Knowledge Graphs that didn’t get as much attention last year as it likely will in 2019, says Irene Polikoff, CEO at Semantic Data Integration vendor TopQuadrant. That is, the use of Knowledge Graphs in the enterprise will be substantially expanded in key areas including data relationship discovery and exploration, semantic interoperability, and especially Data Governance.

Indeed, enterprise-wide Knowledge Graphs that maintain models of collections of entities and their semantic types, properties, and inter-relationships can play a big role to play in addressing the issue. “They provide a uniform framework for Data Governance so that organizations can get value,” says Dave Raggett, W3C Data Activity Lead.

Being flexible, evolvable, semantic, and intelligent, Knowledge Graphs support comprehensive (top-down, bottom-up, and middle-out) Data Governance, which Polikoff describes as a lifecycle-centric asset management activity. “To understand and realize the value of data assets, it is necessary to capture information about them (their metadata) in the connected way.”

A Data Governance environment must represent assets and their role in the enterprise using an open, extensible, and “smart” approach, Polikoff says. Knowledge Graphs are foundational to Data Governance because they catalog diverse enterprise data by capturing their technical and business context and meaning through connections across all assets in the enterprise ecosystem, she says. “Knowledge Graphs are an ideal and, arguably, the only viable foundation for bridging and connecting enterprise metadata silos.”

Semantic Momentum

Amazon’s debut of Graph Database service AWS Neptune in May has turned the spotlight on Graph Databases for creating interactive graph applications. Preview customers used it to build everything from Knowledge Graphs to social networks to recommendation engines. “Amazon’s release of AWS Neptune was a very validating announcement for most of our clients, even those who are not using AWS,” McComb says.

Tom Sawyer Perspectives was chosen by the Amazon Neptune team to support the Graph Database service with an integrated solution for building applications to visualize and analyze data and connections. On the point of the future of graph visualization apps, Aasman notes that:

“Most graph visualization applications show network diagrams in only two dimensions, but it is unnatural to manipulate graphs on a flat computer screen in 2D. Modern R virtual reality will add at least two dimensions to graph visualization, which will create a more natural way to manipulate complex graphs by incorporating more depth and temporal unfolding to understand information within a time perspective.”

Given the increase in the use of Graph Database, the whole category will be prominent in 2019. Raggett points out, for instance, that W3C’s upcoming March workshop on web standardization for graph data will discuss issues including introducing standards for information exchange between the worlds of RDF and Property Graphs. “The idea is that is there is an opportunity for companies to work together on the convergence of graph query language,” he says.

That’s in the works now. “There is now a task force from several industry Graph Database vendors getting together working on a standard graph query language,” remarks Wu. “The task force includes Neo4j, Oracle, TigerGraph and more.”

Speaking of W3C standards, Shapes Constraint Language (SHACL) entered those ranks recently. With SHACL, a schema language now exists for schema-less graphs to become Knowledge Graphs. SHACL provides a set of critical capabilities in the semantic stack, says Holger Knublauch, Principal Engineer at TopQuadrant. “In terms of Knowledge Graphs, this means that SHACL can make sure that the knowledge remains consistent against multiple viewpoints; retains a good quality against validation rules and supports the inference of additional knowledge that is not explicitly stated in the data,” he says. As a standards-based way of increasing the coherence of data, it assists in Data Quality by making things explicit.

“We are seeing early interest in SHACL,” McComb says, and that interest should mount. SHACL “will change the nature of semantic modeling — back to the way it should have been. Ontologists have been letting application specific concerns creep into their models. SHACL allow a nice separation of concerns between meaning (OWL) and usage (SHACL).”

Semantic Tech Effects on Businesses and Industries

It looks like more and more industries are pursuing Semantic Technologies, often keying in on Knowledge Graphs:

“I was frankly surprised recently to find out that Uber uses Knowledge Graphs behind its applications,” Sarris says. “I’d imagine it is to better understand their customers and the things they’re doing — in other words, the user context — so they can be more proactive about meeting their customers’ needs.”

Finance is taking Knowledge Graphs to heart, too, according to Dean Allemang, CEO and Principal Consultant at Working Ontologist LLC. “The momentum in finance is unstoppable. The notion of a Knowledge Graph in a bank is going from a competitive edge to a must-have,” he says.

Allemang also sees agriculture as a new space that Semantic Technologies can be put to use in:

“It is being driven by an awareness of opportunity in global data for agriculture. The easy science has been done there, but hard science needs more data: global data, longitudinal data, well-structured data with clear identities,” Allemang says.

There are a lot of projects around the world that work on this, all of them brought together by AGROVOC (the controlled vocabulary covering all areas of interest of the food and agriculture organization of the United Nations). “They realize that they need a way to join their data together.”

But Smart City is the Semantic Tech-infused up-and-comer that he’s most excited about:

“Smart City is such a new concept to begin with, it doesn’t even quite know what it is. But it is pretty sure that data is a big part of what it means to be a smart city, and that data is heavily distributed, and needs the inclusion of a lot of meaning to make it useful,” he says. “I think Smart City is the frontier beyond the next frontier.”

The inherent power Semantic Technology will continue to drive important trends, says Bergman — trends that will create hundreds of billions of dollars in value heretofore locked up in existing information assets within and across enterprises. There’s “the use behind the scenes of RDF and SPARQL for making major repositories of valuable online data available,” he says.

Wikidata and the museum and biomedical communities are notable for serving up free data in massive amounts daily to users of all stripes:

“We don’t hear much of this use, but it is driving a number of new exciting start-ups and less-known internal enterprise uses. The volume of Wikidata SPARQL queries satisfied daily, for example, is staggering, more than 100 million per month in 2018 and growing fast. They also have a great storehouse of hundreds of off-the-shelf SPARQL queries.”

Welcome to 2019. It will be an exciting time for the Semantic Web and Semantic Technologies.


Image used under license from

Leave a Reply