Search

The protocadherins Fat and Dachsous regulate organ growth, shape, patterning, and planar cell polarity. Although Dachsous and Fat have been described as ligand and receptor, respectively, in a signal transduction pathway, there is also evidence for bidirectional signaling. Here, we assess signaling downstream of Dachsous through analysis of its intracellular domain. Genomic deletions of conserved sequences within dachsous identified regions of the intracellular domain that contribute to Dachsous activity. Deletion of the A motif increased Dachsous protein levels and decreased wing size. Deletion of the D motif decreased Dachsous levels at cell membranes, increased wing size, and disrupted wing, leg and hindgut patterning and planar cell polarity. Co-immunoprecipitation experiments established that the D motif is necessary and sufficient for association of Dachsous with key partners, including Lowfat, Dachs, Spiny-legs, Fat and MyoID. Subdivision of the D motif identified distinct regions that preferentially contribute to different Dachsous activities. Our results identify motifs that are essential for Dachsous function and are consistent with the hypothesis that the key function of Dachsous is regulation of Fat. [ABSTRACT FROM AUTHOR]

This paper illustrates the new structure of artificial neuron based on root-power mean (RPM) aggregation for quaternionic-valued signals and also presented an efficient learning process of neural networks with quaternionic-valued RPM neurons. The main aim of this neuron is to present the potential capability of a nonlinear aggregation operation on the quaternionic-valued signals in neuron cell. A wide spectrum of aggregation ability of RPM in between minima and maxima has a beautiful property of changing its degree of compensation in the natural way which emulates the various existing neuron models as its special cases. Further, the quaternionic resilient propagation algorithm (ℍ-RPROP) with error-dependent weight backtracking step significantly accelerates the training speed and exhibits better approximation accuracy. The wide spectra of benchmark problem are considered to evaluate the performance of proposed quaternionic RPM neuron with ℍ-RPROP learning algorithm. [ABSTRACT FROM AUTHOR]

Computation intelligence is an interesting field having ability to solve many complex problems that are exist in real world. Suitable collaboration of different type of computational learning intelligence techniques viz. fuzzy, evolutionary and neural methods can be more efficient for solving real word complex problems. This paper presents comparative analysis using some intelligence approaches viz. fuzzy c-means clustering(FCM), evolutionary fuzzy clustering with Minkowski distance (EFC-MD) and fusion of EFC-MD with functional modular neural network (EFCMD-FMNN) for determine their efficiencies in real domain (first generation neural network) and complex domain (second generation neural network) both. EFC-MD is introduced for pre-classification, specifying the optimal number of cluster that are assigned in training process. Rather than the Euclidean distance, Minkowski provides the flexibility to algorithm clustering in achieving any size for the cluster keeping the distance matrix in mind. A new approach, hilbert transformation is used for converting real datasets in complex datasets (complex number form) for the computation in complex domain. In this paper initially two existing real domain datasets (wine datasets and monk datasets) is selected and complexities these datasets in complex domain using hilbert approach. Next application of various computational algorithm separately in respective domain to evaluating performance in terms of accuracy. Exploration of these algorithms in complex domain (2-D) investigates improved results and better learning characteristics in the compression of real domain (1-D). [ABSTRACT FROM AUTHOR]

This paper aims to develop a recognition system which incorporates a novel fusion from different areas of computational intelligence along with a proposed artificial neuron structure. Our Deep Hybrid Computational Model (DHCM) is a novel aggregation of fuzzy clustering fused with evolutionary searching and a neural network based on a proposed neuron model TROIKA. The deep architecture with evolutionary fuzzy clustering followed by hybrid neural network, built upon novel neuron TROIKA, yields a powerful tool for face recognition applications. The proposed TROIKA neuron is based on nonlinear aggregation functions which enables our hybrid neural classifier to inculcate the benefits of its computational power resulting in superior learning and generalization. The proposed TROIKA neuron reduces the complexity of DHCM because very few neurons are sufficient to recognize a subject in database. The experimental results completed on two benchmark face datasets-INDIAN and FERET datasets demonstrated the effectiveness of the proposed model. The performance comparisons revealed the outperformance of (i) TROIKA based model over conventional neuron based model (ii) proposed model over other classification methods. [ABSTRACT FROM AUTHOR]

In the last few years there have been a growing number of studies concerning the introduction of quaternions into neural networks, which demand a faster learning technique with superior performance. In this paper, we propose a fast, but novel quaternionic resilient propagation (H-RPROP) algorithm for high dimensional problems. It achieves significantly faster learning over quaternionic domain back propagation (H-BP) algorithm. The slow convergence and stability of weight update around the local minima are the main drawbacks of H-BP. The gradient descent based H-BP algorithm takes the value of partial derivative (error gradient) and scales the weight updates through a learning rate while H-RPROP does not takes the value of partial derivatives, but it considers only the sign of partial derivatives that indicates the direction for each component of quaternionic weight update. The main aim of H-RPROP is to eliminate the value which is a little increased by constant increasing factor in order to accelerate convergence in shallow regions. H-RPROP computes an individual delta for each connection of the network, which determines the size of weight update. Therefore, the faster convergence and higher accuracy are the main key features of proposed algorithm. The intelligent behavior of the proposed learning approach is demonstrated through a wide spectrum of prediction problems with different statistical performance evaluation metrics. In order to illustrate the learning and generalization of 3D motion as its inherent behavior, a solid set of experiments is presented where the training is performed through input-output mapping over a line and the generalization ability is verified over various non-linear geometrical objects. The slow convergence problem of back-propagation algorithm has been well combated by H-RPROP. It has always demonstrated drastic reduction in the training cycles. [ABSTRACT FROM AUTHOR]

This paper describes an artificial neuron structure and an efficient learning procedure in the complex domain. This artificial neuron aims at incorporating an improved aggregation operation on the complex-valued signals. The aggregation operation is based on the idea underlying the weighted root-power mean of input signals. This aggregation operation allows modeling the degree of compensation in a natural manner and includes various aggregation operations as its special cases. The complex resilient propagation algorithm (\BBC-RPROP) with error-dependent weight backtracking step accelerates the training speed significantly and provides better approximation accuracy. Finally, performance evaluation of the proposed complex root-power mean neuron with the \BBC-RPROP learning algorithm on various typical examples is given to understand the motivation. [ABSTRACT FROM AUTHOR]