Main Menu

Research & Integration


WP2: Evaluation, Integration and Standards
WP3: Visual Content Indexing
WP4: Content Description for Audio, Speech and Text
WP5: Multimodal Processing and Interaction
WP6: Machine Learning and Computation Applied to Multimedia
WP7: Dissemination towards Industry

Logo IST
Home arrow Research & Integration arrow E-teams arrow Active Semi-Supervised Learning

Active Semi-Supervised Learning

Contact person: Michel Crucianu, INRIA

  • Multimedia databases are characterized by large amounts of rather cheap data and small amounts of quite expensive knowledge. Among the many existing frameworks for learning, these characteristics support those frameworks that can exploit large amounts of data while using a minimum amount of expensive knowledge. This expensive knowledge can be either available a priori or actively acquired by interaction with users. The aim of this e-team is to better identify the working assumptions of the existing frameworks (supervised learning, transduction, semi-supervised clustering, unsupervised clustering...) in the context of applications to multimedia data, identify gaps in this spectrum of assumptions, put forward new methods for those gaps that deserve being filled and align existing semi-supervised learning methods to the requirements of multimedia content. Unlike other e-teams that directly focus on rather well-identified research themes, this one's first aim (i and ii) is the identification of valuable research directions in learning for multimedia content.

Additonal info: E-Team's webpage