Abstract. We propose a novel method for addressing the model selection problem in the context of kernel methods. In contrast to existing methods which rely on hold-out testing or t...
Recognizing categories of articulated objects in real-world scenarios is a challenging problem for today's vision algorithms. Due to the large appearance changes and intra-cla...
Abstract. In contrast to the standard inductive inference setting of predictive machine learning, in real world learning problems often the test instances are already available at ...
Custom-designed DNA arrays offer the possibility of simultaneously monitoring thousands of hybridization reactions. These arrays show great potential for many medical and scientif...
Amir Ben-Dor, Richard M. Karp, Benno Schwikowski, ...
We now know that the addition of a concept constructor called a Path Functional Dependency (PFD) to the Boolean-complete description logic DLF leads to undecidability of ABox consi...