Dr Serge Sharoff

Serge Sharoff

Position

Turing Fellow

Partner Institution

Bio

Serge Sharoff started working in the field of Artificial Intelligence since 1990s. He obtained his PhD in 1997 from the Moscow State University with the topic on Natural Language Processing tools for Information Extraction. Since then he worked at the Russian Research Institute for Artificial Intelligence, University of Bielefeld (Germany) and finally he has been working at the University of Leeds since 2003.

Research interests

Sharoff's research interests are related to three domains: linguistics, computer science and cognitive science.
Probably the most interesting bit in his recent research concerns digital curation of representative corpora automatically collected from the Web. This implies their annotation in terms of genres, domains or morphosyntactic categories. The current set of resources includes very large corpora for Arabic, Chinese, English, French, German, Italian, Polish, Portuguese, Russian and Spanish. Examples of his recent projects are in the areas automatic text classification, language adaptation and non-parallel resources for translation. Examples of his recent projects are in the areas automatic text classification, language adaptation and non-parallel resources for translation.

On a more general level he is interested in interpretability of linguistic models. For example, neural networks produce empirically efficient prediction models, especially the modern setup which is based on fine-tuning pre-trained transformer models, such as BERT. However, they act as a blackbox, as it is difficult to determine why a model with a given set of training parameters produced a specific prediction. His research is aimed at better understanding of the reasons for predictions and the differences between the humans and the computers in interpreting linguistic tasks.