A White Paper on language technology support for the Maltese language has concluded that research and development in this area is in an “immature state”, despite the “significant efforts made by a small number of well-qualified people”. Furthermore, efforts that have been made are “uncoordinated, short term, and fragmentary”, resulting in slower progress.
“Although there are signs that the situation is improving, investment in language technology remains at a low level, and as a result, despite modest local achievements, the effort is fragmentary, both in terms of coverage of different areas, and in terms of sustainability of research: there have been too many projects involving just one area, just one researcher, and just one or two years. The collective efforts don’t add up as they should,” the report said.
The document was published by the METANET4U project, an EU initiative with the aim of supporting language technology for European languages and multilingualism. Such technology includes speech synthesisers, text-to-speech engines and translation software.
It is part of a constellation of EU projects that together are contributing towards the realisation of META, the Multilingual Europe Technology Alliance, dedicated to building the technological foundations of a multilingual European information society.
The University of Malta is the Maltese partner, which joined a consortium of eight partners on the project that started in February 2011 and is spread over two years. Indeed the White Paper was authored by Michael Rosner and Jan Joachimsen, both from university.
A key part of the strategy for developing multilingual language technology is to assess the status of every language in the consortium. A language White Paper series ‘Languages in the European Information Society’ has been commissioned and each paper reports on the state of each European language with respect to language technology and explains urgent needs and current opportunities. The series covers all official European languages.
The White Paper on the status of the Maltese language takes stock of the situation in every aspect of language technology.
Prototype speech recognition tools have been developed at the University of Malta but are not readily available at the time of writing. However, a government-funded speech engine should yield a working speech synthesiser by 2013. While this is a very positive development, it is highly focused on the synthesis side of speech. Almost now work on speech recognition is planned at this stage.
With respect to resources, the situation is a little more structured. There is the Maltese Language Resource Server (MLRS), a project coordinated by the Institute of Linguistics and the Department of Intelligent Computer Systems at the University of Malta, whose primary aim is to create and make available basic language resources for the Maltese language, as well as develop natural language processing tools for Maltese. The project has benefited from funding from the Malta Council for Science and Technology (MCST) and the University of Malta.
The document notes that the existing MLRS corpus is predominantly textual and monolingual. It is also somewhat non-representative: there is no shortage of legalistic material, but there is currently a lack of academic text and works of fiction.
As things stand, these materials can only be searched and analysed through the server and cannot be accessed directly. The reasons are legalistic. With access restricted in this way, the complications of IPR and copyright have been neatly sidestepped. The price is that these complications will eventually have to be confronted in the future, and in fact META is in the process of formulating a set of licence agreements to suit the distribution of resources, like MLRS.
“In this report, we have tried to convey the paradoxical state of Maltese language technology. The paradox arises because there are significant efforts made by a small number of well-qualified people across a spectrum of language technology-related activities to improve the state of the art, whether this be in terms of tools, or resources, or both.
“It is also clear that within the wider context of educational, commercial and cultural activities in the country, there is a place for language technology to make an important contribution. The problem is that efforts that have been made are uncoordinated, short term, and fragmentary, so progress is slower than it has to be,” the document warned.
“Sustained and directed coordination of effort is, in our opinion, the only way in which the benefits of language technology for Maltese will be realised in a reasonable time. We believe that even in a country as small as Malta, the work needs to be shared out amongst different stakeholders.
“We must arrive at a workable roadmap via a localised version of the tripartite division of labour advocated by META: identification of a community with a shared vision; extension of an infrastructure to facilitate the sharing of resources, and reinforcement of connections between language technology and neighbouring fields of research and development,” the White Paper concluded.
The META-NET White Paper for Maltese can be downloaded freely from the META-NET website: www.meta-net.eu/blog/whitepapers.