60 likes | 173 Views
COCOSDA/WRITE Roadmap for Language Resources and Evaluation in a Multilingual Environment. Ontologies & Semantic Web … Report on work/goals/roadmaps of COCOSDA & Oriental COCOSDA Dafydd Gibbon?? & Khalid Choukri Slides from Chiu-yu Tseng & Shuichi Itahashi ??
E N D
COCOSDA/WRITE Roadmap for Language Resources and Evaluation in a Multilingual Environment • Ontologies & Semantic Web • … • Report on work/goals/roadmaps of COCOSDA & Oriental COCOSDA • Dafydd Gibbon?? & Khalid Choukri • Slides from Chiu-yu Tseng & Shuichi Itahashi ?? • Report on work/goals/roadmaps of WRITE& Regional updates • Nicoletta Calzolari & Steven Krauwer • Christopher Cieri • Chu-Ren Huang • Organisational structures & Coordination • Benjamin K Tsou • Peter Wittenburg • Christopher Cieri • Khalid Choukri • 9.30-11.00: Topic Areas • Introduction by Organisers: • Multimodality • Nick Campbell • Jean-Claude Martin • Stelios Piperidis • Speech-to-Speech • Gianni Lazzari • Industrial perspective • Jan Odijk • Terminology • Pierre Zweigenbaum • 11.00-11.30: coffee • 11.30-13.30: Topic Areas & Organisational structures • Minority Languages • Justus Roux • Annotation • … COCOSDA/WRITE - LREC. Genova, 2006
COCOSDA/WRITE Roadmap for Language Resources and Evaluation in a Multilingual Environment • 13.30-14.30: lunch • 14.30-16.30: Evaluation, Planning & Funding • Evaluation & Quality • Noriko Kando • Steven Krauwer • Funding Agencies • Xavier Gros (EC) • Tatiana D. Korelsky (NSF) • Joseph Mariani (France) • Asian (??) • Conclusions • Nicoletta Calzolari • Khalid Choukri • Christopher Cieri • Dafydd Gibbon • Chu-Ren Huang Report for Funding Agencies COCOSDA/WRITE - LREC. Genova, 2006
COCOSDA/WRITE Roadmap Conclusions – Thematic Priorities • Data collection, more data, many modalities, varieties, .. • languages need more coverage, need more data, costly, no single company can afford • to support also basic linguistic description & documentation • BLARK + entry-level Blarkette: be proactive • Annotation as a science: systematise • basic research on semantic annotation • Levels of metadata • Cross-mediality in multimedia content processing • Knowledge across modalities • Multilinguality, e.g. for cross-language info access • Multilinguality across culture: Pragmatic layer with cultural & social aspects • Translation technology to become an enabling technology COCOSDA/WRITE - LREC. Genova, 2006
COCOSDA/WRITE Roadmap Conclusions – Thematic Priorities • Information access: need better LRs: coverage, richness, quality • Intrinsic & extrinsic evaluation: combination needed • Subjectivity, • Community –based ontology, domain-specific, multi-faceted… • Standardisation & Sustainability of Standards • Harmonisation of existing semantic LRs • Distributed development in wiki style • Dynamic named entity • Language Identification • .. COCOSDA/WRITE - LREC. Genova, 2006
COCOSDA/WRITE Roadmap Infrastructural/Political Priorities • Quality insurance & measures, only relative to the task • Evaluation infrastructure, central neutral store for descriptors, a validation wiki; agency? • Open source platforms, sharing, wiki-mode • LR freely available • Cross-disciplinarity • Links with study of language & linguists: need of an infrastructure • Join communities, cooperation, coordination • Infrastructure/services for a broad community, also with linguists, humanities, getting LR&T out of the domain of experts COCOSDA/WRITE - LREC. Genova, 2006
COCOSDA/WRITE Roadmap Infrastructural/Political Priorities • Wiki to facilitate networking • Data, tools, standards for Linguistics, need of an infrastructure: collaborate for interchangeable data • Improve steering of research, also at international level • Long term research • Open up the field: only large scale collaborations will help, not a competitive area, strategic plans to boost research • National LRs & HLT Centers: need an umbrella • International cooperation: to be suggested to funding agencies • Specific proposals from NSF • LangNet in EU • Talk to politicians COCOSDA/WRITE - LREC. Genova, 2006