Cross-lingual search in medical texts

Speaker:
Shadi Saleh
Abstract:
In this talk, we will present the problem of Cross-lingual Information Retrieval (CLIR) in medical texts and show our contribution to the research in this area. The goal of this task is to find documents in a given language relevant to queries formulated in a different language. Our experiments are conducted using the CLEF eHealth collection containing medical articles in English and queries in Czech, French, German, Hungarian, Polish, Spanish, Swedish. In our work we explore the two main approaches to this task: Document Translation and Query Translation. We exploit both the Statistical Machine and Neural Machine Translation paradigms, and methods for query translation reranking and expansion.  
Length:
01:07:11
Date:
02/03/2020
views: 310

Images:
Attachments: (video, slides, etc.)
62 MB
209 downloads
699 MB
222 downloads
179 MB
211 downloads
378 MB
311 downloads