«Swiss Computer Science Challenges Award» prizewinner honored / Prizewinning Swiss Visions in Information Technology

“The aim of this competition is to encourage the Swiss computer science community and in particular its younger researchers to determine and address the most significant future challenges faced by their chosen field,” says Hilty, who himself studied the subject. He also lectures at Zurich University and is the head of Empa’s «Technology and Society» Laboratory.

A further aim of the prize, according to Hilty, is to increase awareness of information technology as a scientific discipline and to create a basis for visionary research projects in the field of informatics in Switzerland. Hilty and his fellow jurors on the selection committee are delighted not just by the number of projects submitted for consideration this very first time, but also with the high quality of the work on offer.

Prizes intended to encourage the submission of more project proposals

The two other prize-winning research projects were concerned with intelligent, computer based translation systems and with previously unachieved performance improvements in informatics systems through parallel programming models and system architecture. Altogether 57 projects were submitted for consideration by a prize jury consisting of fourteen leading personalities drawn from the Swiss information technology scene under the leadership of Lorenz Hilty. The prize money of CHF 30,000 in total was made available by the Hasler Foundation.

For Juerg Kohlas, jury member and representative of the Hasler Foundation, what is now absolutely essential is that the prize winners receive further support in continuing their research work. “Sustainability is a watchword of the Hasler Foundation,” he said during the media conference at the Swiss Federal Institute of Technology, Lausanne. “This is why we have invited them to apply to us for further support for their research projects,” he continued.

The winning project: culturally adaptive software

Why do South Korean websites look different to North American ones? Would a European use a Chinese search engine? How do Africans in Rwanda work with the internet and with learning software? The user interface of a software product must be able to take into account the cultural backgrounds of its users; otherwise they will have only limited success in using it and may even reject it totally. The cultural differences relevant for software design purposes are, however, generally underestimated. The cultural background of users determines, for instance, if they prefer to be guided by the program step by step or if they would rather take the initiative themselves; whether they prefer clearly laid out functions or would rather navigate through a network structure; whether they would like to see a Spartan design or a complicated, animated interface. Software should be capable of adapting to the individual cultural background of the user, independently of the location, and therefore be internationally applicable.

The prize winning researcher, Katharina Reinecke, a doctoral student at Zurich University Institute of Informatics, convinced the jury that culturally adaptive software represents one of the fundamental challenges facing information technology. She indicated the ways in which this problem could be tackled by means of interdisciplinary research projects. The topic is very relevant in a social context, not least because it affects the ability of developing and emerging countries to take part in the information society – the key term in this context being «Digital Divide». It also impacts the maintenance of cultural diversity. In previous work for her final year undergraduate project Katharina Reinecke developed E-Learning software for agricultural advisers in Rwanda.

«Intuitive Translation Systems» – second place

Automatic translation systems have made great strides forward in recent years. Despite this, and decades of research, good quality machine translation remains elusive. This means, among other problems, that internet searches miss important documents which are written in other languages. The web is no longer just English speaking – far from it! Today Chinese, Spanish, Japanese, French, German, Arabic and Portuguese are jostling for space in the internet.

Second place was taken by a project undertaken by Davide Picca, a doctoral student at Lausanne University currently guest researcher at Columbia University in New York, and Marco Pennacchiotti, a scientific assistant working on computer linguistics at the University of Saarland in Saabruecken. They were able to convince the jury that a future challenge in the field of information technology was the combining of various different aspects of machine translation, which to date have been studied in isolation. These include semantic approaches which attempt to represent the meaning of the text to be translated, and statistical methods, which are used to derive rules from existing translations. These can then be applied to new texts. The intelligent combination of the two approaches would permit both statistically generated rules as well as prior knowledge to be combined in a single process of machine translation. A breakthrough in this field would be a great value to society in general, particularly in multilingual countries such as Switzerland.

In third place: efficient work sharing for computer processors

The rapid advance in the significance of informatics in our lives is in no small measure due to the speed at which improvements in the performance of microelectronic components have been occurring over the past decades. However this technological potential appears to be nearing exhaustion, and in order to make further sustainable increases in performance new approaches are required. Luc Blaeser, an independent software developer and consultant based in Zurich, has recognized that the productive exploitation of a large number of connected processors working in parallel remains one of the great unsolved problems in the field of information technology. Progress at three levels is necessary to find a solution: in the program models, in run-time systems and in the system architecture. Despite the ready availability of cheap networked processors, the standard programming model remains oriented towards the use of a sequentially operating machine. The capability to carry out processing with many processors working in parallel is only exploited in exceptional cases.