Explore the vast range of titles available.

This article presents the application of a recent neural network topology known as the deep echo state network to the prediction and modeling of strongly nonlinear systems typical of the process industry. The article analyzes the results by introducing a comparison with one of the most common and efficient topologies, the long short-term memories, in order to highlight the strengths and weaknesses of a reservoir computing approach compared to one currently considered as a standard of recurrent neural network. As benchmark application, two specific processes common in the integrated steelworks are selected, with the purpose of forecasting the future energy exchanges and transformations. The procedures of training, validation and test are based on data analysis, outlier detection and reconciliation and variable selection starting from real field industrial data. The analysis of results shows the effectiveness of deep echo state networks and their strong forecasting capabilities with respect to standard recurrent methodologies both in terms of training procedures and accuracy.

The dissemination of fake news on social media platforms is an issue of considerable interest, as it can be used to misinform people or lead them astray, which is particularly concerning when it comes to political events. The recent event of Hong Kong protests triggered an outburst of fake news posts that were identified on Twitter, which were then promptly removed and compiled into datasets to promote research. These datasets focusing on linguistic content were used in previous work to classify between tweets spreading fake and real news using traditional machine learning algorithms (Zervopoulos et al., in: IFIP international conference on artificial intelligence applications and innovations, Springer, Berlin, 2020). In this paper, the experimentation process on the previously constructed dataset is extended using deep learning algorithms along with a diverse set of input features, ranging from raw text to handcrafted features. Experiments showed that the deep learning algorithms outperformed the traditional approaches, reaching scores as high as 99.3% F1 Score, with the multilingual state-of-the-art model XLM-RoBERTa outperforming other algorithms using raw untranslated text. The combination of both traditional and deep learning algorithms allows for increased performance through the latter, while also gaining insight regarding tweet structure from the interpretability of the former.

Facial expression recognition (FER) in-the-wild is challenging due to unconstraint settings such as varying head poses, illumination, and occlusions. In addition, the performance of a FER system significantly degrades due to large intra-class variation and inter-class similarity of facial expressions in real-world scenarios. To mitigate these problems, we propose a novel approach, Discriminative Attention-augmented Feature Learning Convolution Neural Network (DAF-CNN), which learns discriminative expression-related representations for FER. Firstly, we develop a 3D attention mechanism for feature refinement which selectively focuses on attentive channel entries and salient spatial regions of a convolution neural network feature map. Moreover, a deep metric loss termed Triplet-Center (TC) loss is incorporated to further enhance the discriminative power of the deeply-learned features with an expression-similarity constraint. It simultaneously minimizes intra-class distance and maximizes inter-class distance to learn both compact and separate features. Extensive experiments have been conducted on two representative facial expression datasets (FER-2013 and SFEW 2.0) to demonstrate that DAF-CNN effectively captures discriminative feature representations and achieves competitive or even superior FER performance compared to state-of-the-art FER methods.

Everybody likes to make a good prediction, in particular, when some sort of personal investment is involved in terms of finance, energy or time. The difficulty is to make a prediction that optimises the reward obtained from the original contribution; this is even more important when investments are the core service offered by a business or pension fund. The complexity of finance is that the human predictor may have other interests or bias than the human investor, the trust between predictor and investor will never be completely established as the investor will never know if the predictor has generated, intentionally or unintentionally, the optimum possible reward. This article presents the random neural network (RNN) in a recurrent configuration for the 4th industrial revolution on a Fintech application; the RNN is proposed to make predictions on time series data, specifically, prices. The biological model inspired by the brain structure and neural interconnections make predictions entirely on previous data from the time series rather than predictions based on several uncorrelated inputs. The model is validated against the property, stock and Fintech market: (1) UK property prices, (2) stock market indice prices, (3) cryptocurrency prices. Experimental results show that the proposed method makes accurate predictions on different investment portfolios. The prediction accuracy of the proposed RNN model is compared against long short-term memory (LSTM) and linear regression (LR) models.

This work describes a novel methodology for creating exergames on an edge-native platform with the integration of multiple deep neural networks. A prototype of the platform, which includes capabilities for innovative gameplay and advanced user interactivity, has been implemented and deployed in a real-world scenario. At core of the proposed methodology is the ad hoc training of classifiers for posture classification which can be dynamically adapted to the specific requirements of the usage scenario, operational and environmental conditions allowing for real-time identification of events and advanced game control. The proposed solution is ideal for individual consumers in a home environment since is supports by-design edge platforms minimizing the cost of the system and enabling in parallel the communication with state-of-the-art hardware (i.e., GPUs, TPUs, computer boards) for real-time operation. The proposed system allows the collection and analysis of game data, which can be exploited by specialized personnel in rehabilitation centers or for other purposes in the areas of healthcare and assisted living.

As neural architecture search (NAS) becomes an increasingly adopted method to design network architectures, various methods have been proposed to speedup the process. Besides proxy evaluation tasks, weight sharing, and scaling down the evaluated architectures, performance-predicting models exhibit multiple advantages. Eliminating the need to train candidate architectures and enabling transfer learning between datasets, researchers can also utilize them as a surrogate function for Bayesian optimization. On the other hand, graph convolutional networks (GCNs) have also been increasingly adopted for various tasks, enabling deep learning techniques on graphs without feature engineering. In this paper, we employ an evolutionary-based NAS method to evolve GCNs for the problem of predicting the relative performance of various architectures included in the NAS-Bench-101 dataset. By fine-tuning the architecture generated by our methodology, we manage to achieve a Kendall’s tau correlation coefficient of 0.907 between 1050 completely unseen architectures, utilizing only 450 samples, while also outperforming a strong baseline on the same task. Furthermore, we validate our method on custom global search space architectures, generated for the Fashion-MNIST dataset.

This article presents iBuilding: distributed artificial intelligence embedded into Intelligent or Smart Buildings in an Industry 4.0 application that enables the adaptation to the external environment and the different building users. Buildings are becoming more intelligent in the way they monitor the usage of its assets, functionality and space. The more efficiently a building can be monitored or predicted, the more return of investment can deliver as unused space or energy can be redeveloped or commercialized, therefore reducing energy consumption while increasing functionality. This article proposes distributed artificial intelligence embedded into a Building based on neural networks with a deep learning structure. (1) Sensorial neurons at the device level are dispersed through the intelligent building to gather, filter environment information and predict its next values. (2) Management neurons based on reinforcement learning algorithm at the edge level make predictions about values and trends for building managers or developers to make commercial or operational informed decisions. (3) Finally, transmission neurons based on the genetic algorithms and the genome codify, transmit iBuilding information and also multiplex its data entirely to generate clusters of buildings interconnected with each other at the cloud level. The proposed iBuilding based on distributed learning is validated with a public research dataset; the results show that artificial intelligence embedded into the intelligent building enables real-time monitoring and successful predictions about its variables. The key concept proposed by this article is that the learned information obtained by iBuilding after its adaptation to the environment is never lost when the building changes over time or is decommissioned but transmitted to future generations.

Data flow learning algorithms must be very efficient in learning and predicting sequences. The model that monitors a sequence of data or events can predict the sequel and can act in such a way that it optimally achieves the desired result. Security and digital risk tracking systems are receiving a constant and unlimited input of observations. These data flows are characterized by high variability, as their properties can change drastically and unpredictably over time. Each incoming example can only be processed once, or it must be summarized with a small memory imprint. This research paper proposes the development of an intelligent system, for real-time detection of data flow anomalies related to information systems’ security. Specifically, it describes the implementation of an efficient and high-precision energy-based Online Sequential Extreme Learning Machine (e-b OSELM) that is proposed for the first time in the literature. It is an intelligent model that can detect data dependencies, by applying a measure of compatibility (scalable energy) to each configuration of its variables. It assigns low energy to the correct values and higher energy to the divergent (abnormal) ones. The innovative combination of energy models and ELMs offers high learning speed, ease of execution, minimum human involvement and minimum computational power and resources for anomaly detection and identification.

The special nature, volume and broadness of biomedical literature pose barriers for automated classification methods. On the other hand, manually indexing is time-consuming, costly and error prone. We argue that current word embedding algorithms can be efficiently used to support the task of biomedical text classification even in a multilabel setting, with many distinct labels. The ontology representation of Medical Subject Headings provides machine-readable labels and specifies the dimensionality of the problem space. Both deep- and shallow network approaches are implemented. Predictions are determined by the similarity between extracted features from contextualized representations of abstracts and headings. The addition of a separate classifier for transfer learning is also proposed and evaluated. Large datasets of biomedical citations are harvested for their metadata and used for training and testing. These automated approaches are still far from entirely substituting human experts, yet they can be useful as a mechanism for validation and recommendation. Dataset balancing, distributed processing and training parallelization in GPUs, all play an important part regarding the effectiveness and performance of proposed methods.

A data-driven two-branch deep neural network (DNN), to be referred to as λ -DNN, used to predict scalar fields is presented. The network architecture consists of two separate branches (input layers) connected to the main one towards its output. In multi-disciplinary shape optimization problems, such as those this paper is dealing with, the input to the λ -DNN contains data relevant to the geometrical shape and the case itself. Herein, the λ -DNN is used in conjugate heat transfer (CHT) analysis and shape optimization problems, synergistically with codes simulating flows over the fluid domain and solving the heat conduction equations over the solid one. It is used to optimize a duct and an internally cooled turbine blade-airfoil surrounded by hot gas. The λ -DNNs, after being trained on fields computed using the CHT solver, are used as surrogates for either the heat conduction equation solver of the solid domain, replicating either one out of the two disciplines of the problem or the coupled CHT solver.