AI-generated artworks are rapidly improving in quality, and bring many ethical issues to the forefront of discussion. Data scarcity leaves many individuals under-represented due to aspects such as age and ethnicity, which can provide useful context when transferring artistic styles to an image. In this study, we consider current issues through the engineering of an AI art model trained on work inspired by Vincent van Gogh. The model is fine-tuned from a dataset of nearly 6 billion images and thus enables style transfer to individuals and entities not present in the art dataset given the knowledge of context. All models in this work are trained on consumer-level computing hardware with presented hyperparameters and configurations. Finally, we explore the application of computer vision models that can detect when an artwork has been created by human or machine with 98.14% accuracy. The dataset and models are open-sourced for future work.

Spiking Neural Networks (SNNs) emerged as a promising solution in the field of Artificial Neural Networks (ANNs), attracting the attention of researchers due to their ability to mimic the human brain and process complex information with remarkable speed and accuracy. This research aimed to optimise the training process of Liquid State Machines (LSMs), a recurrent architecture of SNNs, by identifying the most effective weight range to be assigned in SNN to achieve the least difference between desired and actual output. The experimental results showed that by using spike metrics and a range of weights, the desired output and the actual output of spiking neurons could be effectively optimised, leading to improved performance of SNNs. The results were tested and confirmed using three different weight initialisation approaches, with the best results obtained using the Barabasi-Albert random graph method.

A fuzzy Logic System (FLS) has the unique potential of intelligent decision-making capability with the input data, which are imprecise and unpredictable but can be expressed using linguistic rules. However, such FLS strongly depends on the hardware architecture of the systems under consideration. A language called Fuzzy Markup Language (FML), which is based on XML, has the ability to express fuzzy systems without taking into account either hardware or software restrictions, thus enabling sharing effectively among distributed environments. FML creates a unified and clearly defined FLS that is independent of its previous representation. It also increases the capabilities of system designers by giving them access to a range of services that accelerate the process of creating a distributed or centralised fuzzy system. This can be done either through a centralised cloud-based platform or through a distributed cloud-based platform. FML-related software libraries, such as Simpful (a Python-based library) and Java FML (JFML, a Java-based library), have added to the novelty in it. However, JFML and Simpful enable fuzzy system modeling for programmers only in the abovementioned programming languages. Hence a need for the development of an open-source FLS which will be capable of enabling fuzzy system-based modelling is felt. The present communication introduces an enhancement of JFML through a visual environment based on GUI following IEEE Std 1855-2016. The functionalities of user-friendly visual FML are developed considering the data related to Human Activity Recognition (HAR) and other physiological data like SpO2 and pulse rate obtained from wearable sensors to compare the performances.

Despite extensive research on human-centric domestic behaviour monitoring for various essential applications, there are still significant challenges to deploying such systems on a larger scale. One of the main obstacles is adjusting the gap between privacy, performance, and the cost of assistive technologies to support older adults living independently in their homes. For instance, while traditional vision-based sensing approaches offer high performance, they compromise human privacy in domestic environments. On the other hand, ambient sensing approaches, such as the use of Passive Infra-Red (PIR) sensors, maintain human privacy but are hindered significantly in real-world scenarios, such as multi-occupancy environments. Inspired by our previous research work, this paper proposes a holistic system approach of several functional phases that can be used together to monitor and facilitate the independence of older adults using the Thermal Sensor Array.

The role of machine learning in neuroscience has been increasing through the years, in aiding diagnosis, biomarker discovery, signal analysis, and other applications. However, the lack of information of the decision-making of the models restricts their use and adoption by the community. In the process of neuronal signal acquisition, other electrical signals can distort the recording, for which a review process is necessary. Machine learning can aid by automatically detecting affected segments, speeding up the review process. However, as the ground-truth labelling is done manually or via a threshold, researchers must be able to identify the causes of false negatives and positives. This paper looks into explainable machine learning for artefact detection in invasively recorded neural signals through the use of different classifiers, trained with a feature subset produced by the combination of feature selection algorithms to reduce the dimensionality by two orders of magnitude. Our results show that the bagging decision tree model is best suited for creating a generalised model that is capable of classifying artefactual patterns in a multi-state dataset, which achieves an accuracy of 96.1%. Lastly, the predictor importance, Shapely values, and reduced feature space visualisation are used to gain insight into the model.

Brain signals are recorded using different techniques to aid an accurate understanding of brain function and to treat its disorders. Untargeted internal and external sources contaminate the acquired signals during the recording process. Often termed as artefacts, these contaminations cause serious hindrances in decoding the recorded signals; hence, they must be removed to facilitate unbiased decision making for a given investigation. Due to the complex and elusive manifestation of artefacts in neuronal signals, computational techniques serve as powerful tools for their detection and removal. Machine learning (ML) based methods have been successfully applied in this task. Due to ML's popularity, many articles are published every year, making it challenging to find, compare and select the most appropriate method for a given experiment. To this end, this paper presents ABOT (Artefact removal Benchmarking Online Tool) as an online benchmarking tool which allows users to compare existing ML-driven artefact detection and removal methods from the literature. The characteristics and related information about the existing methods have been compiled as a knowledgebase (KB) and presented through a user-friendly interface with interactive plots and tables for users to search it using several criteria. Key characteristics extracted from over 120 articles from the literature have been used in the KB to help compare the specific ML models. To comply with the FAIR (Findable, Accessible, Interoperable and Reusable) principle, the source code and documentation of the toolbox have been made available via an open-access repository.

Neural signals are the recordings of the electrical activity individual or groups of neurons, and they are used for disease staging, brain-computer interface control and understanding the neural processes. When carrying out a functional connectivity study in rodents, processing must be done to eliminate disturbance in the data in order to have the most faithful representation of the neural activity. This step mainly includes filtering and artefact removal, where the latter can be approached by diverse methods. Furthermore, it is important to identify when the rodent is stressed, as the local field potentials can be coupled to theta oscillations. To this end, we set out to develop a machine learning-based model for the detection of stress in rodents with multi-modal recordings, namely local field potentials, respiration and electrocardiography. We explore subject-specific and cross-subject models, as well as employing an artefact detection model as a generic anomaly detector. Results show that subject-specific models can achieve a good performance, but the variability is significant across all three signals among rodents of the same age, gender and species.

Mobile applications in the area of human-centered applications are based on fuzzy logic have exhibited their effectiveness in managing intelligent environments, however the deployment of mobile fuzzy logic systems has been usually associated with dedicated hardware and software packages. Introducing openness for fuzzy logic systems offers exciting features such as system independence, simplicity, load balancing, and controlled resource allocation. On the other hand, while major cloud service providers support readymade commercial services for AI techniques such as for deep neural networks, there is no similar services for fuzzy logic systems. This study aims to develop a cloud-based fuzzy logic system under Microsoft Azure, employing Simpful as the cloud-side Python library and FML as data exchange standard. The developed cloud service is shown to effectively serve mobile phone applications for human monitoring purposes. Also in the present study, two types of fuzzy inference systems namely Mamdani and TSK have been utilized wherein both these systems have been compared on the basis of their processing time and accuracy of result. Results indicated that Mamdani fuzzy inference system outperformed TSK fuzzy inference system in terms of processing time by 0.456 seconds. Moreover, the detection accuracy of Mamdani system was found to be higher than that of TSK system by 6.82%.

Traditionally, fuzzy logic systems are linked to specific hardware or software systems. Observations reveal that dispersed and distributed designs of intelligent systems are gaining attraction. Due to the possible complexities of fuzzy logic computations, distributed architectures have the potential to add value to the development of fuzzy systems. However, the absence of best practices and standard methodologies may prevent widespread adoption. By broadening the IEEE-1855 (2016) standard in terms of system definition and data exchange, this research offers a standard solution for building a Service-Oriented Architecture (SOA) as a novel method of implementing fuzzy logic systems by means of a cloud-based collecting, processing, and examining data over the web. A comparison between the performances of a stand-alone hardware-dependent solution and a cloud-based solution (known as fuzzy-as-a-service) is performed. The analysis is also carried out on two different cloud service providers and software libraries (Amazon Web Services using JFML as a java-based library and Azure Web Services using Simpful as a python-based library). The analysis and evaluation are performed on a human fall detection scenario involving wearable sensors. The proposed algorithm can identify between fall and non-fall events. However, the results show that the processing time taken per 10,000 samples using smartwatch and mobile was 2220 s and 101 s for a cloud-based non-fuzzy machine learning system, 1111 s and 45 s for a cloud-based fuzzy system with AWS and JFML, and 1250 s and 97 s for a cloud-based fuzzy system with Microsoft Azure and Simpful libraries. It has been observed that a smartwatch with a fuzzy stand-alone crashed after processing 5000 samples and a mobile phone requires 179.42 s to process 10,000 samples.

Fuzzy logic system (FLS) based applications in Ambient Intelligence (AmI) have exhibited their capabilities in realizing intelligent environments. FLS deployment is usually associated with non-scalable hardware and software platforms. For intensive computation-based FLSs, distributing FLS openness as web services is known as fuzzy-as-a-service (FaaS), wherein a service is self-sufficiently developed from FLS itself, permits system independence, ingenuousness, load balancing and well-organised resource allocation. The fuzzy mark-up language (FML) and its related software libraries such as Simpful and Java FML (JFML) have added to the novelty. This study aims to develop and compare service-oriented architecture based FaaS by employing Microsoft Azure and Amazon Web Service (AWS) for AmI implementation in e-healthcare. Data is collected, processed and monitored via the Internet. Heart rate and SpO2 data are used in the study. The obtained data is evaluated by a fuzzy inference system using Microsoft Azure and AWS and displayed via a mobile application. A simulation scenario of real-time human activity recognition using a rule-based FLS is demonstrated. Results indicate that the proposed FaaS requires an average of 0.306 seconds to process a single real-time data using AWS and JFML. Conversely, Azure and Simpful require 0.456 seconds for processing. Thus, as per the selected scenario in this research, it can be concluded that there is a variation in the processing time using different cloud-based services. For example, AWS in one of the use case scenarios takes 19.65% less processing time of compared to Microsoft Azure.

Acquisition of neuronal signals involves a wide range of devices with specific electrical properties. Combined with other physiological sources within the body, the signals sensed by the devices are often distorted. Sometimes these distortions are visually identifiable, other times, they overlay with the signal characteristics making them very difficult to detect. To remove these distortions, the recordings are visually inspected and manually processed. However, this manual annotation process is time-consuming and automatic computational methods are needed to identify and remove these artefacts. Most of the existing artefact removal approaches rely on additional information from other recorded channels and fail when global artefacts are present or the affected channels constitute the majority of the recording system. Addressing this issue, this paper reports a novel channel-independent machine learning model to accurately identify and replace the artefactual segments present in the signals. Discarding these artifactual segments by the existing approaches causes discontinuities in the reproduced signals which may introduce errors in subsequent analyses. To avoid this, the proposed method predicts multiple values of the artefactual region using long–short term memory network to recreate the temporal and spectral properties of the recorded signal. The method has been tested on two open-access data sets and incorporated into the open-access SANTIA (SigMate Advanced: a Novel Tool for Identification of Artefacts in Neuronal Signals) toolbox for community use.

Tremor is one of the cardinal symptoms of Parkinson’s disease. Currently, tremor severity is scored based on the Movement Disorders Society’s Unified Parkinson’s Disease Rating Scale, MDS-UPDRS, which is subjective and unreliable. Therefore, several studies have tried to measure tremor objectively using machine learning techniques. However, a limited number of studies have explored or compared medication state (ON or OFF) effect on objective measurement of tremor severity. Also, few studies have compared different types of wearable devices for tremor measurement. In this study, the medication state effect on tremor measurement is explored using different machine learning algorithms utilising different datasets that have been collected from different sensors. The results showed that the objective measurement of tremor severity is higher when patients are on medication using the Pebble smartwatch. The highest accuracy achieved was when patients were on medication and obtained 80% accuracy using Random Forest classifier, while the highest accuracy achieved when patients were off medication was 77% using Random Forest and Artificial Neural Network based on Multi-Layer Perceptron.

This research aims to present a technical analysis of the Traversal Optimisation Algorithm (TOA) for clustering and K-means clustering algorithm. The goal is to rigorously test this algorithm against different data specifications beyond what has previously been used with K-means without artificially and subjectively setting the initial number of clusters. The experimental evaluation involve the use of diverse cluster optimisation techniques for K-means while applying a wider range of internal validation methods such as Davies-Bouldin Index, Dunn Index and Silhouette Method, for appraising cluster quality of the Traversal Optimisation Algorithm, while at the same time not compromising the configuration of the default algorithm. The findings in this work shows that the optimisation algorithm’s clustering quality as calculated by multiple internal validity indices can be very poor when operating on datasets with varying characteristics. This is owing to the algorithm’s lack of any add-on mechanism for computing the optimal number of clusters that a dataset needs apriori. The results reveal that in a data processing contexts where the number of clusters are specified, the TOA yields a favourable cost-benefit in terms of run-time complexity and clustering quality.

Tremor is an indicative symptom of Parkinson’s disease (PD). Healthcare professionals have clinically evaluated the tremor as part of the Unified Parkinson’s disease rating scale (UPDRS) which is inaccurate, subjective and unreliable. In this study, a novel approach to enhance the tremor severity classification is proposed. The proposed approach is a combination of signal processing and resampling techniques; over-sampling, under-sampling and a hybrid combination. Resampling techniques are integrated with well-known classifiers, such as artificial neural network based on multi-layer perceptron (ANN-MLP) and random forest (RF). Advanced metrics are calculated to evaluate the proposed approaches such as area under the curve (AUC), geometric mean (Gmean) and index of balanced accuracy (IBA). The results show that over-sampling techniques performed better than other resampling techniques, also hybrid techniques performed better than under-sampling techniques. The proposed approach improved tremor severity classification significantly and show that the best approach to classify tremor severity is the combination of ANN-MLP with Borderline SMOTE which has obtained 93.81% overall accuracy, 96% Gmean, 91% IBA and 99% AUC. Besides, it is found that different resampling techniques performed differently with different classifiers.

Recent advancements in the Internet of Things and the miniaturisation of low-cost sensing devices allow for the unobtrusive collection of data for human activity recognition and behaviour modelling. A useful application of this in the context of ambient assisted living is in the monitoring of older adults daily for improved wellbeing and quality of life. The existing solutions are based on per-individual monitoring, therefore the systems are managed independently for each ambient intelligent environment. In this paper, we proposed a centralised system for the collective monitoring of individuals in a community. The proposed approach is based on a cloud-based solution where data collection and processing are centralised. Since the data are aggregated for all the residents, the system has the potential of promoting social interaction among the community residents. Additionally, the cost of the in-home monitoring system can be reduced since only the sensing devices are required for data collection, while the processing is carried out on the cloud infrastructure. This also reduces the tedious tasks required in setting up individual home monitoring systems. The role of assistive robots, the possibility of remote monitoring and potential challenges of the proposed approach are explored.

Fuzzy logic systems are customarily related to specific hardware or software systems. Nevertheless, it has been observed that distributed and cloud-based architectures of various intelligent systems are pouring intensifying attention. While the distributed architectures can potentially add values in developing fuzzy systems, a lack of standard methods and practices may limit their public use. This study aims to provide a standard solution for developing cloud-based service-oriented architectures for fuzzy logic systems, based on extending IEEE-1855 (2016) in the defining system and exchanging data. Experiments were performed employing simulation concerning collection, processing and monitoring of data in a distributed manner over the web. A real-time human activity recognition simulated scenario is also demonstrated through a cloud-based fuzzy system.

The process of recording local fields potentials can be influenced by many internal and external sources of electrical noise. To successfully use these recordings, noise must be removed, for which an automatic detection tool is needed to speed up the reviewing process. This research aims to develop an interpretable model, and among the many machine learning models, decision trees stand out due to the innate ability to allow insight into the classification criteria. As they require extracted features instead of the raw signal to achieve good performance, the adaptation of features was proposed originally intended for electroencephalography classification for the detection of artefacts in local fields potentials. Afterwards, they are grouped, and three different filtering feature selection algorithms are applied to obtain the most relevant ones and compare their choices. Classification accuracy of 88.1% is obtained with a single feature. Hence, it is concluded that interpretable models can be obtained with a performance similar to that of deep learning approaches.

Local fields potentials (LFP) can be contaminated by different internal and external sources of noise during their recordings. In cases where artefacts are present in them, automatic detection tools are needed to speed up the high accuracy detection process, followed by their removal to successfully use these recordings. This process is facilitated by a pool of supervised machine learning based tools which require labelled data for training. These algorithms have the capacity to distinguish between normal brain patterns and artefacts from an individual or a group, which is more flexible than template matching and subtraction. In addition, their portability has seen developments from both the software and hardware perspective. For many tools, LFP signal power is used as a gauge to measure and eventually label the artefacts portion of the LFP. This work explores how signal power affects the detection and classification accuracy of artefacts. Results show that a higher threshold value impacts positively the accuracy, due to less false positives in the data, without compromising the specificity in balanced datasets.

To support the independent living and improve the quality of life for the increasing ageing population, system for monitoring their daily routine and detecting anomalies in the routine is required. Existing anomaly detection systems are unable to identify the sources of the abnormalities, thereby hindering the development of adaptive monitoring systems with reduced false prediction rate. In this paper, an approach for identifying the sources of abnormalities in human activities of daily living is proposed. Anomalies are detected by modelling the existing activity data representing the usual behavioural routine of an individual to serve as a baseline model. Subsequent activities deviating from the baseline are then classified as outliers or anomalies. An ensemble of one-class support vector machine, isolation forest, robust covariance estimator and local outlier factor is utilised for the anomaly detection achieving an accuracy of $$98\%$$ 98% . The proposed approach for identifying anomaly sources is based on the concept of similarity measure using distance functions. Two methods for measuring the pairwise distance of the features of the activity data termed as one vs one similarity measure and one vs all similarity measure are proposed. Experimental evaluation of the proposed approach on activities of daily living datasets has shown the credibility of the proposed approach for utilisation in an in-home monitoring system.