Indicative areas for Bachelor / Master / PhD projects

Language technologies for contrastive linguistic studies, e.g., comparing collocations across languages; properties of collocations in translated language; negative polar questions in parallel corpora; syntagmatic fixedness of multi-word expressions crosslinguistically.

Information structure and grammatical interfaces, e.g., topicalisation strategies in contrast; word order phenomena.

Computational modelling of Slavic languages, in particular: grammar engineering, treebanking and resource development; interface between treebanking and practical grammar engineering; cross-linguistic and cross-formalism generalisations and comparisons.

Slavic translational equivalents of German composita: implications for grammar induction and grammar engineering, e.g., a corpus-based study Russian - German; a corpus-based study Bulgarian - German; a corpus-based study Polish - German; a corpus-based study Czech - German, etc.

Linguistically informed NLP, e.g., for the purposes of information retrieval and question answering; machine translation; dialogue systems; computer-assisted language learning; processing of structured proper nouns, named entities with internal structure, multiword expressions, collocations.

Sample Topics

    1. Developing a (Slavic) component for the DFKI IE system SProUT

• Russian version of the information extraction system SProUT
• Polish version of the information extraction system SProUT
• Bulgarian version of the information extraction system SProUT

    2. Developing a (Slavic) resource grammar for DELPH-IN

• Construction of Russian HPSG in the DELPH-IN environment combining the Grammar Matrix and a Slavic Core Grammar with corpus-based grammar elaboration; exploiting the Russian National Corpus (RNC) in Russian Ressource Grammar (RRG) engineering.
• Construction of Bulgarian HPSG in the DELPH-IN environment combining the Grammar Matrix and a Slavic Core Grammar with corpus-based grammar elaboration; exploiting the BulTreeBank in Bulgarian Ressource Grammar (BRG) engineering.
• Construction of Polish HPSG in the DELPH-IN environment combining the Grammar Matrix and a Slavic Core Grammar with corpus-based grammar elaboration; exploiting the IPI PAN Corpus in Polish Ressource Grammar (PRG) engineering.
• Construction of Czech HPSG in the DELPH-IN environment combining the Grammar Matrix and a Slavic Core Grammar with corpus-based grammar elaboration; exploiting the Prague Dependency Treebank (PDT) in Czech Ressource Grammar (CzRG) engineering.

    3. Corpora and Grammatical Theory

• Translational equivalents of German composita: implications for grammar induction and grammar engineering.
• Linguistically informed processing of named entities with internal structure.
• Corpus-aided building of a multi-purpose valence dictionary.

Supervsed theses