Accepted Paper at SIGIR 2020 “Supporting Non-Expert Annotators with Dynamic Examples from Experts”

by Posted on

This week kicked-off with the good news of a paper acceptance at SIGIR, one of the core conferences for the Information Retrieval community. The paper is a result of collaboration with our colleague Markus Zlabinger.

TITLE: DEXA: Supporting Non-Expert Annotators with  Dynamic Examples from Experts

AUTHORS: Markus Zlabinger, Marta Sabou, Sebastian Hofstätter, Mete Sertkan and Allan Hanbury

The paper investigates improving crowd-based annotation of textual documents. It proposes a new approach in which crowd-workers are shown training examples which share similarities with the currently annotated instance. The method was tested on a medical document annotation task and showed very promising results.

Published by Ayu Irsyam