Dreams do come true, especially if you are Tim Berners-Lee. His vision of a “web of data” processed by machines has left the realm of science fiction and entered the real world, as new applications, tools and business models give today’s “semantic web” its much-needed edge. One EUREKA project is in the thick of it.
Tim Berners-Lee’s ambitions for the ‘world wide web’ he created went far beyond the internet that we now know and, arguably, can’t live without. He famously said that his dream is for computers to be able to analyse all data on the web – the content, links, and transactions between people and devices.
“A ‘semantic web’, which makes this possible, has yet to emerge, but when it does, the day-to-day mechanisms of trade, bureaucracy and our daily lives will be handled by machines talking to machines. The ‘intelligent agents’ people have touted for ages will finally materialise,” he was quoted as saying.
Many thought such technology was beyond us. But with developments in artificial intelligence (AI) and text mining, and the power of machine-learning now manifesting, the semantic web is very much a reality. Banking and finance, human sciences, the publishing industry, and health care have been early adopters. Engineering and manufacturing, life sciences, and even retail are tipped to be next, as AI developments speed ahead and more value is attached to high-quality information available in linked data form, including knowledge graphs, according to Dataversity’s ‘Semantic web and semantic technology trends in 2018’.
Indeed, work by the teams in the EUREKA LDL4HELTA project has capitalised on these advances by combining lexicography – creating dictionary resources – with linked data, and integrating high-quality closed data sources with open data to develop new tools and a raft of services meeting growing demand for language technologies (LT).
Viennese SME, Semantic Web Company (SWC), which is a leading provider of graph-based metadata, search, and analytic solutions, joined forces in LDL4HELTA with Tel Aviv-based SME, K Dictionaries (KD), whose diverse content coverage and know-how in lexicography provided a solid base for improved text analysis and processing for numerous languages. The two-year project was carried out as part of the EUREKA bilateral Austria-Israel R&D framework.
Innovation, ingenuity, resourcefulness … a platform for the best of these qualities to shine through
Learning to speak
To provide such solutions, a multilingual metadata and data management approach is needed, and this is where SWC’s PoolParty Semantic Suite comes into play. As PoolParty follows W3C semantic web standards like SKOS, it already integrates language-independent-based technologies.
Strong cooperation with KD was key to the project’s success and a continuing business relationship enables PoolParty to carry on “learning to speak” more and more languages with ever-increasing accuracy, according to SWC’s Martin Kaltenböck, Managing Director and CFO.
LDL4HELTA focused on ‘word-sense disambiguation’ – a crucial LT for identifying a sentence’s meaning when different interpretations are possible. By combining KD’s rich lexical resources with SWC’s linked data and machine-learning techniques, the team came up with innovative approaches, which improved products and services, and led to new ones.
The project boosted KD’s transformation from dictionary creator to a technology-driven content creator, establishing new products and services for Natural Language Processing and other language software under the new brand Lexicala, which includes an API for multi-layered, cross-lingual lexical data covering 50 languages.
The collaboration also gave SWC a clear competitive edge in the fast-growing market for multilingual (or language-neutral) search and data management. This is timely because the global language industry is predicted to reach some US$45 billion (€36 billion) by 2020, according to the Common Sense Advisory’s ‘Top 100 language service providers: 2016’ report.
Not surprisingly, SWC’s turnover has increased by around 25%, thanks in part to the cooperation, as well as intensified marketing of the improved PoolParty, and a growing number of multilingual solutions being rolled out. The company is currently serving more than 20 Fortune 500 companies, with a total of around 150 installations in Europe, North America, Australia, New Zealand and Southeast Asia, in the semantic AI field.
“We’re using KD’s data in PoolParty, and we teamed up with them – together with Madrid Polytechnic University, which was also involved in LDL4HELTA, as well as seven other partners from across Europe – to launch a new EU Horizon 2020 project (Lynx) at the end of 2017,” reveals Kaltenböck.
Banner of knowledge
“Deep knowledge about a particular field or industry, and the capacity to apply [it] in new fields, will create value in the new ecosystem. Language technology is key for such innovation, entwined as it is with human knowledge and the ability to understand and interpret meaning in speech and text,” notes LT-Innovate, one of Europe’s leading language technology industry forums.
According to Knowledge Management World, this so-called “banner of knowledge management” spans diverse territory, ranging from language-learning and augmented human-machine conversation to full AI for myriad applications, from science to sales.
SWC was recently named one of KMWorld’s top 100 companies that matter in knowledge management, citing such qualities as innovation, ingenuity, resourcefulness, usefulness, collaboration, community, and knowledge-management expertise.
The LDLHELTA project provided a platform for the best of these qualities to shine through, and the relationships formed in the project are carrying on in further research and innovation projects, as well as commercially. The EUREKA initiative thus helped its two SME partners strengthen their offer and deepen their skills and knowledge in the burgeoning language technology market.