Methodology and software for semi-automatic multi-domain annotations: annotating, exploring and sharing data

Brigitte Bigi

Grégoire de Montcheuil

Roma - September, 14th, 2015

ORTOLANG is a beneficiary of French State support via its «Investissements d’avenir» program (ANR-11-EQPX-0032)

Presenters

Brigitte Bigi

Brigitte Bigi

Computer Scientist working in the field of corpora and annotations:
- formalization/constitution of corpora,
- automatic annotation (mainly at the phonetic level, also at the discourse level),
- multimodality (annotation, exploration, extraction of annotated data),
- multilinguality (methods and algorithms).

SPPAS

Grégoire de Montcheuil

Research Engineer at the CNRS - Ortolang, Laboratoire Parole et Langage, Aix-en-Provence, France
Computer Scientist working in the field of Natural Language Processing:
- written and spoken French morphosyntactic analisys (MarsaTag)
- treebank creation (4Couv) and analisys (MarsaGram)
and in the past:
- information extraction (resume parser)
- document classification (medical record, competitive intelligence)
- word sense disambiguation

Content

Softwares
- Selection
- Examples : Praat, Elan, SPPAS
An annotation workflow
- Recording
- IPU & Transcription
- Phonetics: tokens, phonemes, syllables
- Morphosyntax: POS, lemma, chunk
- Discourse: repetitions
- Prosody: Momel & INTSINT
- Gestures
Exploring
Sharing
- Why ? How ?
- Data Repositories
- Metadata