Automatic Summarization for All-Digital ERA – ASADERA
Communication Stream analysis has become a crucial strategic issue. Quantity as well as quality of information to analyze increase every day for the all-digital era. So summarization and filtering techniques have become decisive. Information sources diversity forces new processings: multilingual, cross-lingual, and new needs emerge: opinion summarization, comparative summarization, audio and video documents... ASADERA project fits into this context and proposes several tasks linked to automatic summarization. The first aims at developping an open automatic summarization platform that includes several state-of-the-art methods. Such a Platform does not yet exist and its development at LIASD laboratory (University Paris 8) sets out to improve LIASD's visibility at an international level. The second and main task focuses on comparative cross-lingual and multilingual summarization (CLMS). CLMS consists in generating a summary in a target language that sums up the différences in the processing of a same topic between documents written in different languages. This kind of summay is new, as it has never been studied, and can have important impact on social sciences, where comparative analysis of documents is a recurrent task. The third task aims at exploring the use of different kinds of resources, acquired automatically or built manually for summarizing specialized documents. Scientific risks associated to this project are limited, but the main task poses the problem of how to extract knowledge from automatically translated and so noisy documents.
Aurélien Bossard, whose research has mainly focused on automatic summarization (multidocument, opinion summarization...) and has also worked in the automatic translation field when he tackled the named entity annotation rules transfer in multilingual context, will coordinate ASADERA project. This explains the multilingual aspect of the main task. ASADERA project has been thought to federate researchers based at the technological universitary institute of Montreuil (93). These researchers have complementary profiles: natural language processing, automatic summarization, information retrieval, knowledge management and reasoning, fuzzy logic... ASADERA project's goal is to develop a new research field (automatic summarization) in a way that will make use of existing and living research fields Inside LIASD's branch based at Montreuil.
Project coordination
Aurélien Bossard (Laboratoire d'informatique avancée de Saint-Denis)
The author of this summary is the project coordinator, who is responsible for the content of this summary. The ANR declines any responsibility as for its contents.
Partner
LIASD Laboratoire d'informatique avancée de Saint-Denis
Help of the ANR 163,266 euros
Beginning and duration of the scientific project:
October 2016
- 36 Months