This paper deals with the exploration of biomedical multivariate time series to construct typical parameter evolution or scenarios. This task is known to be difficult: the tempora...
Abstract— This paper presents i-AA1 , a constructive, incremental learning algorithm for a special class of weightless, self-organizing networks. In i-AA1 , learning consists of ...
The concept of Dynamic Neural Networks (DNN) is a new approach within the Neural Network paradigm, which is based on the dynamic construction of Neural Networks during the processi...
Chrobak (1986) proved that a language accepted by a given nondeterministic finite automaton with one-letter alphabet, i.e., a unary NFA, with n states can be represented as the uni...
This paper describes a new method for building compact context-dependency transducers for finite-state transducer-based ASR decoders. Instead of the conventional phonetic decision...