The article titled “Annotate Smarter, Not Harder: Using Active Learning to Reduce Emotional Annotation Effort” was published in IEEE Transactions on Affective Computing, a top-tier journal ranked in the top 10% of both Scimago Journal Ranking and Google Scholar.

In this article, a team of researchers including LASIGE integrated members Soraia M. Alarcão, Vânia Mendonça, and Manuel J. Fonseca , and Cláudia Sevivas, LASIGE former member, proposed the use of active learning to reduce the annotation effort in emotional recognition tasks. Moreover, the authors also propose to cope with one of active learning’s shortcomings, by dynamically combining multiple active learning strategies through the use of online learning. This approach allowed to reduce the annotation effort, reaching or surpassing the performance of a supervised baseline with as little as 3% to 18% of the baseline’s training set.

The paper is available here.