TY - GEN AU - Kassner, Hendrik ID - 10680 TI - MPI-CUDA Codegenerierung für Nanophoton Simulationen auf Clustern ER - TY - CHAP AU - Kaufmann, Paul AU - Platzner, Marco ED - Müller-Schloer, Christian ED - Schmeck, Hartmut ED - Ungerer, Theo ID - 10687 T2 - Organic Computing---A Paradigm Shift for Complex Systems TI - Multi-objective Intrinsic Evolution of Embedded Systems VL - 1 ER - TY - GEN AU - Schwabe, Arne ID - 10736 TI - Analysis of Algorithmic Approaches for Temporal Partitioning ER - TY - CHAP AU - Sekanina, Lukas AU - Walker, James Alfred AU - Kaufmann, Paul AU - Plessl, Christian AU - Platzner, Marco ID - 10737 T2 - Cartesian Genetic Programming TI - Evolution of Electronic Circuits ER - TY - CHAP AU - Walker, James Alfred AU - Miller, Julian F. AU - Kaufmann, Paul AU - Platzner, Marco ID - 10748 T2 - Cartesian Genetic Programming TI - Problem Decomposition in Cartesian Genetic Programming ER - TY - GEN AU - Welp, Daniel ID - 10750 TI - User Space Scheduling for Heterogeneous Systems ER - TY - GEN AU - Feldotto, Matthias ID - 1092 TI - Parallele Kompression von XML mit XML-Schema-Subtraktion ER - TY - CONF AU - Bevermeier, Maik AU - Flanke, Stephan AU - Haeb-Umbach, Reinhold AU - Stehr, Jan ID - 11721 T2 - International Workshop on Intelligent Transportation (WIT 2011) TI - A Platform for efficient Supply Chain Management Support in Logistics ER - TY - CHAP AB - In this contribution classification rules for HMM-based speech recognition in the presence of a mismatch between training and test data are presented. The observed feature vectors are regarded as corrupted versions of underlying and unobservable clean feature vectors, which have the same statistics as the training data. Optimal classification then consists of two steps. First, the posterior density of the clean feature vector, given the observed feature vectors, has to be determined, and second, this posterior is employed in a modified classification rule, which accounts for imperfect estimates. We discuss different variants of the classification rule and further elaborate on the estimation of the clean speech feature posterior, using conditional Bayesian estimation. It is shown that this concept is fairly general and can be applied to different scenarios, such as noisy or reverberant speech recognition. AU - Haeb-Umbach, Reinhold ED - Haeb-Umbach, Reinhold ED - Kolossa, Dorothea ID - 11774 T2 - Robust Speech Recognition of Uncertain or Missing Data TI - Uncertainty Decoding and Conditional Bayesian Estimation ER - TY - CHAP AU - Haeb-Umbach, Reinhold ID - 11775 T2 - Baustelle Informationsgesellschaft und Universität heute TI - Können Computer sprechen und hören, sollen sie es überhaupt können? Sprachverarbeitung und ambiente Intelligenz ER -