Women in Science Profiles

Dr. Mirjam Sepesy Maucec

Research curiosity led me to an interdisciplinary field combining two sciences which at first sight seem contradicting: mathematics and linguistics. This field is computational linguistics. Its main objective is to find mathematical (mostly statistical) mechanisms for describing languages. We build language models. How do we recognize a good language model? The smaller the entropy of a text the better the model. When we define the entropy of language, we enter into the field of information theory. In the group in which I work we are building a model of the Slovenian language and are thus trying to pave its way to those world languages which have already been explored much more closely. Our language has some distinctive features which cannot be found in other languages, particularly not in non-Slavic languages. The research is not possible without language sources which are mostly developed by ourselves. A language source is a collection of information about a language (e.g. a dictionary or a corpus). We build language sources which are compiled on the basis of written texts and language sources which are derived from spoken texts. A language model is not built for its own purpose. It represents a source of information for numerous target systems. Let me mention just the largest one which is being developed in our laboratories: this is a system for translating speech to speech which includes a speech recognizer, a translator, which translates spoken text from one language into another, and a speech synthesizer. All co-workers in our group participate in the construction of this system. The international recognition of our research work is shown in many publications in scientific magazines and at international conferences. We also successfully take part in international projects. We are very pleased that the international research of world languages also includes Slovenian and that this is the result of our efforts.

My scientific career was not carefully planned – one could say that it just happened. The first turning point was the study programme I chose. Since I liked to solve mathematical problems I was looking for an occupation that would be most closely associated with this task. At home we had a computer, however, only my father used it as he worked in the field of computer science. I was not keen on computers, I often even thought they were "too stupid". I preferred to solve mathematical problems on paper which could handle anything. My father convinced me that my objection to computers was only a teenage whim and he asked me to give him just a few minutes so that he could present the computer to me in its true colours. This is when I wrote my first algorithm. It described the steering of a car at the crossroads. These few minutes which later on turned into hours were all I needed to decide without hesitation to study computer science. After finishing my studies, I stayed in the field of research because of the invitations I received from Prof. Horvat and Prof. Kačič who also accepted me into his research group.

Despite the challenges posed by research, my first and foremost commitment is the role of a mother. I have two children. Katarina is seven years old and was born in the middle of my PhD studies. Her brother Timotej is two years old. Combining scientific work and family life requires a lot of sacrifice in both fields. Only now I realize what luxury were my studies when that was my only obligation. Today I know when I have to turn off my computer and devote my time to my children. What is more, the accomplishments in scientific work are even more noteworthy when you share them with a happy family. All of this, however, would not be possible if I did not have the support of all of my family members, my co-workers and superiors.

Let me finish by saying that my path is made up of happy and bitter moments and I have learnt a lot from all of them.

Women in Science Profiles