Ioanna Lykourentzou - Guided crowdsourcing: Introduction, current approaches and open issues

Organized by: 



Ioanna Lykourentzou


- Mardi 18/06/2013 à 10h00

- Amphithéatre F107, INRIA Grenoble Rhône-Alpes, Montbonnot

- Intervenante : Ioanna Lykourentzou

- Researcher at CRP Henri Tudor/INRIA 
- On 
- Speaker’s biography : Dr. Ioanna Lykourentzou is a researcher at the Centre de Recherche Public Henri Tudor, Luxembourg in collaboration with Inria Nancy - Grand Est. Her research focuses on the performance improvement of crowd-involving systems, both on-line and off-line. Indicatively, she has worked on quality enhancement in wikis using machine learning and expert finding algorithms, on the provision of performance guarantees for crowdsourcing systems using task matching methods and, regarding off-line crowds, on the improvement of user routing in indoor physical spaces using scheduling techniques. She is also interested in large-scale knowledge retrieval and management. She received her Ph. D. in Collective Intelligence System Optimization from the National Technical University of Athens in 2009, and her Electrical and Computer Engineer diploma from the same university in 2005. She is an ERCIM and Marie Curie fellow. Her research interests are in the domains of collective intelligence, crowdsourcing optimization, corporate wikis, crowd simulation, machine learning and agent-based modeling.

Crowdsourcing is increasingly gaining momentum, as one of the most promising forms of large-scale dynamic collective work. Yet, current crowdsourcing approaches do not offer guarantees often demanded by consumers or by the involved community, for example regarding minimum job quality, maximum cost or job accomplishment time. Guided crowdsourcing can be an alternative to overcome these issues. This term refers to the use of algorithmic mechanisms to coordinate participants in crowdsourcing settings, in order to ensure collective performance goals such as quality, cost or time. In this 30-minute presentation we will start with an introduction to the crowdsourcing technology and examine it through three distinct examples : real-time, paid crowdsourcing (such as image recognition in AMT), knowledge-intensive crowdsourcing (such as Wikipedia) and crowdsourcing for solving open-ended problems (such as ontology matching). We will then see the main guided crowdsourcing approaches that have currently been proposed for optimizing some of the above cases. Finally, we will examine and discuss open issues and future research directions.