University of Helsinki
Department of General Linguistics
21 August 2001

Language Technology: Evaluation of Education, 2001

1. Background

A curriculum for language technology in Finland was established in autumn 1988 and the first courses were given in autumn 1989. Starting from mid 1990 there have been permanent teaching resources consisting of one professor. Since summer 2000 more resources are available thanks to a special three year funding by the Ministry of Education. The Faculty of Arts at the University of Helsinki still is the only faculty in Finland licensed to provide advanced studies (syventävät opinnot) in this subject.

During the initial years, the discipline was called "computational linguistics" ("tietokonelingvistiikka") and was renamed in 1999 as "language technology" ("kieliteknologia") to reflect a new and broader orientation.

The first language technology students were enrolled in 1988, and the number of new students has varied between 1 (1995) and 9 (2000) with the average of 5; the number of freshman students in autumn 2001 will be 14.

Out of the 68 students registered having language technology as their major subject, 57 % were men and 43 % women in spring 2000. The number of students registered with the discipline as a minor subject was 32, half of which are women.

There were also four postgraduate students and five foreign exchange students enrolled during the academic year 2000-2001. Some 202 students from other disciplines (including some from LANGNET, Graduate School for Language Studies) have taken language technology courses as part of their studies at the university. Six students from other universities participated in the courses during autumn 2000 and spring 2001 as KIT-students (Finnish Network for Language Technology Studies).

Two PhD dissertations have been defended within the discipline during 1999 and 2000 and four students have completed their Masters Degrees in computational linguistics between 1995 and 1999. Since 1995, Four Bachelor Degrees have been completed in computational linguistics and two in language technology.

1.1 Planning, aims, contents

The Programme for the Development of Teaching and Studies 2001-2003 (as approved by the Senate of the University of Helsinki on 17 May 2000) lists objectives of teaching. These objectives are enumerated below, followed by comments on how these objectives are reflected in the specification of the curricula in language technology:

Objective: Profound competence and expertise in the student's field form the core of each degree.

Objective: Skills related to the production of new information and the ability to improve one's learning skills are qualities that are essential for academic experts.

Objective: Degree programmes of high quality also contribute to the general qualities of an academically educated person by emphasising cooperative skills, communication skills, knowledge of different cultures and openness to international interaction.

Objective: The University of Helsinki will have a stimulating atmosphere for teaching and learning.

Objective: Research forms the basis for all other activities of the University.

Objective: The trend in the development of teaching and learning is away from a teacher-oriented approach toward one which is student-oriented.

The following goals and considerations can probably be derived from the law for universities (yliopistolaki) rather than from the development plan mentioned above. (§ 4. "Yliopistojen tehtävänä on ... sekä kasvattaa nuorisoa palvelemaan isänmaata ja ihmiskuntaa." i.e. the task of the universities is to raise youth to serve the nation and mankind.)

During the early years, the discipline aimed at educating a small number of high quality MAs to be future researchers, as there was no prospect of wider commercial or other applications of the theories. However, the methods and theories were expected to become important sooner or later. In the late 1990s, a substantial language technology industry emerged worldwide and in Finland. This led to a radical reorientation of the discipline and the aims of the education.

The present aim of the education is to produce a sufficient number of language technology MAs (FM) and PhDs (FT) along with minors in language technology combined with various other degrees. The present scheme of education is multidisciplinary in two senses: (1) students in a host of other disciplines participate in our courses, and (2) our students combine studies from other disciplines in their degrees.

The education has two aims: to provide the students with sufficient skills and knowledge for their future jobs outside the university, and for a smaller part, to produce high quality researchers. The former objective has been combined with practical periods of training (apprenticeships) in commercial companies or other application areas. In addition to the training included in the curricula, many students finance their studies by part time work in language technology companies.

Apprenticeship or work experience have been included in the formal requirements of the intermediate and the advanced studies in language technology.

The curricula and requirements of the studies, introductory studies (perusopinnot), intermediate studies (aineopinnot) and advanced studies (syventävät opinnot) have been planned in cooperation with our students and the KIT network and they are consistent among several universities (at the level of basic studies).

One particular aim has been to increase the interaction and common studies between humanists and technologists.

1.2 Implementation

The education has been international to some extent during the years, but this aspect will be more important in the near future. There have been occasional visiting students from Germany, Sweden, Denmark and Estonia. The number of our own students sent abroad within exchange programs has been smaller. We expect that the emergence of the Nordic graduate studies within the Nordic language technology research program will radically increase both kinds of exchange. The department's appointment of a lecturer (Graham Wilcock) from Britain will, in part, help this process of internationalization. Cooperation with various language area studies includes African studies (Arvi Hurskainen, several joint international courses).

There have been visiting scholars from the U.S. (Ruth and Alton Sanders both 1 yr, Roman Yangarber), Norway (Torbjørn Nordgård), Denmark (Thomas Bilgram), Sweden (Jussi Karlgren, Björn Gambäck both 2 yrs, Christer Samuelsson). Especially within computational morphology and surface syntactic analysis, there has been substantial cooperation with scholars of common and exotic languages: Udmurt, Mari, Sanskrit, Accadian, Swahili, Kwanyama, Basque, Inuit, etc.

Students are encouraged to participate in international conferences like NODALIDA (Nordiska datalingvistikdagarna), COLING and summer schools like ESSLLI. Accepted conference papers are counted as credits in advanced studies.

1.3 Evaluation and development

The number of completed Master's degrees has been remarkably low. This reflects the small number of students on one hand, but two other factors can be identified:

  1. Many students are working full or part time in industry, and their studies take longer. Their degrees will be completed later but often with distinction.
  2. The average number of new students accepted in years 1988 - 1999 was only 4 per year.
  3. Especially during the early years, some students apparently failed to identify themselves with the discipline because no true courses in the subject were scheduled until the third or fourth semester.

Nothing but an economic recession will affect the first cause. The second one has been recognized and the studies reorganized accordingly. Many more students change their major subject from other disciplines to language technology than the other way. The national KIT network has also greatly increased the visibility and attraction of our subject.

The number of doctoral students and completed PhD degrees is satisfactory.

Delays in publishing results of course examinations have been unusually prominent. No acceptable excuse can be provided for this.

The overall picture conveyed by the survey

Some 30 students returned the questionnaire, nine of them had first started in some other subject and then moved over to language technology. Most of these 30 students have studied less than 3 years.

The ratings given as numbers in the following are on a scale from 1 ... 10, where 1 is full disagreement or the lowest rating, and 10 is complete agreement or the highest rating. Preference orders are indicated with Roman numerals.

Entrance examination

Comment: The entrance examination only tests logical reasoning and general ability to understand languages or their structures.

Teaching and assessment

The students preferred various forms of completing courses in the following order (best first): (i) exercises/project works, (ii) lectures, (iii) study groups (opintopiiri), (iv) essays, (v) book examinations, (vi) study diary (opintopäiväkirja). The forms currently used were considered satisfactory (6).

Areas in which the students wished more courses: (i) speech technology, (ii) language learning, (iii) other programming (than Perl), (iv) Perl programming.

Supervision and assistance

Skills for successful employment

Career prospects

Comment: The survey failed to ask whether the students were currently full/part time employed. This information would be useful. The general understanding is that the demand for competent people on this field greatly exceeds the supply by a factor of (up to) 10.

Wishes and opinions concerning the department

The areas of the department where the most urgent need for improvement was indicated were: (i) improving the library, (ii) improving the accessibility of lecture material, (iii) getting more premises for teaching.

Note: the staff of the department was split in three separate buildings at the time of the survey.

1.4 Future outlooks and plans

This was the first survey of education in language technology, so comparisons are not possible yet.

The Ministry of Education has granted funding for a new national graduate school in language technology starting form January 2002 for five doctoral students. This will be a multidisciplinary effort parallel to the KIT network, and a great improvement in the graduate studies in language technology. In addition to this, the Nordic language technology research program offers lots of doctoral studies for these purposes.

At least one completed new PhD (FT) degree in language technology per two years is expected during the next few years until the organized graduate school activities start to produce more.

The number of MA (FM) degrees completed each year is expected to increase to 10-15 per year in some 3-5 years as the students of the network start to graduate. The number of completed 20 credit minors (basic studies) is expected to increase up to 50 in the some 3 years, if the KIT network is successful.

2 Current themes

The KIT network has been granted adequate funding but it also has very serious challenges and commitments in order to succeed as a national educational network. The most difficult task is to provide courses and education which would be truly accessible for students at various universities, some at great distance.

Both professors (Koskenniemi and Carlson) and the coordinator (Westerlund) have participated in several courses on new teaching methods, and our new lecturer (Wilcock) has substantial experience of distance learning methods in Britain. In addition to conventional methods using the world wide web, digital video equipment has been acquired by the department. In addition, a full time professional for assisting in the production of multimedia material for courses will be recruited. These activities will take up a substantial part of our time and resources.

Some other activities are ongoing. Efforts to create widely materials and corpora which can be widely used in research and education have been initiated by the staff. One has resulted in a Finnish Language Bank located at the national centre CSC and used by most Finnish universities. This collection has some 200 million words of Finnish texts. Another effort aims at building a similar archive for Finnish speech.

3 How this report came about

In a meeting between the teachers and representatives of students, the principles of conducting this survey were outlined. Student Tero Aalto and coordinator Hanna Westerlund were responsible for carrying out the questionnaire and providing the basic data. A questionnaire was prepared and distributed to students and the students were able to answer either through a web page or with conventional paper. Using either way, the answers were anonymous. Kimmo Koskenniemi wrote most of the text and several others helped with their comments.

Strengths, weaknesses, threats and opportunities