next up previous contents
Next: Basic concepts Up: Background to EAGLES Previous: Background to EAGLES

What is EAGLES?

The Expert Advisory Group on Language Engineering Standards is an initiative of the European Commission (EC). It was launched, in February 1993, within EC DG XIII's Linguistic Research and Engineering programme, as project LRE-61-100. The aim of EAGLES is to accelerate the provision of standards for:

Leading industrial and academic players in the Language Engineering field have actively participated in the definition of this initiative and have lent invaluable support to its execution. Moreover, the initiative is a direct result of a series of recommendations made to the EC over several years. Reports from recent EC Language Engineering strategy committees have strongly endorsed standardisation efforts in Language Engineering:

The importance of working towards standards and protocols is well recognised, with the establishment of the EAGLES project that brings together senior representatives of all the major speech and language development projects in Europe.
(Oakley, 1993, 71)

Moreover, there is a recognition that standardisation work is not only important, but is a necessary component of any strategic programme to create a coherent market, which demands sustained effort and investment:

The EAGLES initiative to encourage the development of standards and protocols for all aspects of language and speech engineering should be continued as a priority.(Oakley, 1993, 73)

It is important to note that the work of EAGLES must be seen in a long-term perspective. This is especially true for any attempt aiming at standardisation in terms of International Standards. EAGLES will not and cannot result in such Standards: this is the preserve of the International Organization for Standardization (ISO). However, the road to standardisation is a long one. Moreover, successful Standards are those which respond to commonly perceived needs or aid in overcoming common problems. In terms of offering workable, compromise solutions, they must be based on some solid platform of accepted facts and acceptable practices.

With Language Engineering, we have a relatively immature field. It is not possible to propose Standards across the board in such a circumstance. It is not even appropriate to consider Standards for narrowly defined areas of Language Engineering, insofar as such areas involve interpretation of abstract knowledge about language: research is still confronted with hard problems about the formal nature of language in general and languages in particular, as used in many varied communicative situations, by people from different linguistic backgrounds, and has not arrived at any deep understanding or common approach (Bates and Weischedel, 1993). Equally, there are hard problems to be met at the computational level, when one tries to find flexible, expressive computational means to support discovery, formulation and formalisation of knowledge about language, means to constrain general purpose machines to interpret and apply such knowledge as language experts to language processing tasks and means for users of resultant systems to participate appropriately in these tasks. One should also not forget that, whereas in a text-based world, messages conviently appear in a recognisable and tractable (although not necessarily understandable) form, in a speech-based world there are severe problems in simply knowing what parts of a complex acoustic signal, possibly produced by many sources of air vibration, convey a message. The degree of variation in speech behaviour from individual to individual and even for one individual over time is moreover immense, which compounds the problem. Thus, looked at from the perspective of how close one is to achieving fully robust, flexible, accurate language processing technology, the field is indeed immature and not at all ready to propose Standards of the International Standard variety, except in areas associated largely with the physical handling of data.

However, the field is nevertheless highly active and has met with modest success in several areas. These successes have indeed led to a welcome transfer of knowledge to industry, for example through EC sponsored programmes such as ESPRIT, LRE, MLAP and LE. There is a growing number of companies specialising in natural language processing and/or speech processing, especially through European initiatives.

However, although industry is capable of producing language engineering applications, the widespread development and adoption of such applications is threatened due to the lack of standards in the domain.

EAGLES was set up to ameliorate this situation, through bringing together representatives of major collaborative European R&D projects in relevant areas, to determine which aspects of our field are open to short-term de facto standardisation and to encourage the development of such standards for the benefit of consumers and producers of language technology. This work is being conducted with a view to providing the foundation for any future recommendations for International Standards that may be formulated under the aegis of ISO.

At the very least, EAGLES work towards de facto standards will allow the field to establish broad consensus on key issues, providing thus an opportunity for consolidation and a basis for technological advance and expansion of knowledge.

next up previous contents
Next: Basic concepts Up: Background to EAGLES Previous: Background to EAGLES