The autoencoder is an artificial neural network that learns hidden representations of unlabeled data. With a linear transfer function it is similar to the principal component analysis (PCA). While both methods use weight vectors for linear transformations, the autoencoder does not come with any indication similar to the eigenvalues in PCA that are paired with eigenvectors. We propose a novel su...

Shortcut connections are a popular architectural feature of multi-layer perceptrons. It is generally assumed that by implementing a linear sub-mapping, shortcuts assist the learning process in the remainder of the network. Here we nd that this is not always the case: shortcut weights may also act as distractors that slow down convergence and can lead to inferior solutions. This problem can be a...

In this paper the authors present a simulation study of two 802.11e network scenarios. The presented analysis is not only novel but most of all crucial for understanding how a theoretically simple star topology network can be degraded by the presence of hidden nodes. The authors discuss the results obtained during the analysis of two different star topologies where the hidden node problem exist...

CR and PU hidden terminals in multi-channel Cognitive MAC protocols result in increased packet drops. This is due to inefficient node synchronization with existing “Control Channel” design. To date, In-band and Out-of-band CCC based MAC protocols are proposed to avoid PU and CR hidden terminals. But, In-band CCC based CR-MAC protocols cannot efficiently resolve the hidden terminal packet drops ...

According to conventional neural network theories, the feature of single-hidden-layer feedforward neural networks(SLFNs) resorts to parameters of the weighted connections and hidden nodes. SLFNs are universal approximators when at least the parameters of the networks including hidden-node parameter and output weight are exist. Unlike above neural network theories, this paper indicates that in o...

In connectionist networks, newly-learned information destroys previously-learned information unless the network is continually retrained on the old information. This behavior, known as catastrophic forgetting, is unacceptable both for practical purposes and as a model of mind. This paper advances the claim that catastrophic forgetting is a direct consequence of the overlap of the system’s distr...