LivingKnowledge goal is to bring a new quality into search and knowledge
management technology for more concise, complete and contextualised search results.
The paper “Large-Scale Support Vector Learning with Structural Kernels” co-written by A. Severyn and A. Moschitti has been presented at ECML/PKDD 2010 in Barcelona, Spain on September 20th - 24th, 2010.
ECML PKDD 2010
Large scale training: Existing solutions
Linear SVMs for binary classiﬁcation - O(n) scaling
• SVMGperf [Joachims, 2006]
• OCAS[Fran and Sonneburg, 2008]
• Pegasos [Schwartz, 2007]
Non=linear(e.g. Kernels) - scaling
• Sparse Kernel SVMs [Joachims, 2009]
• Basis pursuit approach
• SVMGstruct using Sampled Cuts [Yu and Joachims, 2008]
• Uses sampling to approximate the sub-gradient
The LK project is funded by the European Commission under Project No. 231126