When using online tools and platforms, EU citizens and businesses often face communication barriers which restrict access to certain goods and services, preventing citizens, businesses and governments from reaping all the benefits of the Digital Single Market, namely, access to new opportunities, reduction in differences between online and offline worlds and elimination of barriers to cross-border online activity.
Within this arena, Artificial Intelligence (AI) and Language Technology (LT) have key roles to play. Both are deeply connected to the problem of creating natural and intelligent communication between man and machine with AI focused upon more general aspects of the problem such as understanding, learning, reasoning, vision, and LT upon language-related aspects like question answering, explanation and translation.
In certain application areas, such as intelligent chatbots which engage the user in conversation, the two are so closely integrated that it is difficult to see the border between the two.
With 24 official EU languages, there are 552 possible language combinations, each of which potentially adds another barrier to bilingual interaction. Together with the explosion in documentation and online presence, it is now accepted that the problem of multilingual access cannot realistically be addressed without help from technology.
Clearly, the LT known as machine translation is of key importance but LT actually refers to a whole family of technologies and, although translation is considered the biggest revenue contributor, it is closely followed by other LTs, all of which are relevant to transparent cross-border interaction. Speech technologies (speech recognition, speaker identification and speech synthesis), for example, lower the man/machine barrier by reducing the need for a keyboard.
Similarly, other LTs such as text classification, named-entity recognition, sentiment analysis, topic identification, opinion mining are all building blocks for very powerful business decision support and policy formulation tools which can underlie ever more intelligent interaction.
The market forecasts for all forms of LT are very promising. Growing demand for enhanced customer experience, burgeoning use of smart devices and increased choice in application areas are all expected to drive growth. According to a recent report for the period 2019-2025, the global market size is expected to reach $29.5 billion by 2025, a market growth of 20.5 per cent annually.
With 24 official EU languages, there are 552 possible language combinations, each of which potentially adds another barrier to bilingual interaction- Michael Rosner
Most suppliers are quite positive when looking towards the future and expect the LT market to grow as AI will increasingly become an integral part of LT. Natural language understanding, question answering and chatbot applications in particular are often mentioned as emerging technologies to look out for and are expected to become increasingly widespread. Given these strong forecasts, one might expect to find a growth in interest for LT in relevant local market sectors.
In many countries, it is the ICT industries that host such LT development and deployment and in Malta there is certainly no shortage of ICT players. In 2017, the industry was represented by more than 300 companies, employing over 7,300 persons, and the ICT sector had grown to represent 6.6 per cent of the GVA generated by the Maltese economy.
However, in Malta the number of ICT players involved with LT is extremely limited. Prima facie, this looks like a missed opportunity. There are different reasons for this, one being the cost of the continuous innovation needed to keep up with the latest developments. Another is that LT is perceived to be peripheral to the direct sphere of competence of most ICT players and, therefore, difficult and complex to manage. Then there is the particular situation of the Maltese language, for which LT market is small, limiting business opportunities for vendors that focus uniquely on Maltese. Finally, there is the more general observation that, from a European perspective, the European LT industry is fragmented, dominated by hundreds of SMEs divided by nation states, languages, domains and sectors, significantly holding back its impact.
The European Language Grid (ELG) project is a scalable cloud platform, providing access to hundreds of commercial and non-commercial resources for all European languages. Unlike other repositories, which mainly contain static datasets, ELG also includes tools and services can be run as cloud-based services provided that the tools implement a standard API.
In a nutshell, ELG addresses many of the problems preventing potential LT players from participating. Its functionality directly confronts the fragmentation problem by facilitating the sharing, linking and reuse of existing components. When fully operational, ELG will enable the European LT community to upload their technologies and data sets in an easy and efficient way, to deploy them through the grid and to connect with other resources.
The ELG will, thus, boost the Multilingual Digital Single Market towards a thriving European LT community, creating new jobs and opportunities and also addressing the threat of digital language extinction.
A half-day online seminar will take place on May 20 to present ELG and examine some key issues concerning its potential to kick-start the further take-up of LT in Malta. It will include short presentations and demonstrations by international members of the project consortium as well as by local representatives from industry, government and academia.
Participation is free but pre-registration is required. For further information send an e-mail to Michael Rosner (mike.rosner@um.edu.mt).
Michael Rosner, senior lecturer, Department of Artificial Intelligence