Towards the Standardisation of Curricula in Computational Linguistics in Europe

Andy Way,
School of Computer Applications,
Dublin City University,
Dublin 9, Ireland.

Phone: +353-1-7045644
Fax: +353-1-7045442
away@compapp.dcu.ie
http://www.compapp.dcu.ie/~away/away.html


From its inception in 1993, the Erasmus ICP in the area of NLP existed primarily to facilitate student mobility in the Computational Linguistics fraternity within Europe, but at the same time a common interest was pursued in Curriculum Development within the same field. Now that under Socrates the notion of ICP's is defunct, the innovative work done on curriculum development has continued under the ageis of the ACO*HUM thematic network project, (http://www.uib.no/acohum ) and these initiatives are reported on here.

The joint curriculum development effort aims at developing a modular curriculum in NLP consisting of a basic module of courses in CL to be implemented in each participating university (17 in total), and several specialisation modules, different for each university, depending on their expertise, taking the entrance condition as nothing more than the basic module. Students travelling in the network would then be able to begin their NLP studies locally and, depending on their area of interest, take any specialisation module at any site.

Given this, our ultimate desideratum was harmonisation of curricula. This was seen as a four-stage process, namely:

        1. Individual recognition of courses on a case by case basis.
        2. Mutual agreement on lists of periodically offered courses to be recognised.
        3. Mutual recognition of complete components, divisions or stages in the curriculum, leading to joint development and co-ordination of different administrative requirements.
        4. Mutual recognition of complete curricula, including all courses, exams and degrees,   leading (ultimately) to treaties between universities for mutual recognition.

In the first phase, the undergraduate curricula of DCU, Saarbruecken and Tilburg were standardised. Then in 1994-95, the contents of the basic module were defined and agreed upon by a subset of the partners in the ICP. The proposal was discussed, adapted, and agreed upon later by a larger subset of the ICP, with the result that introductory modules were developed in six main areas, namely:

        1. Basic Linguistics
        2. Computational Linguistics
        3. Symbolic Computation
        4. Mathematics/Logic
        5. Grammar Formalisms
        6. Artificial Intelligence

with Empirical Methods added at a later date as a further core component. It is envisaged that these courses could be taken by students from any university at any other.

Of course, as well as these core components, there are a number of specialisation modules, including Vision/Robotics, Philosophy, Corpus Analysis, Psycholinguistics, Sociolinguistics, Speech, Machine Translation, and Information Retrieval, to name but a few. These obviously reflect the different expertise of each site, and make possible a wider choice of subjects to all students. Other modules which merit inclusion are language-dependent NLP (i.e. French, Dutch, German, English NLP etc), together with application-specific NLP modules such as Dialogue systems and Software Localisation.

The envisaged modus operandi of the programme is structured as follows:

        1. Full ECTS compatibility in the credit system.
        2. European dimension in language-specific modules.
        3. Introduce ODL teaching techniques.
        4. Develop guidelines for sites starting with NLP curricula.

Of these, all have been attempted, and some have been achieved in full.Three years ago, the contents of the basic module was adapted to run on the lines of ECTS (i.e. 6 modules per semester, 5 credits per module). Most of the additional material has also graduated to this level of detail, and has been completed at a number of sites, primarily those having an undergraduate degree in this area.One of our partners has implemented a long-distance introductory course on NLP basedon the results of the joint curriculum development, and the ICP is developing courseware adapted specifically for the web.

The implementation of this curriculum development effort is an incremental effort, and work continues within ACO*HUM as well as in conjunction with ELSNET. We are also interested in broadening the focus of such work in areas such as postgraduate level certification, investigating further the use of innovative teaching methodologies, broadening the user base, and having our guidelines tested by other institutions setting up similar degree schemes.

Of course, while we in the network are operating within the broad area of the humanities, some of the institutions involved come from a different background altogether, so there are local differences to overcome in attempting such a process of integration. As these courses could be embedded in different types of curricula (e.g. languages, linguistics, computer science, psychology etc), the exact number of courses and the place in the curriculum where they are taught, therefore, naturally differ from site to site. Nevertheless, we consider that our efforts are contributing effectively to the internationalisation of  higher education in Europe, whilst maintaining academic, cultural and professional distinctions wherethis is desirable.