Sensitivity analysis is a method for extracting the cause and effect relationship between the inputs and outputs of the network. After training a neural network, one may want to k...
Abstract. The recurrent associative memory networks with complexvalued Hebbian matrices of connections are designed from interacting limitcycle oscillators. These oscillatory netwo...
Margarita Kuzmina, Eduard A. Manykin, Irina Surina
This paper is about a better understanding on the structure and dynamics of science and the usage of these insights for compensating the typical problems that arises in metadata-d...
To accelerate the learning of reinforcement learning, many types of function approximation are used to represent state value. However function approximation reduces the accuracy o...
Conventional methods for disambiguation problems have been using statistical methods with co-occurrence of words in their contexts. It seems that human-beings assign appropriate w...