arrow drop search cross

INCEpTION: Efficient text annotation using human-in-the-loop technology [Workshop]

Workshop run by Ute Winchenbach (UKP - Ubiquitous Knowledge Processing Lab, INCEpTION Project) on May 7, 2020


19/05/2020   :   10h00
 Speaker: Ute Winchenbach (UKP - Ubiquitous Knowledge Processing Lab, INCEpTION Project)
Publication : 19/05/2020
Partager cet article :

Speaker: Ute Winchenbach (UKP - Ubiquitous Knowledge Processing Lab, INCEpTION Projet)

A wide range of research domains e.g. digital humanities are in need of easy and efficient ways to annotate specialized text corpora, e.g. for text exploration and interpretation or gold standard creation for machine learning algorithms. Creating corpora is not an easy task though: text annotation is a time consuming and expensive process; new annotation projects often require custom annotation schemata and workflows. The general purpose text annotation tool INCEpTION addresses these problems by supporting all phases of corpus creation: document selection, annotation, knowledge management and curation, as well as being easy to integrate into external workflows via remote API. INCEpTION employs machine learning to suggest potential annotations to users to reduce overall annotation time and effort. Many projects that used INCEpTION have shown that it is very extensible and can be easily adapted to the annotation work at hand.

Participants do not need to install INCEpTION to attend this seminar. Nevertheless, they may wish to have a look at some projects on this website and try INCEpTION beforehand. In this case, they can download it here and run it locally on their laptop via Java.

Due to current lockdown measures to fight against the spread of COVID-19, this seminar will be held on-line via Microsoft Teams. To register, please send an e-mail to Marco.CORNELI@univ-cotedazur.fr