Helps users annotate large text corpora through a web interface. To improve analysis and annotation of those corpora, bilingual linguistic, lexical and semantic analysis services are available for text annotation. The platform also includes an active learning service enabling semi-automatic annotation of text attributes by letting the user interact with a machine learning module that will identify the most significant data to annotate. This service will thus predict annotation of some text attributes that were not manually annotated. Visualization and research tools are also integrated to facilitate the researcher's work.

Contributor:

CRIM

Research Subject:

Natural language and speech understanding

Support Email:

support-pacte@crim.ca

Type:

Managed

Version:

1.0.5

Release Date (UTC):

2018-03-01 12:00:00

CANARIE Project ID:

RS-10-1

Tags:

collaborative corpus disambiguation document lexicon named entity natural language processing text annotation

How to Cite:

Ménard, P. A. et Barrière, C. "PACTE: a collaborative platform for textual annotation" dans Proceedings of the 12th International Conference on Computational Semantics (IWCS 2017). Montpellier, France, du 19 au 22 septembre 2017