In the paper we present a generalized discriminative multiple instance learning algorithm (GD-MIL) for multimedia semantic concept detection. It combines the capability of the MIL...
Current approaches for proving the correctness of business processes focus on either soundness, weak soundness, or relaxed soundness. Soundness states that each activity should be ...
In this paper, we tackle learning in distributed systems and the fact that learning does not necessarily involve the participation of agents directly in the inductive process itse...
A novel automatic image annotation system is proposed, which integrates two sets of SVMs (Support Vector Machines), namely the MIL-based (Multiple Instance Learning) and global-fe...
The method based on local features has an advantage that the important local motion feature is represented as bag-of-features, but lacks the location information. Additionally, in ...