Search

High-quality audio data is a critical prerequisite for training robust text-to-speech models, which often limits the use of opportunistic or crowdsourced datasets. This paper presents an approach to overcome this limitation by implementing a denoising pipeline on the Catalan subset of Commonvoice, a crowd-sourced corpus known for its inherent noise and variability. The pipeline incorporates an audio enhancement phase followed by a selective filtering strategy. We developed an automatic filtering mechanism leveraging Non-Intrusive Speech Quality Assessment (NISQA) models to identify and retain the highest quality samples post-enhancement. To evaluate the efficacy of this approach, we trained a state of the art diffusion-based TTS model on the processed dataset. The results show a significant improvement, with an increase of 0.4 in the UTMOS Score compared to the baseline dataset without enhancement. This methodology shows promise for expanding the utility of crowdsourced data in TTS applications, particularly for mid to low resource languages like Catalan., Comment: Submitted to Iberspeech 2024

The increase in travelers and stays in tourist destinations is leading hotels to be aware of their ecological management and the need for efficient energy consumption. To achieve this, hotels are increasingly using digitalized systems and more frequent measurements are made of the variables that affect their management. Electricity can play a significant role, predicting electricity usage in hotels, which in turn can enhance their circularity - an approach aimed at sustainable and efficient resource use. In this study, neural networks are trained to predict electricity usage patterns in two hotels based on historical data. The results indicate that the predictions have a good accuracy level of around 2.5% in MAPE, showing the potential of using these techniques for electricity forecasting in hotels. Additionally, neural network models can use climatological data to improve predictions. By accurately forecasting energy demand, hotels can optimize their energy procurement and usage, moving energy-intensive activities to off-peak hours to reduce costs and strain on the grid, assisting in the better integration of renewable energy sources, or identifying patterns and anomalies in energy consumption, suggesting areas for efficiency improvements, among other. Hence, by optimizing the allocation of resources, reducing waste and improving efficiency these models can improve hotel's circularity.

Tabular data have been extensively used in different knowledge domains. Convolutional neural networks (CNNs) have been successfully used in many applications where important information about data is embedded in the order of features (images), outperforming predictive results of traditional models. Recently, several researchers have proposed transforming tabular data into images to leverage the potential of CNNs and obtain high results in predictive tasks such as classification and regression. In this paper, we present a novel and effective approach for transforming tabular data into images, addressing the inherent limitations associated with low-dimensional and mixed-type datasets. Our method, named Low Mixed-Image Generator for Tabular Data (LM-IGTD), integrates a stochastic feature generation process and a modified version of the IGTD. We introduce an automatic and interpretable end-to-end pipeline, enabling the creation of images from tabular data. A mapping between original features and the generated images is established, and post hoc interpretability methods are employed to identify crucial areas of these images, enhancing interpretability for predictive tasks. An extensive evaluation of the tabular-to-image generation approach proposed on 12 low-dimensional and mixed-type datasets, including binary and multi-class classification scenarios. In particular, our method outperformed all traditional ML models trained on tabular data in five out of twelve datasets when using images generated with LM-IGTD and CNN. In the remaining datasets, LM-IGTD images and CNN consistently surpassed three out of four traditional ML models, achieving similar results to the fourth model.

We present a specific-purpose globalized and preconditioned Newton-CG solver to minimize a metric-aware curved high-order mesh distortion. The solver is specially devised to optimize curved high-order meshes for high polynomial degrees with a target metric featuring non-uniform sizing, high stretching ratios, and curved alignment -- exactly the features that stiffen the optimization problem. To this end, we consider two ingredients: a specific-purpose globalization and a specific-purpose Jacobi-$\text{iLDL}^{\text{T}}(0)$ preconditioning with varying accuracy and curvature tolerances (dynamic forcing terms) for the CG method. These improvements are critical in stiff problems because, without them, the large number of non-linear and linear iterations makes curved optimization impractical. Finally, to analyze the performance of our method, the results compare the specific-purpose solver with standard optimization methods. For this, we measure the matrix-vector products indicating the solver computational cost and the line-search iterations indicating the total amount of objective function evaluations. When we combine the globalization and the linear solver ingredients, we conclude that the specific-purpose Newton-CG solver reduces the total number of matrix-vector products by one order of magnitude. Moreover, the number of non-linear and line-search iterations is mainly smaller but of similar magnitude., Comment: 64 pages, 9 figures

We define a regularized size-shape distortion (quality) measure for curved high-order elements on a Riemannian space. To this end, we measure the deviation of a given element, straight-sided or curved, from the stretching, alignment, and sizing determined by a target metric. The defined distortion (quality) is suitable to check the validity and the quality of straight-sided and curved elements on Riemannian spaces determined by constant and point-wise varying metrics. The examples illustrate that the distortion can be minimized to curve (deform) the elements of a given high-order (linear) mesh and try to match with curved (linear) elements the point-wise stretching, alignment, and sizing of a discrete target metric tensor. In addition, the resulting meshes simultaneously match the curved features of the target metric and boundary. Finally, to verify if the minimization of the metric-aware size-shape distortion leads to meshes approximating the target metric, we compute the Riemannian measures for the element edges, faces, and cells. The results show that, when compared to anisotropic straight-sided meshes, the Riemannian measures of the curved high-order mesh entities are closer to unit. Furthermore, the optimized meshes illustrate the potential of curved $r$-adaptation to improve the accuracy of a function representation., Comment: 49 pages, 23 figures

Transmission system operators have a growing need for more accurate forecasting of electricity demand. Current electricity systems largely require demand forecasting so that the electricity market establishes electricity prices as well as the programming of production units. The companies that are part of the electrical system use exclusive software to obtain predictions, based on the use of time series and prediction tools, whether statistical or artificial intelligence. However, the most common form of prediction is based on hybrid models that use both technologies. In any case, it is software with a complicated structure, with a large number of associated variables and that requires a high computational load to make predictions. The predictions they can offer are not much better than those that simple models can offer. In this paper we present a MATLAB toolbox created for the prediction of electrical demand. The toolbox implements multiple seasonal Holt-Winters exponential smoothing models and neural network models. The models used include the use of discrete interval mobile seasonalities (DIMS) to improve forecasting on special days. Additionally, the results of its application in various electrical systems in Europe are shown, where the results obtained can be seen. The use of this library opens a new avenue of research for the use of models with discrete and complex seasonalities in other fields of application., Comment: 40 pages, 8 figures, 6 tables

There are several international standards that define the way to evaluate the attenuation capacity of noise reducing devices, by single-number quantities representing airborne sound insulation and insertion loss. These two single-value ratings define the quality and performance of acoustic barriers, the former being related to intrinsic and the latter to both intrinsic and extrinsic acoustic characteristics of the devices. However, not many studies can be found on whether these objective parameters correlate to the perception of annoyance reduction. The aim of the present work is to analyze the adequacy of these objective ratings to indicate the performance of noise barriers, by comparing their values with the perception of annoyance reduction. For this purpose, ninety individuals of two different nationalities (Spanish and Portuguese) were asked to rate the perceived annoyance reduction in a listening experimental test, in which they were exposed, under controlled conditions, to several environmental noises and acoustic screened stimuli simulated by audio filters. The obtained results show a high correlation between objective ratings and subjective annoyance perception, with a better correlation being observed for insertion loss single-number parameter than for the airborne sound insulation single-number rating. Furthermore, significant differences were found depending on the gender and nationality of the respondents. The results, from this ongoing research work, may be of great interest for future acoustic barriers design.

Sonic crystal acoustic screens have been in progressive research and development in the last two decades as a technical solution for mitigating traffic noise. Their behaviour is quite different from that observed in classical barriers, with the latter being based on physically blocking the direct sound propagation path (only allowing diffracted noise to reach sensible receivers), and sonic crystals providing attenuation efficiency based on the creation of band-gaps at specific frequency ranges, due to the Bragg's interference phenomenon. The distinct physical mechanisms of these two types of noise barriers complicates the use of classical simplified or even numerical models developed for traditional barriers to simulate and predict the attenuation performance of a sonic crystal, and alternative methods become thus required. In the acoustics scientific literature, several authors have proposed estimation and simulation methods based on different numerical tools to predict the sound insulation provided by these new noise abatement solutions. This paper presents a comparative evaluation of some of these methods, with emphasis to the assessment of their accuracy versus memory usage in order to determine which one is the most suitable for optimization methodologies in the design of new devices with improved acoustic performance., Comment: 18 pages, 4 figures

This article presents the use of advanced tools applied to the design of devices that can solve specific acoustic problems, improving the already existing devices based on classic technologies. Specifically, we have used two different configurations of a material called Sonic Crystals, which is formed by arrays of acoustic scatterers, to obtain acoustic screens with high diffusion properties by means of an optimization process. This design procedure has been carried out using a multiobjective evolutionary algorithm along to an acoustic simulation model developed with the numerical method called Finite Difference Time Domain (FDTD). The results obtained are discussed in terms of both the acoustic performance and the robustness of the devices achieved., Comment: 19 pages, 5 figures

We detail how to use Newton's method for distortion-based curved $r$-adaption to a discrete high-order metric field while matching a target geometry. Specifically, we combine two terms: a distortion measuring the deviation from the target metric; and a penalty term measuring the deviation from the target boundary. For this combination, we consider four ingredients. First, to represent the metric field, we detail a log-Euclidean high-order metric interpolation on a curved (straight-edged) mesh. Second, for this metric interpolation, we detail the first and second derivatives in physical coordinates. Third, to represent the domain boundaries, we propose an implicit representation for 2D and 3D NURBS models. Fourth, for this implicit representation, we obtain the first and second derivatives. The derivatives of the metric interpolation and the implicit representation allow minimizing the objective function with Newton's method. For this second-order minimization, the resulting meshes simultaneously match the curved features of the target metric and boundary. Matching the metric and the geometry using second-order optimization is an unprecedented capability in curved (straight-edged) $r$-adaption. This capability will be critical in global and cavity-based curved (straight-edged) high-order mesh adaption.

We present a nodal interpolation method to approximate a subdivision model. The main application is to model and represent curved geometry without gaps and preserving the required simulation intent. Accordingly, we devise the technique to maintain the necessary sharp features and smooth the indicated ones. This sharp-to-smooth modeling capability handles unstructured configurations of the simulation points, curves, and surfaces. The surfaces correspond to initial linear triangulations that determine the sharp point and curve features. The method automatically suggests a subset of sharp features to smooth which the user modifies to obtain a limit model preserving the initial points. This model reconstructs the curvature by subdivision of the initial mesh, with no need of an underlying curved geometry model. Finally, given a polynomial degree and a nodal distribution, the method generates a piece-wise polynomial representation interpolating the limit model. We show numerical evidence that this approximation, naturally aligned to the subdivision features, converges to the model geometrically with the polynomial degree for nodal distributions with sub-optimal Lebesgue constant. We also apply the method to prescribe the curved boundary of a high-order volume mesh. We conclude that our sharp-to-smooth modeling capability leads to curved geometry representations with enhanced preservation of the simulation intent.

We present an $n$-dimensional marked bisection method for unstructured conformal meshes. We devise the method for local refinement in adaptive $n$-dimensional applications. To this end, we propose a mesh marking pre-process and three marked bisection stages. The pre-process marks the initial mesh conformingly. Then, in the first $n-1$ bisections, the method accumulates in reverse order a list of new vertices. In the second stage, the $n$-th bisection, the method uses the reversed list to cast the bisected simplices as reflected simplices, a simplex type suitable for newest vertex bisection. In the final stage, beyond the $n$-th bisection, the method switches to newest vertex bisection. To allow this switch, after the second stage, we check that under uniform bisection the mesh simplices are conformal and reflected. These conditions are sufficient to use newest vertex bisection, a bisection scheme guaranteeing key advantages for local refinement. Finally, the results show that the proposed bisection is well-suited for local refinement of unstructured conformal meshes.

Multi-functional wireless networks are rapidly evolving and aspire to become a promising attribute of the upcoming 6G networks. Enabling multiple simultaneous networking functions with a single radio fosters the development of more integrated and simpler equipment, overcoming design and technology barriers inherited from radio systems of the past. We are seeing numerous trends exploiting these features in newly designed radios, such as those operating on the mmWave band. In this article, however, we carefully analyze the challenges and opportunities for multi-functional wireless networks in UHF bands, advocating the reuse of existing infrastructures and technologies, and exploring the possibilities of expanding their functionality without requiring architectural changes. We believe that both modern and legacy technologies can be turned into multi-functional systems if the right scientific and technological challenges are properly addressed. This transformation can foster the development of new applications and extend the useful life of these systems, contributing to a more sustainable digitization by delaying equipment obsolescence.