Explore the vast range of titles available.

In this work, we show that a late fusion approach to multimodality in sign language recognition improves the overall ability of the model in comparison to the singular approaches of image classification (88.14%) and Leap Motion data classification (72.73%). With a large synchronous dataset of 18 BSL gestures collected from multiple subjects, two deep neural networks are benchmarked and compared to derive a best topology for each. The Vision model is implemented by a Convolutional Neural Network and optimised Artificial Neural Network, and the Leap Motion model is implemented by an evolutionary search of Artificial Neural Network topology. Next, the two best networks are fused for synchronised processing, which results in a better overall result (94.44%) as complementary features are learnt in addition to the original task. The hypothesis is further supported by application of the three models to a set of completely unseen data where a multimodality approach achieves the best results relative to the single sensor method. When transfer learning with the weights trained via British Sign Language, all three models outperform standard random weight distribution when classifying American Sign Language (ASL), and the best model overall for ASL classification was the transfer learning multimodality approach, which scored 82.55% accuracy.

This study suggests a new approach to EEG data classification by exploring the idea of using evolutionary computation to both select useful discriminative EEG features and optimise the topology of Artificial Neural Networks. An evolutionary algorithm is applied to select the most informative features from an initial set of 2550 EEG statistical features. Optimisation of a Multilayer Perceptron (MLP) is performed with an evolutionary approach before classification to estimate the best hyperparameters of the network. Deep learning and tuning with Long Short-Term Memory (LSTM) are also explored, and Adaptive Boosting of the two types of models is tested for each problem. Three experiments are provided for comparison using different classifiers: one for attention state classification, one for emotional sentiment classification, and a third experiment in which the goal is to guess the number a subject is thinking of. The obtained results show that an Adaptive Boosted LSTM can achieve an accuracy of 84.44%, 97.06%, and 9.94% on the attentional, emotional, and number datasets, respectively. An evolutionary-optimised MLP achieves results close to the Adaptive Boosted LSTM for the two first experiments and significantly higher for the number-guessing experiment with an Adaptive Boosted DEvo MLP reaching 31.35%, while being significantly quicker to train and classify. In particular, the accuracy of the nonboosted DEvo MLP was of 79.81%, 96.11%, and 27.07% in the same benchmarks. Two datasets for the experiments were gathered using a Muse EEG headband with four electrodes corresponding to TP9, AF7, AF8, and TP10 locations of the international EEG placement standard. The EEG MindBigData digits dataset was gathered from the TP9, FP1, FP2, and TP10 locations.

In this work we present a three-stage Machine Learning strategy to country-level risk classification based on countries that are reporting COVID-19 information. A K% binning discretisation (K = 25) is used to create four risk groups of countries based on the risk of transmission (coronavirus cases per million population), risk of mortality (coronavirus deaths per million population), and risk of inability to test (coronavirus tests per million population). The four risk groups produced by K% binning are labelled as 'low', 'medium-low', 'medium-high', and 'high'. Coronavirus-related data are then removed and the attributes for prediction of the three types of risk are given as the geopolitical and demographic data describing each country. Thus, the calculation of class label is based on coronavirus data but the input attributes are country-level information regardless of coronavirus data. The three four-class classification problems are then explored and benchmarked through leave-one-country-out cross validation to find the strongest model, producing a Stack of Gradient Boosting and Decision Tree algorithms for risk of transmission, a Stack of Support Vector Machine and Extra Trees for risk of mortality, and a Gradient Boosting algorithm for the risk of inability to test. It is noted that high risk for inability to test is often coupled with low risks for transmission and mortality, therefore the risk of inability to test should be interpreted first, before consideration is given to the predicted transmission and mortality risks. Finally, the approach is applied to more recent risk levels to data from September 2020 and weaker results are noted due to the growth of international collaboration detracting useful knowledge from country-level attributes which suggests that similar machine learning approaches are more useful prior to situations later unfolding.

In this study, we present a transfer learning method for gesture classification via an inductive and supervised transductive approach with an electromyographic dataset gathered via the Myo armband. A ternary gesture classification problem is presented by states of ’thumbs up’ , ’thumbs down’ , and ’relax’ in order to communicate in the affirmative or negative in a non-verbal fashion to a machine. Of the nine statistical learning paradigms benchmarked over 10-fold cross validation (with three methods of feature selection), an ensemble of Random Forest and Support Vector Machine through voting achieves the best score of 91.74% with a rule-based feature selection method. When new subjects are considered, this machine learning approach fails to generalise new data, and thus the processes of Inductive and Supervised Transductive Transfer Learning are introduced with a short calibration exercise (15 s). Failure of generalisation shows that 5 s of data per-class is the strongest for classification (versus one through seven seconds) with only an accuracy of 55%, but when a short 5 s per class calibration task is introduced via the suggested transfer method, a Random Forest can then classify unseen data from the calibrated subject at an accuracy of around 97%, outperforming the 83% accuracy boasted by the proprietary Myo system. Finally, a preliminary application is presented through social interaction with a humanoid Pepper robot, where the use of our approach and a most-common-class metaclassifier achieves 100% accuracy for all trials of a ‘20 Questions’ game.

In this work we present the Chatbot Interaction with Artificial Intelligence (CI-AI) framework as an approach to the training of a transformer based chatbot-like architecture for task classification with a focus on natural human interaction with a machine as opposed to interfaces, code, or formal commands. The intelligent system augments human-sourced data via artificial paraphrasing in order to generate a large set of training data for further classical, attention, and language transformation-based learning approaches for Natural Language Processing (NLP). Human beings are asked to paraphrase commands and questions for task identification for further execution of algorithms as skills. The commands and questions are split into training and validation sets. A total of 483 responses were recorded. Secondly, the training set is paraphrased by the T5 model in order to augment it with further data. Seven state-of-the-art transformer-based text classification algorithms (BERT, DistilBERT, RoBERTa, DistilRoBERTa, XLM, XLM-RoBERTa, and XLNet) are benchmarked for both sets after fine-tuning on the training data for two epochs. We find that all models are improved when training data is augmented by the T5 model, with an average increase of classification accuracy by 4.01%. The best result was the RoBERTa model trained on T5 augmented data which achieved 98.96% classification accuracy. Finally, we found that an ensemble of the five best-performing transformer models via Logistic Regression of output label predictions led to an accuracy of 99.59% on the dataset of human responses. A highly-performing model allows the intelligent system to interpret human commands at the social-interaction level through a chatbot-like interface (e.g. “Robot, can we have a conversation?”) and allows for better accessibility to AI by non-technical users.

Stress has emerged as a major concern in modern society, significantly impacting human health and well-being. Statistical evidence underscores the extensive social influence of stress, especially in terms of work-related stress and associated healthcare costs. This paper addresses the critical need for accurate stress detection, emphasising its far-reaching effects on health and social dynamics. Focusing on remote stress monitoring, it proposes an efficient deep learning approach for stress detection from facial videos. In contrast to the research on wearable devices, this paper proposes novel Hybrid Deep Learning (DL) networks for stress detection based on remote photoplethysmography (rPPG), employing (Long Short-Term Memory (LSTM), Gated Recurrent Units (GRU), 1D Convolutional Neural Network (1D-CNN)) models with hyperparameter optimisation and augmentation techniques to enhance performance. The proposed approach yields a substantial improvement in accuracy and efficiency in stress detection, achieving up to 95.83% accuracy with the UBFC-Phys dataset while maintaining excellent computational efficiency. The experimental results demonstrate the effectiveness of the proposed Hybrid DL models for rPPG-based-stress detection.

In this work, we argue that the implications of pseudorandom and quantum-random number generators (PRNG and QRNG) inexplicably affect the performances and behaviours of various machine learning models that require a random input. These implications are yet to be explored in soft computing until this work. We use a CPU and a QPU to generate random numbers for multiple machine learning techniques. Random numbers are employed in the random initial weight distributions of dense and convolutional neural networks, in which results show a profound difference in learning patterns for the two. In 50 dense neural networks (25 PRNG/25 QRNG), QRNG increases over PRNG for accent classification at + 0.1%, and QRNG exceeded PRNG for mental state EEG classification by + 2.82%. In 50 convolutional neural networks (25 PRNG/25 QRNG), the MNIST and CIFAR-10 problems are benchmarked, and in MNIST the QRNG experiences a higher starting accuracy than the PRNG but ultimately only exceeds it by 0.02%. In CIFAR-10, the QRNG outperforms PRNG by + 0.92%. The n -random split of a Random Tree is enhanced towards and new Quantum Random Tree (QRT) model, which has differing classification abilities to its classical counterpart, 200 trees are trained and compared (100 PRNG/100 QRNG). Using the accent and EEG classification data sets, a QRT seemed inferior to a RT as it performed on average worse by − 0.12%. This pattern is also seen in the EEG classification problem, where a QRT performs worse than a RT by − 0.28%. Finally, the QRT is ensembled into a Quantum Random Forest (QRF), which also has a noticeable effect when compared to the standard Random Forest (RF). Ten to 100 ensembles of trees are benchmarked for the accent and EEG classification problems. In accent classification, the best RF (100 RT) outperforms the best QRF (100 QRF) by 0.14% accuracy. In EEG classification, the best RF (100 RT) outperforms the best QRF (100 QRT) by 0.08% but is extremely more complex, requiring twice the amount of trees in committee. All differences are observed to be situationally positive or negative and thus are likely data dependent in their observed functional behaviour.

Biophilic design is a well-recognised discipline aimed at enhancing health and well-being, however, most buildings lack adequate representation of nature or nature-inspired art. Notable barriers exist such as wealth, education, and physical ability restricting people’s accessibility to nature and associated artworks. An AI-based Biophilic arts curation and personalised recommendation system were developed in this study to improve accessibility to biophilic arts. Existing Biophilic research mainly focuses on building design principles, limited research exists to examine biophilic arts and associated emotional responses. In this paper, an interdisciplinary study addresses this gap by developing metrics for Biophilic art attributes and potential emotional responses, drawing on existing Biophilic architecture attributes and PANAS items. A public survey of 200 participants was developed in this study. The survey collected art viewers’ ratings of Biophilic attributes and associated emotional responses to establish statistical correlations between Biophilic attributes and emotional responses. The statistical analysis established a positive correlation between Biophilic attributes and positive emotions. The public survey results show significant positive emotional impacts (p-value <0.05) after exposure to Biophilic images, supporting further research and development of the Biophilic art curation system. This digital curation system employs Computer Vision algorithms (ResNet50) to automate Biophilic art categorisation and generate personalised recommendations. This study emphasises the importance of integrating nature into built environments. It proposes that artificial intelligence could significantly enhance the categorisation and recommendation of Biophilic art, advocating for expanding Biophilic art databases for emotionally responsive art display systems, benefiting mental health, and making art more accessible.

With growing societal acceptance and increasing cost efficiency due to mass production, service robots are beginning to cross from the industrial to the social domain. Currently, customer service robots tend to be digital and emulate social interactions through on-screen text, but state-of-the-art research points towards physical robots soon providing customer service in person. This article explores two possibilities. Firstly, whether transfer learning can aid in the improvement of customer service chatbots between business domains. Second, the implementation of a framework for physical robots for in-person interaction. Modelled on social interaction with Twitter customer support accounts, transformer-based chatbot models are initially assigned to learn one domain from an initial random weight distribution. Given shared vocabulary, each model is then tasked with learning another domain by transferring knowledge from the previous. Following studies on 19 different businesses, results show that the majority of models are improved when transferring weights from at least one other domain, in particular those that are more data-scarce than others. General language transfer learning occurs, as well as higher-level transfer of similar domain knowledge, in several cases. The chatbots are finally implemented on Temi and Pepper robots, with feasibility issues encountered and solutions are proposed to overcome them.

This study explores how robots and generative approaches can be used to mount successful false-acceptance adversarial attacks on signature verification systems. Initially, a convolutional neural network topology and data augmentation strategy are explored and tuned, producing an 87.12% accurate model for the verification of 2,640 human signatures. Two robots are then tasked with forging 50 signatures, where 25 are used for the verification attack, and the remaining 25 are used for tuning of the model to defend against them. Adversarial attacks on the system show that there exists an information security risk; the Line-us robotic arm can fool the system 24% of the time and the iDraw 2.0 robot 32% of the time. A conditional GAN finds similar success, with around 30% forged signatures misclassified as genuine. Following fine-tune transfer learning of robotic and generative data, adversarial attacks are reduced below the model threshold by both robots and the GAN. It is observed that tuning the model reduces the risk of attack by robots to 8% and 12%, and that conditional generative adversarial attacks can be reduced to 4% when 25 images are presented and 5% when 1000 images are presented.