<> "The repository administrator has not yet configured an RDF license."^^ . <> . . "Translation-based Ranking in Cross-Language Information Retrieval"^^ . "Today's amount of user-generated, multilingual textual data generates the necessity for information processing\r\nsystems, where cross-linguality, i.e the ability to work on more than one\r\nlanguage, is fully integrated into the underlying models. In the particular\r\ncontext of Information Retrieval (IR), this amounts to rank and retrieve relevant\r\ndocuments from a large repository in language A, given a user's information\r\nneed expressed in a query in language B. This kind of application is commonly\r\ntermed a Cross-Language Information Retrieval (CLIR) system. Such\r\nCLIR systems typically involve a translation component of varying complexity,\r\nwhich is responsible for translating the user input into the document\r\nlanguage. Using query translations from modern, phrase-based Statistical\r\nMachine Translation (SMT) systems, and subsequently retrieving monolingually\r\nis thus a straightforward choice. However, the amount of work committed to\r\nintegrate such SMT models into CLIR, or even jointly model translation and\r\nretrieval, is rather small.\r\n\r\nIn this thesis, I focus on the shared aspect of ranking in translation-based\r\nCLIR: Both, translation and retrieval models, induce rankings over a set of\r\ncandidate structures through assignment of scores. The subject of this thesis\r\nis to exploit this commonality in three different ranking tasks: (1) \"Mate-ranking\" refers to the\r\ntask of mining comparable data for SMT domain adaptation through translation-based\r\nCLIR. \"Cross-lingual mates\" are direct or close translations of the query.\r\nI will show that such a CLIR system is able to find\r\nin-domain comparable data from noisy user-generated corpora and improves\r\nin-domain translation performance of an SMT system. Conversely, the CLIR system\r\nrelies itself on a translation model that is tailored for retrieval. This\r\nleads to the second direction of research, in which I develop two ways to\r\noptimize an SMT model for retrieval, namely (2) by SMT parameter optimization\r\ntowards a retrieval objective (\"translation ranking\"), and (3) by presenting\r\na joint model of translation and retrieval for \"document ranking\". The latter\r\nabandons the common architecture of modeling both components separately. The\r\nformer task refers to optimizing for preference of\r\ntranslation candidates that work well for retrieval. In the core task of \"document ranking\" for CLIR, I present a model that directly ranks documents using an SMT decoder. I present substantial improvements\r\nover state-of-the-art translation-based CLIR baseline systems, indicating that\r\na joint model of translation and retrieval is a promising direction of\r\nresearch in the field of CLIR."^^ . "2015" . . . . . . . "Felix"^^ . "Hieber"^^ . "Felix Hieber"^^ . . . . . . "Translation-based Ranking in Cross-Language Information Retrieval (PDF)"^^ . . . "thesis.pdf"^^ . . . "Translation-based Ranking in Cross-Language Information Retrieval (Other)"^^ . . . . . . "lightbox.jpg"^^ . . . "Translation-based Ranking in Cross-Language Information Retrieval (Other)"^^ . . . . . . "preview.jpg"^^ . . . "Translation-based Ranking in Cross-Language Information Retrieval (Other)"^^ . . . . . . "medium.jpg"^^ . . . "Translation-based Ranking in Cross-Language Information Retrieval (Other)"^^ . . . . . . "small.jpg"^^ . . . "Translation-based Ranking in Cross-Language Information Retrieval (Other)"^^ . . . . . . "indexcodes.txt"^^ . . "HTML Summary of #18696 \n\nTranslation-based Ranking in Cross-Language Information Retrieval\n\n" . "text/html" . . . "004 Informatik"@de . "004 Data processing Computer science"@en . .