logo epfl    School of Computer and Communication Sciences I&C LIA logo
Ecole Polytechnique Fédérale de Lausanne

EPFL > IC > LIA > Application Domains > Electronic Commerce

Artificial Intelligence Laboratory LIA


Electronic Commerce

Electronic commerce presents numerous challenges for intelligent systems. Many of these involve the interactions of customers with merchants through the internet, in partiuclar searching for items and obtaining recommendations. We have worked on the example-critiquing approach to preference-based search and developed commercial applications through the startup company Iconomic Systems SA. For recommender systems, we have developed the new approach of ontology filtering. It solves the cold-start problem of having to require new users to provide a large number or ratings before meaningful recommendations can be made. Ontology filtering is commercialized through the startup company Prediggo. Another application related to e-commerce is monitoring seller behavior through reputation systems. As part of a project involving several groups ( Computational Reputation Mechanisms for Enabling Peer-To-Peer Commerce in Decentralized Networks.), we have developed various techniques for obtaining truthful feedback in reputation systems. Based on these, we have developed a novel method for >Reliable Quality of Service Monitoring Based on Client Feedback that can be used for example for service-level agreements related to electronic services.
© 2011 EPFL all rights reserved Login