Artificial neural networks (ANNs) have become a popular tool in digital signal processing (DSP). Among the widespread ANN architectures, complex-valued neural networks (CVNNs) have been extensively studied in image processing and telecommunications. Unlike their real-valued counterparts, CVNNs can handle signals directly in the complex domain. Due to this capability, CVNNs usually exhibit higher accuracy and improved convergence compared to real-valued neural networks (RVNNs). Despite their improved performance in several applications, CVNNs still lag behind RVNNs in terms of learning techniques and heuristics. In this context, we propose adaptive learning rate approaches for CVNNs, extending the well-known adaptive gradient (AdaGrad), root-mean-square propagation (RMSProp), AdaMax, AMSGrad, softplus AMSGrad (SAMSGrad), Nesterov-accelerated adaptive moment estimation (Nadam), and DiffGrad to the complex domain. Computational complexities of the proposed optimizers are analyzed for CVNN architectures. Results are compared in terms of mean-squared-error convergence.

This work presents Chat-IBN-RASA, a conversational AI chatbot based on the open-source RASA framework, acting as an Intent Translator component for Intent-Based Networking (IBN) architectures. The IBN-driven RASA chatbot allows users to interact with the network management system using high-level language communication without the need for in-depth technical knowledge of the network. The chatbot is trained using Natural Language Understanding (NLU) models, a defined domain, stories, and custom actions for API communication and database queries. We present the prototype implementation in a use case of survivability intents in packet-optical networks. The custom actions featured include communication with the network database, path computation, and the recommended path for intent deployment, considering availability, protection type, data rate, and link distances.

The objective of this paper is to present a generic and extensible access framework architecture for WebLab integration. In this framework each Weblab becomes accessible by means of a preinstalled plug-in. This modular approach makes it possible to add, remove or modify a plug-in, and its corresponding Weblab, without framework recompilation.

We propose a novel policy called proportional sharing with load reservation (PSLR) for scheduling upstream flows in Ethernet passive optical networks. This policy provides bandwidth guarantees on a per flow basis and redistributes the unused bandwidth among active flows in proportion to their priority level.

In this paper, an envelope process called fractional bounded arrival process (FBAP) is proposed fro self-similar traffic representation. The queuing analysis for FBAP traffic is developed, and upper bounds for the backlog and for the delay are obtained. Finally, the use of the leaky bucket algorithm for policing FBAP traffic is investigated.

It has been proved that it is possible to guarantee delay bounds in a generalized processor sharing (GPS) server when traffic is regulated by the leaky bucket algorithm. However, the leaky bucket is not an appropriate regulator for long-range dependent (LRD) traffic. This paper introduces the computation of backlog and delay bounds for a GPS server under LRD traffic. Traffic is assumed to be regulated by the fractal leaky bucket, which is proved to be an effective policing mechanism for LRD traffic.

This paper introduces the computation of end-to-end delay and backlog bounds in a network of generalized processor sharing (GPS) servers with long-range dependent traffic. The traffic is supposed to be regulated by the fractal leaky bucket, which is an appropriate regulator for such traffic. The network of GPS servers has arbitrary topology and the stability of these networks is analysed.

The constant modulus algorithm (CMA) is recognized as the most widely used algorithm in blind channel equalization practice. However, the CMA cost function exhibits local minima, which often leads to ill-convergence. This paper proposes a concurrent equalizer, in which a least mean square (LMS) equalizer operates cooperatively with a CMA equalizer, controlled through a non-linear link that depends on the system a priory state. Simulation results using M-QAM signalling have shown that the concurrent equalizer presents a much lower sensitivity to local minima than the CMA approach.

This paper proposes an image interpolation model based on a multiresolution scheme. The method uses two different known image resolutions to establish a mapping between them. The mapping is then used to extrapolate the unknown resolution to produce an expanded image with increased definition and with a mean square reconstruction error lower than for the conventional methods. In other words, details in a high resolution image are implicitly predicted based only on its low-frequency resolutions. A high subjective quality is obtained due to the similar nature of the mappings, which are constructed by means of an adaptive bidimensional FIR filter.

This paper investigates the application of genetic algorithms to the decoding problem of error-correcting codes. The basic operations of crossover, mutation and selection are appropriately defined for the decoding problem of binary linear block codes without the knowledge of any algebraic structure. Simulation results indicate that these evolutionary techniques, if properly applied, can provide very interesting results for this difficult combinatorial problem. The result were obtained for hard-decision decoding of binary linear block codes, but the algorithms can be generalized to soft-decision decoding and to other kinds of linear codes. The algorithms presented are in fact evolutionary versions of the well known information set decoding.

A technique for assigning the Gaussian centers to a radial basis function neural network (RBFNN) based on the Karhunen-Loeve transform (KLT), is proposed by applying this technique to time series prediction, a significant performance improvement is obtained in comparison with usual prediction methods that use RBFNNs. For instance, by assigning the KLT scaled eigenvectors to the RBFNN centers yields lower prediction normalized mean squared error and requires less previous known samples than the usual technique that applies the own training set vectors to the centers. The present technique has also shown improved performance when compared with prediction based on RBFNNs that uses the K-means clustering algorithm.

The authors have developed an interactive environment for the creation and maintenance of dynamic, active multimedia-based teaching mechanisms. The environment is designed to be user-friendly and to facilitate the creation of educational material. This tool has already been used to create courses in multidimensional signal processing (MDSP) by researchers working together while geographically separated.

A systematic technique for the optimal design of phase-locked loops for synchronous networks is presented. The method is based on Kalman estimation theory under self-similar random noise processes. This approach is optimal for certain noise models and for linear phase-detectors. The results are then extended in order to maintain the minimum mean-square phase error when the reference signal of a master-slave network is lost.

A method for spectral calculation ofq-ary digital signals encoded byp-ary convolutional codes is developed. We have assumed a parallel encoding process such that for a code with ratek/nand constraint lengthm, the encoder can be modeled as a Mealy machine ofkinputs,noutputs, and(m - 1) kq-ary memory cells. From this model and from the generator matrix of the code, the characterizing matrices of the Mealy machine are found in a straightforward way. Two assumptions are made on the statistics of the input signal. 1) The input sequence is a wide-sense stationary vector process of dimensionk. 2) The input vectors are statistically independent. With these assumptions and from the characterizing matrices of the Mealy machine, the method developed by Cariolaro and Tronca can be readily used. The continuous and discrete components of the spectrum of the coded signal are expressed in matrix form, which is particularly appropriate for computer programming.

A new strategy for frame synchronization of PCM muitlplex systems which use pulse stuffing (justification) is proposed. In devising this technique, the compatibility with existing CCITT systems was taken into consideration. Comparisons between the proposed strategy and the CCITT recommended system are made in terms of average holding time, average time for detection of synchronization loss, and average synchronization recovery time. The proposed strategy proved to be significantly better in total synchronization time without sacrificing its holding time capacity.