In this paper, two supervised methods for enhancing the classification accuracy of the Nonnegative Matrix Factorization (NMF) algorithm are presented. The idea is to extend the NMF algorithm in order to extract features that enforce not only the spatial locality, but also the separability between classes in a discriminant manner. The first method employs discriminant analysis in the features derived from NMF. In this way, a two-phase discriminant feature extraction procedure is implemented, namely NMF plus Linear Discriminant Analysis (LDA). The second method incorporates the discriminant constraints inside the NMF decomposition. Thus, a decomposition of a face to its discriminant parts is obtained and new update rules for both the weights and the basis images are derived. The introduced methods have been applied to the problem of frontal face verification using the well-known XM2VTS database. Both methods greatly enhance the performance of NMF for frontal face verification

Deep learning (DL) models can be used to tackle time series analysis tasks with great success. However, the performance of DL models can degenerate rapidly if the data are not appropriately normalized. This issue is even more apparent when DL is used for financial time series forecasting tasks, where the nonstationary and multimodal nature of the data pose significant challenges and severely affect the performance of DL models. In this brief, a simple, yet effective, neural layer that is capable of adaptively normalizing the input time series, while taking into account the distribution of the data, is proposed. The proposed layer is trained in an end-to-end fashion using backpropagation and leads to significant performance improvements compared to other evaluated normalization schemes. The proposed method differs from traditional normalization methods since it learns how to perform normalization for a given task instead of using a fixed normalization scheme. At the same time, it can be directly applied to any new time series without requiring retraining. The effectiveness of the proposed method is demonstrated using a large-scale limit order book data set, as well as a load forecasting data set.

Photonic artificial neural networks have garnered enormous attention due to their potential to perform multiply-accumulate (MAC) operations at much higher clock rates and consuming significantly lower power and chip real-estate compared to digital electronic alternatives. Herein, we present a comprehensive power consumption analysis of photonic neurons, taking into account global design parameters and concluding to analytical expressions for the neuron's energy- and footprint efficiencies. We identify the optimal design-space and analyze the performance plateaus and their dependence on a range of physical parameters, highlighting the existence of an optimal data-rate for maximizing the energy efficiency. Following a survey of the best-in-class integrated photonic devices, including on-chip lasers, photodetectors, modulators and weighting elements, the mathematically calculated energy and footprint efficiencies are mapped into real photonic neuron deployment scenarios. We reveal that silicon photonics can compete with the best-performing currently available digital electronic neural network engines, reaching ${\rm{TMAC/s/mm^{2}}}$ footprint- and sub-pJ/MAC energy efficiencies. Simultaneously, neuromorphic plasmonics, plasmo-photonics and sub-wavelength photonics hold the credentials for 1 to 3 orders of magnitude improvements even when the laser requirements and a reasonable waveguide pitch are accounted for, promising performance at a few fJ/MAC and up to a few ${\rm{TMAC/s/mm^{2}}}$ .

Machine learning methods have recently seen a growing number of applications in financial trading. Being able to automatically extract patterns from past price data and consistently apply them in the future has been the focus of many quantitative trading applications. However, developing machine learning-based methods for financial trading is not straightforward, requiring carefully designed targets/rewards, hyperparameter fine-tuning, and so on. Furthermore, most of the existing methods are unable to effectively exploit the information available across various financial instruments. In this article, we propose a deep reinforcement learning-based approach, which ensures that consistent rewards are provided to the trading agent, mitigating the noisy nature of profit-and-loss rewards that are usually used. To this end, we employ a novel price trailing-based reward shaping approach, significantly improving the performance of the agent in terms of profit, Sharpe ratio, and maximum drawdown. Furthermore, we carefully designed a data preprocessing method that allows for training the agent on different FOREX currency pairs, providing a way for developing market-wide RL agents and allowing, at the same time, to exploit more powerful recurrent deep learning models without the risk of overfitting. The ability of the proposed methods to improve various performance metrics is demonstrated using a challenging large-scale data set, containing 28 instruments, provided by Speedlab AG.

Time-series forecasting has various applications in a wide range of domains, e.g., forecasting stock markets using limit order book data. Limit order book data provide much richer information about the behavior of stocks than its price alone, but also bear several challenges, such as dealing with multiple price depths and processing very large amounts of data of high dimensionality, velocity, and variety. A well-known approach for efficiently handling large amounts of high-dimensional data is the bag-of-features (BoF) model. However, the BoF method was designed to handle multimedia data such as images. In this paper, a novel temporal-aware neural BoF model is proposed tailored to the needs of time-series forecasting using high frequency limit order book data. Two separate sets of radial basis function and accumulation layers are used in the temporal BoF to capture both the short-term behavior and the long-term dynamics of time series. This allows for modeling complex temporal phenomena that occur in time-series data and further increase the forecasting ability of the model. Any other neural layer, such as feature transformation layers, or classifiers, such as multilayer perceptrons, can be combined with the proposed deep learning approach, which can be trained end-to-end using the back-propagation algorithm. The effectiveness of the proposed method is validated using a large-scale limit order book dataset, containing over 4.5 million limit orders, and it is demonstrated that it greatly outperforms all the other evaluated methods.

In this paper, we propose a novel human crowd detection method that uses deep convolutional neural networks for drone flight safety purposes. The first contribution of this paper is to provide lightweight architectures, as restricted by the computational capacity of the specific application, capable of effectively distinguishing between crowded and non-crowded scenes from drone-captured images, and provide crowd heatmaps which can be used to semantically enrich the flight maps by defining no-fly zones. The second contribution of this paper is to propose a novel generic regularization technique, based on the graph embedding framework, applicable to different deep architectures for generic classification problems. The experimental validation is performed on a new dataset constructed for the task of human crowd detection from drone-captured images, and indicates the effectiveness of the proposed detector, as well as of the proposed regularizers in terms of classification accuracy. Finally, since the proposed regularization scheme is applicable in generic classification problems, we have also conducted experiments on two additional datasets, where the enhanced performance of the regularizers is also validated.