Search

Tinnitus is the phantom perception of sound, experienced by 10%–15% of the global population. Computational models have been used to investigate the mechanisms underlying the generation of tinnitus-related activity. However, existing computational models have rarely benchmarked the modeled perception of a phantom sound against recorded data relating to a person’s perception of tinnitus characteristics, such as pitch or loudness. This article details the development of two perceptual models of tinnitus. The models are validated using empirical data from people with tinnitus and the models’ performance is compared with the existing perceptual models of tinnitus pitch. The first model extends existing perceptual models of tinnitus, while the second model utilizes an entirely novel approach to modeling tinnitus perception using a linear mixed effects (LMEs) model. The LME model is also used to model the perceived loudness of the phantom sound which has not been considered in previous models. The LME model creates an accurate model of tinnitus pitch and loudness and shows that both tinnitus-related activity and individual perception of sound are factors in the perception of the phantom sound that characterizes tinnitus.

Robots need to exploit high-quality information on grasped objects to interact with the physical environment. Haptic data can therefore be used for supplementing the visual modality. This paper investigates the use of Convolutional Neural Networks (CNN) and Long-Short Term Memory (LSTM) neural network architectures for object classification on Spatio-temporal tactile grasping data. Furthermore, we compared these methods using data from two different fingertip sensors (namely the BioTac SP and WTS-FT) in the same physical setup, allowing for a realistic comparison across methods and sensors for the same tactile object classification dataset. Additionally, we propose a way to create more training examples from the recorded data. The results show that the proposed method improves the maximum accuracy from 82.4 % (BioTac SP fingertips) and 90.7 % (WTS-FT fingertips) with complete time-series data to about 94 % for both sensor types.

Time series data from multiple modalities such as physiological and motion sensor data have proven to be integral for measuring mental wellbeing however individual differences between people limit the generalisability of deep learning models especially for those with intellectual disabilities. It is impractical, time consuming and extremely challenging to collect large realworld datasets of individuals’ wellbeing in their everyday life. Therefore, to address this challenge, we propose a Transfer Learning (TL) approach that develops personalised real-world affective models using few labelled samples by adapting a controlled stressor model. This approach to personalise models and improve cross-domain performance is completed on-device, automating the traditionally manual process saving time and labour. The results show adopting the TL approach significantly increased model performance with the multivariate physiological and motion affective model achieving an average accuracy of 93.5 % compared with the comparative non-TL model accuracy of 71.7%. The proposed methodology helps overcome problems with affective model personalisation, thus improving on the performance of conventional deep learning methods.

Spiking neural networks (SNNs) as the third generation of artificial neural networks are closer to their biological counterparts than their predecessors. SNNs have a higher computational capacity and lower power requirements than networks of sigmoidal neurons. In this paper, a new spiking neural network for pattern classification referred to as Degree of Belonging SNN (DoB-SNN) is introduced. DoB-SNN is inspired by a neuronal assembly where each neuron has a degree of belonging to every class of data being process. DoB-SNN clusters the neurons during the training process using DoBs to allocate a group of neurons to each class. A new training algorithm is presented to adjust DoBs along with the network's synaptic weights, based on Spike-Timing Dependent Plasticity (STDP) and neurons' activity for training samples. The performance of DoB-SNN is evaluated on five datasets from the UCI machine learning repository. Nested Cross-Validation is employed to determine the network's hyperparameters for each dataset and thoroughly assess generalisation capability. A detailed comparison on these datasets with three other supervised learning algorithms, including SpikeProp, SWAT, and SRESN is provided. The results show that no algorithm significantly outperforms DoB-SNN, Whereas DoB-SNN has significantly better performance than others for Liver disorders dataset (>6.10%, $p ). Accuracies obtained by DoB-SNN are significantly greater than SWAT for both Iris and Breast Cancer (>1.69%, $p ) and significantly better than SpikeProp for Iris (1.62%, $p=0.04$ ). In all comparisons, DoB-SNN used the smallest network, among others. DoB-SNN therefore offers significant potential as alternative SNN architecture and learning algorithm.

Joint manipulation and object exchange are common in many everyday scenarios. Although they are trivial tasks for humans, they are still very challenging for robots. Existing approaches for robot-to-human object handover assume that there is no fault during the transfer. However, unintentional perturbation forces can be occasionally applied to the object, resulting in the robot and the object being damaged, for example by being dropped. In this paper we present a novel approach to handover objects in a reliable manner while ensuring the safety of the robot and the object. Relying on tactile sensing, the system uses an effort controller to adapt the grasp forces in the presence of perturbations. Moreover, the proposed approach identifies a perturbation being applied on the object. When a perturbation event is detected, the algorithm classifies the direction of the pulling forces to decide whether to release it or not. The reliable handover system was implemented using a Shadow Robot hand equipped with BioTAC tactile sensors. Our results show that the system correctly adapts to the forces applied on the object to maintain the grasp and only releases the object if the human receiver pulls in the right direction.

Artificial vision using computational models that can mimic biological vision is an area of ongoing research. One of the main themes within this research is the study of the retina and in particular, retinal ganglion cells which are responsible for encoding the visual stimuli. A common approach to modelling the internal processes of retinal ganglion cells is the use of a linear - non-linear cascade model, which models the cell's response using a linear filter followed by a static non-linearity. However, the resulting model is generally restrictive as it is often a poor estimator of the neuron's response. In this paper we present an alternative to the linear - non-linear model by modelling retinal ganglion cells using a number of machine learning techniques which have a proven track record for learning complex non-linearities in many different domains. A comparison of the model predicted spike rate shows that the machine learning models perform better than the standard linear - non-linear approach in the case of temporal white noise stimuli.

A Deep Boltzmann Machine is a model of a Deep Neural Network formed from multiple layers of neurons with nonlinear activation functions. The structure of a Deep Boltzmann Machine enables it to learn very complex relationships between features and facilitates advanced performance in learning of high-level representation of features, compared to conventional Artificial Neural Networks. Feature selection at the input level of Deep Neural Networks has not been well studied, despite its importance in reducing the input features processed by the deep learning model, which facilitates understanding of the data. This paper proposes a novel algorithm, Deep Feature Selection (Deep-FS), which is capable of removing irrelevant features from large datasets in order to reduce the number of inputs which are modelled during the learning process. The proposed Deep-FS algorithm utilizes a Deep Boltzmann Machine, and uses knowledge which is acquired during training to remove features at the beginning of the learning process. Reducing inputs is important because it prevents the network from learning the associations between the irrelevant features which negatively impact on the acquired knowledge of the network about the overall distribution of the data. The Deep-FS method embeds feature selection in a Restricted Boltzmann Machine which is used for training a Deep Boltzmann Machine. The generative property of the Restricted Boltzmann Machine is used to reconstruct eliminated features and calculate reconstructed errors, in order to evaluate the impact of eliminating features. The performance of the proposed approach was evaluated with experiments conducted using the MNIST, MIR-Flickr, GISETTE, MADELON and PANCAN datasets. The results revealed that the proposed Deep-FS method enables improved feature selection without loss of accuracy on the MIR-Flickr dataset, where Deep-FS reduced the number of input features by removing 775 features without reduction in performance. With regards to the MNIST dataset, Deep-FS reduced the number of input features by more than 45%; it reduced the network error from 0.97% to 0.90%, and also reduced processing and classification time by more than 5.5%. Additionally, when compared to classical feature selection methods, Deep-FS returned higher accuracy. The experimental results on GISETTE, MADELON and PANCAN showed that Deep-FS reduced 81%, 57% and 77% of the number of input features, respectively. Moreover, the proposed feature selection method reduced the classifier training time by 82%, 70% and 85% on GISETTE, MADELON and PANCAN datasets, respectively. Experiments with various datasets, comprising a large number of features and samples, revealed that the proposed Deep-FS algorithm overcomes the main limitations of classical feature selection algorithms. More specifically, most classical methods require, as a prerequisite, a pre-specified number of features to retain, however in Deep-FS this number is identified automatically. Deep-FS performs the feature selection task faster than classical feature selection algorithms which makes it suitable for deep learning tasks. In addition, Deep-FS is suitable for finding features in large and big datasets which are normally stored in data batches for faster and more efficient processing.

Mental health problems are on the rise globally and strain national health systems worldwide. Mental disorders are closely associated with fear of stigma, structural barriers such as financial burden, and lack of available services and resources which often prohibit the delivery of frequent clinical advice and monitoring. Technologies for mental well-being exhibit a range of attractive properties, which facilitate the delivery of state-of-the-art clinical monitoring. This review article provides an overview of traditional techniques followed by their technological alternatives, sensing devices, behaviour changing tools, and feedback interfaces. The challenges presented by these technologies are then discussed with data collection, privacy, and battery life being some of the key issues which need to be carefully considered for the successful deployment of mental health toolkits. Finally, the opportunities this growing research area presents are discussed including the use of portable tangible interfaces combining sensing and feedback technologies. Capitalising on the data these ubiquitous devices can record, state of the art machine learning algorithms can lead to the development of robust clinical decision support tools towards diagnosis and improvement of mental well-being delivery in real-time.

Model optimisation is a key step in model development and traditionally this was limited to parameter tuning. However, recent developments and enhanced understanding of internal dynamics of model architectures have led to various exploration to optimise and enhance performance through model extension and development. In this paper, we extend the architecture of the Multi-recurrent Neural Network (MRN) to incorporate self-learning recurrent link ratios and periodically attentive hidden units. We contrast and show the superiority of these extensions to the standard MRN for a complex financial prediction task. The superiority is attributed to i) the ability of the self-learning recurrent link ratios to dynamically utilise data to identify optimal parameters of its memory mechanism and ii) the periodically attentive units enabling the hidden layer capture temporal features that are sensitive to different periods of time. Finally, we evaluate our extended MRNs (Self-Learning MRN (SL-MRN) and Periodically Attentive MRN (PA-MRN)), against two current state-of the-art models (Long-Short Term Memory and Support Vector Machines) for an eye state detection task. Our preliminary results demonstrate that the PA-MRN and SL-MRN outperform both state-of-the-art models. These results demonstrate that the MRN extensions are suitable models for machine learning applications and these findings would be further explored.

The work was funded by The Leverhulme Trust Research Project Grant RPG-2016-252 entitled “Novel Approaches for Constructing Optimised Multimodal Data Spaces”. Ensemble models achieve high accuracy by combining a number of base estimators and can increase the reliability of machine learning compared to a single estimator. Additionally, an ensemble model enables a machine learning method to deal with imbalanced data, which is considered to be one of the most challenging problems in machine learning. In this paper, the capability of Adaptive Boosting (AdaBoost) is integrated with a Convolutional Neural Network (CNN) to design a new machine learning method, AdaBoost-CNN, which can deal with large imbalanced datasets with high accuracy. AdaBoost is an ensemble method where a sequence of classifiers is trained. In AdaBoost, each training sample is assigned a weight, and a higher weight is set for a training sample that has not been trained by the previous classifier. The proposed AdaBoost-CNN is designed to reduce the computational cost of the classical AdaBoost when dealing with large sets of training data, through reducing the required number of learning epochs for its ingredient estimator. AdaBoost-CNN applies transfer learning to sequentially transfer the trained knowledge of a CNN estimator to the next CNN estimator, while updating the weights of the samples in the training set to improve accuracy and to reduce training time. Experimental results revealed that the proposed AdaBoost-CNN achieved 16.98% higher accuracy compared to the classical AdaBoost method on a synthetic imbalanced dataset. Additionally, AdaBoost-CNN reached an accuracy of 94.08% on 10,000 testing samples of the synthetic imbalanced dataset, which is higher than the accuracy of the baseline CNN method, i.e. 92.05%. AdaBoost-CNN is computationally efficient, as evidenced by the fact that the training simulation time of the proposed method is 47.33 seconds, which is lower than the training simulation time required for a similar AdaBoost method without transfer learning, i.e. 225.83 seconds on the imbalanced dataset. Moreover, when compared to the baseline CNN, AdaBoost-CNN achieved higher accuracy when applied to five other benchmark datasets including CIFAR-10 and Fashion-MNIST. AdaBoost-CNN was also applied to the EMNIST datasets, to determine its impact on large imbalanced classes, and the results demonstrate the superiority of the proposed method compared to CNN.

Crude oil is fundamental for global growth and stability. The factors influencing crude oil prices and more generally, the oil market, are well known to be dynamic, volatile and evolving. Subsequently, crude oil prediction is a complex and notoriously difficult task. In this paper, we evaluate the Multi-recurrent Network (MRN), a simple yet powerful recurrent neural network, for oil price forecasting at various forecast horizons. Although similar models, such as Long Short-Term Memory (LSTM) networks, have shown some success in this domain, the MRN is a comparatively simplified neural network model which exhibits complex state-based memories that are both flexible and rigid. We evaluate the MRN against the standard Feedforward Multilayered Perceptron (FFMLP) and the Simple Recurrent Network (SRN) in addition to the current state-of-the-art LSTM for specifically modelling the shocks in oil prices caused by the financial crisis. The in-sample data consists of key indicator variables sampled across the pre-financial crisis period (July 1969 to September 2003) and the out-sample data used to evaluate the models, is before, during and beyond the crisis (October 2003 to March 2015). We show that such simple sluggish state-based models are superior to the FFMLP, SRN and LSTM models. Furthermore, the MRN appears to have discovered important latent features embedded within the input signal five years prior to the 2008 financial crisis. This suggests that the indicator variables could provide Central Banks and governments with early warning indicators of impending financial perturbations which we consider an invaluable finding and worthy of further exploration.

Computer vision has been revolutionised in recent years by increased research in convolutional neural networks (CNNs); however, many challenges remain to be addressed in order to ensure fast and accurate image processing when applying these techniques to robotics. These challenges consist of handling extreme changes in scale, illumination, noise, and viewing angles of a moving object. The project main contribution is to provide insight on how to properly train a convolutional neural network (CNN), a specific type of DNN, for object tracking in the context of industrial robotics. The proposed solution aims to use a combination of documented approaches to replicate a pick-and-place task with an industrial robot using computer vision feeding a YOLOv3 CNN. Experimental tests, designed to investigate the requirements of training the CNN in this context, were performed using a variety of objects that differed in shape and size in a controlled environment. The general focus was to detect the objects based on their shape; as a result, a suitable and secure grasp could be selected by the robot. The findings in this article reflect the challenges of training the CNN through brute force. It also highlights the different methods of annotating images and the ensuing results obtained after training the neural network.

A system that can recognise materials based on tactile sensing alone is presented.Identification achieved by analysis of thermal conductivity and surface texture.Evaluation of various machine learning algorithms and developed hybrid classifiers.The system outperforms humans in the task of material recognition. Identification of the material from which an object is made is of significant value for effective robotic grasping and manipulation. Characteristics of the material can be retrieved using different sensory modalities: vision based, tactile based or sound based. Compressibility, surface texture and thermal properties can each be retrieved from physical contact with an object using tactile sensors. This paper presents a method for collecting data using a biomimetic fingertip in contact with various materials and then using these data to classify the materials both individually and into groups of their type. Following acquisition of data, principal component analysis (PCA) is used to extract features. These features are used to train seven different classifiers and hybrid structures of these classifiers for comparison. For all materials, the artificial systems were evaluated against each other, compared with human performance and were all found to outperform human participants average performance. These results highlighted the sensitive nature of the BioTAC sensors and pave the way for research that requires a sensitive and accurate approach such as vital signs monitoring using robotic systems.

The ability to quickly and accurately triage a person’s medical condition in an emergency situation or other critical scenarios could mean the difference between life and death. Endowing a robotic system with vision and tactile capabilities, similar to those of medical professionals, and thus enabling robots to assess a patient’s status in an emergency is a highly sought after characteristic in healthcare robotics. This paper presents a novel fuzzy triage system exploiting visual and tactile sensing, to equip a robot with the skills to accurately determine key vital signs in humans. There are three key signs of human health: respiratory rate, pulse rate (Beats Per Minute (BPM)) and capillary refill time. Using ground truth from a medical professional, the fuzzy triage system is trained and validated initially with informed synthetic data and then further evaluated using vital signs data collected from subjects in a pilot study. Results from this pilot study indicate that the fuzzy triage system is capable of classifying a patient’s health using the novel approaches for collecting BPM, Respiratory Rate (RR) and Capillary Refill Time (CRT) which replicate, to some extent, the approaches used by medical professionals for measuring vital signs. Furthermore, the intelligent system proved capable of determining whether a pulse was regular or arrhythmic, whether respiratory rate was regular or irregular, and determining the subject’s capillary refill time. Such results imply that this system could ultimately be used, for example, in a home assistance robot for elderly or disabled persons, or as a first responder robot. Ultimately the aim would be that these methods could be utilised by robotic systems in emergency scenarios or disaster zones.

This paper presents an enhanced rank-order-based learning algorithm, called SpikeTemp, for spiking neural networks (SNNs) with a dynamically adaptive structure. The trained feed-forward SNN consists of two layers of spiking neurons: 1) an encoding layer which temporally encodes real-valued features into spatio-temporal spike patterns and 2) an output layer of dynamically grown neurons which perform spatio-temporal classification. Both Gaussian receptive fields and square cosine population encoding schemes are employed to encode real-valued features into spatio-temporal spike patterns. Unlike the rank-order-based learning approach, SpikeTemp uses the precise times of the incoming spikes for adjusting the synaptic weights such that early spikes result in a large weight change and late spikes lead to a smaller weight change. This removes the need to rank all the incoming spikes and, thus, reduces the computational cost of SpikeTemp. The proposed SpikeTemp algorithm is demonstrated on several benchmark data sets and on an image recognition task. The results show that SpikeTemp can achieve better classification performance and is much faster than the existing rank-order-based learning approach. In addition, the number of output neurons is much smaller when the square cosine encoding scheme is employed. Furthermore, SpikeTemp is benchmarked against a selection of existing machine learning algorithms, and the results demonstrate the ability of SpikeTemp to classify different data sets after just one presentation of the training samples with comparable classification performance.

The file attached to this record is the author's final peer reviewed version. The Publisher's final version can be found by following the DOI link. Artificial neural networks have been used as a powerful processing tool in various areas such as pattern recognition, control, robotics, and bioinformatics. Their wide applicability has encouraged researchers to improve artificial neural networks by investigating the biological brain. Neurological research has significantly progressed in recent years and continues to reveal new characteristics of biological neurons. New technologies can now capture temporal changes in the internal activity of the brain in more detail and help clarify the relationship between brain activity and the perception of a given stimulus. This new knowledge has led to a new type of artificial neural network, the Spiking Neural Network (SNN), that draws more faithfully on biological properties to provide higher processing abilities. A review of recent developments in learning of spiking neurons is presented in this paper. First the biological background of SNN learning algorithms is reviewed. The important elements of a learning algorithm such as the neuron model, synaptic plasticity, information encoding and SNN topologies are then presented. Then, a critical review of the state-of-the-art learning algorithms for SNNs using single and multiple spikes is presented. Additionally, deep spiking neural networks are reviewed, and challenges and opportunities in the SNN field are discussed.

Accurate estimation of interactions in neuronal circuits is critical in understanding neural information processing and the neuronal dynamics of emergent networks. Transfer entropy(TE) is a model-free information theoretic measure of flow of information between two random processes. TE has recently gained much popularity due to its effectiveness in estimating effective connectivity among neurons from simulated networks. However, experimental recordings inherently lack the ground truth information of neural connectivity - making it difficult to identify true connections from spurious connections. To tackle this problem, we present a superimposition method where the estimated connectivity is superimposed onto the spatial firing density plot. The firing density plot, on itself, does not provide connectivity information but we assume that frequently firing channels are more likely to have more active interactions. A neuron’s firing depends on the presynaptic inputs, and highly interconnected neurons tend to have more presynaptic inputs resulting in more frequent firing. The firing density plot is organised in a spatial layout, which corresponds substantially to the structural layout. We demonstrated that the connectivity estimated closely corresponds to the firing density plot when superimposed onto the firing density plot. To strengthen the arguement, we analyse the firing count and the connectivity inferred for each randomly sampled channel and found a positively correlated relationship.

Biological image processing is performed by complex neural networks composed of thousands of neurons interconnected via thousands of synapses, some of which are excitatory and others inhibitory. Spiking neural models are distinguished from classical neurons by being biological plausible and exhibiting the same dynamics as those observed in biological neurons. This paper proposes a Natural Convolutional Neural Network (NatCSNN) which is a 3-layer bio-inspired Convolutional Spiking Neural Network (CSNN), for classifying objects extracted from natural images. A two-stage training algorithm is proposed using unsupervised Spike Timing Dependent Plasticity (STDP) learning (phase 1) and ReSuMe supervised learning (phase 2). The NatCSNN was trained and tested on the CIFAR-10 dataset and achieved an average testing accuracy of 84.7% which is an improvement over the 2-layer neural networks previously applied to this dataset.

Multi-modal data extracted from different sensors in a smart home can be fused to build models that recognize the daily living activities of residents. This paper proposes a Deep Convolutional Neural Network to perform the activity recognition task using the multi-modal data collected from a smart residential home. The dataset contains accelerometer data (composed of three perpendicular components of acceleration and the strength of the accelerometer signal received by four receivers), video data (15 time-series related to 2D and 3D center of mass and bounding box extracted from an RGB-D camera), and Passive Infra-Red sensor data. The performance of the Deep Convolutional Neural Network is compared to the Deep Belief Network. Experimental results revealed that the Deep Convolutional Neural Network with two pairs of convolutional and max pooling layers achieved better classification accuracy than the Deep Belief Network. The Deep Belief Network uses Restricted Boltzmann Machines for pre-training the network. When training deep learning models using classes with a high number of training samples, the DBN achieved 65.97% classification accuracy, whereas the CNN achieved 75.33% accuracy. The experimental results demonstrate the challenges of dealing with multi-modal data and highlight the importance of having a suitable number of samples within each class for sufficiently training and testing deep learning models.

Obstacle avoidance is a critical aspect of control for a mobile robot when navigating towards a goal in an unfamiliar environment. Where traditional methods for obstacle avoidance depend heavily on path planning to reach a specific goal location whilst avoiding obstacles, the method proposed uses an adaption of a potential fields algorithm to successfully avoid obstacles whilst the robot is being guided to a non-specific goal location. Details of a generalised finite state machine based control algorithm that enable the robot to pursue a dynamic goal location, whilst avoiding obstacles and without the need for specific path planning, are presented. We have developed a novel potential fields algorithm for obstacle avoidance for use within Robot Operating Software (ROS) and made it available for download within the open source community. We evaluated the control algorithm in a high-speed predator-prey scenario in which the predator could successfully catch the moving prey whilst avoiding collision with all obstacles within the environment.

The raw EEG signal acquired non-invasively from the sensorimotor cortex during the motor imagery (MI) performed by a brain-computer interface (BCI) user is naturally embedded with noise while the actual noise-free EEG is still unattainable. This paper compares the enhancement in information when filtering these noisy EEG signals while using a Schrodinger wave equation (SWE) based Recurrent Quantum Neural Network (RQNN) model and a Savitzky-Golay (SG) filtering model, while investigating over multiple classification techniques on several datasets. The RQNN model is designed to efficiently capture the statistical behavior of the input signal using an unsupervised learning scheme. The algorithm is robust to parametric sensitivity and does not make any a priori assumption about the true signal type or the embedded noise. The performance of both the filtering approaches, investigated for the BCI competition IV 2b dataset as well as the offline datasets on subjects in the BCI laboratory, over multiple classifiers shows that the RQNN can potentially be a flexible technique that can suit different classifiers for real-time EEG signal filtering. The average classification accuracy performance across all the subjects with the RQNN technique is better than the SG (and the unfiltered signal) by approximately 5% (and 7%) and 1% (and 4%) during the training and the evaluation stages respectively.

Tinnitus is a problem that affects a diverse range of people. One common trait amongst people with tinnitus is the presence of hearing loss, which is apparent in over 90% of the cohort. It is postulated that the remainder of people with tinnitus have hidden hearing loss in the form of cochlear synaptopathy. The loss of hearing sensation is thought to cause a reduction in the bottom-up excitatory signals of the auditory pathway leading to a change in the frequency of thalamocortical oscillations known as thalamocortical dysrhythmia (TCD). The downward shift in oscillatory behavior, characteristic of TCD, has been recorded experimentally but the underlying mechanisms responsible for TCD in tinnitus subjects cannot be directly observed. This paper investigates these underlying mechanisms by creating a biologically faithful model of the auditory periphery and thalamocortical network, called the central auditory processing (CAP) model. The proposed model replicates tinnitus related activity in the presence of hearing loss and hidden hearing loss in the form of cochlear synaptopathy. The results of this paper show that, both the bottom-up and top-down changes are required in the auditory system for tinnitus related hyperactivity to coexist with TCD, contrary to the theoretical model for TCD. The CAP model provides a novel modeling approach to account for tinnitus related activity with and without hearing loss. Moreover, the results provide additional clarity to the understanding of TCD and tinnitus and provide direction for future approaches to treating tinnitus.

Real time classification of objects using computer vision techniques are becoming relevant with emergence of advanced perceptions systems required by, surveillance systems, industry 4.0 robotics and agricultural robots. Conventional video surveillance basically detects and tracks moving object whereas there is no indication of whether the object is approaching or receding the camera (looming). Looming detection and classification of object movements aids in knowing the position of the object and plays a crucial role in military, vehicle traffic management, robotics, etc. To accomplish real-time object trajectory classification, a contour tracking algorithm is necessary. In this paper, an application is made to perform looming detection and to detect imminent collision on a system-on-chip field-programmable gate array (SoC- FPGA) hardware. The work presented in this paper was designed for running in Robotic platforms, Unmanned Aerial Vehicles, Advanced Driver Assistance System, etc. Due to several advantages of SoC-FPGA the proposed work is performed on the hardware. The proposed work focusses on capturing images, processing, classifying the movements of the object and issues an imminent collision warning on-the-fly. This paper details the proposed software algorithm used for the classification of the movement of the object, simulation of the results and future work.

Determining the receptive field of a retinal ganglion cell is critically important when formulating a computational model that maps the relationship between the stimulus and response. This process is traditionally undertaken using reverse correlation to estimate the receptive field. By stimulating the retina with artificial stimuli, such as alternating checkerboards, bars or gratings and recording the neural response it is possible to estimate the cell’s receptive field by analysing the stimuli that produced the response. Artificial stimuli such as white noise is known to not stimulate the full range of the cell’s responses. By using natural image stimuli, it is possible to estimate the receptive field and obtain a resulting model that more accurately mimics the cells’ responses to natural stimuli. This paper extends on previous work to seek further improvements in estimating a ganglion cell’s receptive field by considering that the receptive field can be divided into subunits. It is thought that these subunits may relate to receptive fields which are associated with bipolar retinal cells. The findings of this preliminary study show that by using subunits to define the receptive field we achieve a significant improvement over existing approaches when deriving computational models of the cell’s response.

The retina performs the earlier stages of image processing in living beings and is composed of six different groups of cells, namely, the rods, cones, horizontal, bipolar, amacrine and ganglion cells. Each of those group of cells can be sub-divided into other types of cells that vary in shape, size, connectivity and functionality. Each cell is responsible for performing specific tasks in these early stages of biological image processing. Some of those cells are sensitive to horizontal and vertical movements. This paper proposes a multi-hierarchical spiking neural network architecture for detecting horizontal and vertical movements using a custom dataset which was generated in laboratory settings. The proposed architecture was designed to reflect the connectivity, behaviour and the number of layers found in the majority of vertebrates retinas, including humans. The architecture was trained using 2303 images and tested using 816 images. Simulation results revealed that each cell model is sensitive to vertical and horizontal movements with a detection error of 6.75 percent.

A human's Capillary Refill Time (CRT) is a key indicator of their current health status. Being able to accurately assess a human's cardiovascular system peripherally by assessing their CRT in an emergency or search and rescue situation could, in critical scenarios, mean the difference between life and death. This paper presents a novel algorithm that enables a Shadow Robot Hand equipped with BioTAC biomimetic tactile fingertip sensors and a red, green, blue (RGB) camera to measure the CRT of humans by making contact with their forehead, regardless of their skin tone. The method presented replicates, to some extent, the methods carried out by medical professionals when measuring CRT and could be used to equip a first responder robot. Furthermore, the algorithms determine whether a person has a healthy cardiovascular system or whether the blood supply has been cut off from the skin indicating various issues such as shock or severe dehydration. The method presented in this work allows for a more accurate measurement of CRT than that of a medical professional.

Biometric systems detect authenticity based on users' distinct physiological or behavioral characteristics for purposes of identification and access control. These pattern recognition systems are difficult to bypass when compared to traditional token or password based systems. This paper is proposing a new deep learning architecture for fingerprint recognition. The proposed architecture comprises of a pre-processing stage for extracting texture features from fingerprints, and this stage is performed by using histogram equalization, Gabor enhancement and fingerprint thinning. The pre-processed fingerprints are input into a Deep Convolutional Neural Network classifier. The proposed approach has achieved 98.21% classification accuracy with 0.9 loss. The obtained accuracy is significantly higher than previously reported results on the same dataset, 77%.

Predictive modeling in an electrical power systems is currently gaining momentum especially as Phasor Measurement Units (PMUs) are being deployed in modern electrical grids to replace the Supervisory Control and Data Acquisition System (SCADA). This paper evaluates machine learning algorithms for the task of monitoring voltage instability for online decision making. In particular the performance of the Naive Bayesian, K-Nearest Neighbors, Decision Tree and Ensembles classifiers (XGBoost, Bagging, Random Forest, and AdaBoost) were compared. Performance evaluation measures of Precision, Recall, F1-score, and Accuracy were adopted to evaluate the performance of the classifiers. In this paper, a number of voltage stability operating points were generated with different variations of load/generation, using the PSSE Power-Voltage (PV) analysis tool. An IEEE 39 bus was used as a test system. Sufficient training patterns that captured different Operating Points (OPs) at base-case and at multiple contingencies (N-k) were gathered to train machine learning methods to identify acceptable operating conditions and near collapse situations. Experimental results show that AdaBoost achieved the highest classification accuracy, i.e. 96.02%, compared to the other classifiers.

Hand gestures provide a natural way for humans to interact with computers to perform a variety of different applications. However, factors such as the complexity of hand gesture structures, differences in hand size, hand posture, and environmental illumination can influence the performance of hand gesture recognition algorithms. Recent advances in Deep Learning have significantly advanced the performance of image recognition systems. In particular, the Deep Convolutional Neural Network has demonstrated superior performance in image representation and classification, compared to conventional machine learning approaches. This paper proposes an Adapted Deep Convolutional Neural Network (ADCNN) suitable for hand gesture recognition tasks. Data augmentation is initially applied which shifts images both horizontally and vertically to an extent of 20% of the original dimensions randomly, in order to numerically increase the size of the dataset and to add the robustness needed for a deep learning approach. These images are input into the proposed ADCNN model which is empowered by the presence of network initialization (ReLU and Softmax) and L2 Regularization to eliminate the problem of data overfitting. With these modifications, the experimental results using the ADCNN model demonstrate that it is an effective method of increasing the performance of CNN for hand gesture recognition. The model was trained and tested using 3750 static hand gesture images, which incorporate variations in features such as scale, rotation, translation, illumination and noise. The proposed ADCNN was compared to a baseline Convolutional Neural Network and the results show that the proposed ADCNN achieved a classification recognition accuracy of 99.73%, and a 4% improvement over the baseline Convolutional Neural Network model (95.73%).

Inflammatory Bowel Disease (IBD) is an umbrella term for a group of inflammatory diseases of the human gastrointestinal tract, including Crohn’s Disease (CD) and ulcerative colitis (UC). Changes to the intestinal microbiome, the community of micro-organisms that resides in the human gut, have been shown to contribute to the pathogenesis of IBD. IBD diagnosis is often delayed due its non-specific symptoms (e.g. abdominal pain) and an invasive colonoscopy is required for confirmation. Delayed diagnosis is linked to poor growth in children and worse treatment outcomes. Microbial communities are extremely complex and feature selection algorithms are often applied to identify key bacterial groups that drive disease. It has been shown that aggregating Ensemble Feature Selection (EFS) can be used to improve the robustness of feature selection algorithms. The robustness of a feature selector is defined as the variation of feature selector output caused by small changes to the dataset. Typical feature selection algorithms can be used to help build simpler, faster, and easier to understand models - but suffer from poor robustness. Having confidence in the output of a feature selector algorithm is key for enabling knowledge discovery from complex biological datasets. In this work we apply a two-step filter and an EFS process to generate robust feature subsets that can non-invasively predict IBD subtypes from high-resolution microbiome data. The predictive power of the robust feature subsets is the highest reported in literature to date. Furthermore, we identify five biologically plausible bacterial species that have not previously been implicated in IBD aetiology.

A system of interacting agents is, by definition, very demanding in terms of computational resources. Although multi-agent systems have been used to solve complex problems in many areas, it is usually very difficult to perform large-scale simulations in their targeted serial computing platforms. Reconfigurable hardware, in particular Field Programmable Gate Arrays devices, have been successfully used in High Performance Computing applications due to their inherent flexibility, data parallelism, and algorithm acceleration capabilities. Indeed, reconfigurable hardware seems to be the next logical step in the agency paradigm, but only a few attempts have been successful in implementing multi-agent systems in these platforms. This article discusses the problem of inter-agent communications in Field Programmable Gate Arrays. It proposes a Network-on-Chip in a hierarchical star topology to enable agents’ transactions through message broadcasting using the Open Core Protocol as an interface between hardware modules. A customizable router microarchitecture is described and a multi-agent system is created to simulate and analyse message exchanges in a generic heavy traffic load agent-based application. Experiments have shown a throughput of 1.6Gbps per port at 100MHz without packet loss and seamless scalability characteristics.

Intrinsic motivations play an important role in human learning, particularly in the early stages of childhood development, and ideas from this research field have influenced robotic learning and adaptability. In this paper we investigate one specific type of intrinsic motivation, that of novelty detection and we discuss the reasons that make it a powerful facility for continuous learning. We formulate and present one original type of biologically inspired novelty detection architecture and implement it on a robotic system engaged in a perceptual classification task. The results of real-world robot experiments we conducted show how this original architecture conforms to behavioural observations and demonstrate its effectiveness in terms of focusing the system's attention in areas that are potential for effective learning. We proposed novelty detection as intrinsic motivation for autonomous learning.Novelty detection model is bio-inspired by the behavioural phenomenon of habituation.We demonstrated the effectiveness of the model in a real-world experiment.We showed how habituation is used as an intrinsic motivation for robot learning.

Price manipulation refers to the activities of those traders who use carefully designed trading behaviors to manually push up or down the underlying equity prices for making profits. With increasing volumes and frequency of trading, price manipulation can be extremely damaging to the proper functioning and integrity of capital markets. The existing literature focuses on either empirical studies of market abuse cases or analysis of particular manipulation types based on certain assumptions. Effective approaches for analyzing and detecting price manipulation in real time are yet to be developed. This paper proposes a novel approach, called adaptive hidden Markov model with anomaly states (AHMMAS) for modeling and detecting price manipulation activities. Together with wavelet transformations and gradients as the feature extraction methods, the AHMMAS model caters to price manipulation detection and basic manipulation type recognition. The evaluation experiments conducted on seven stock tick data from NASDAQ and the London Stock Exchange and 10 simulated stock prices by stochastic differential equation show that the proposed AHMMAS model can effectively detect price manipulation patterns and outperforms the selected benchmark models.

Various ambient assisted living (AAL) technologies have been proposed for improving the living conditions of elderly people. One of them is to introduce robots to reduce dependency on support staff. The tasks commonly encountered in a healthcare facility such as a care home for elderly people are heterogeneous and are of different priorities. A care home environment is also dynamic and new emergency priority tasks, which if not attended shortly may result in fatal situations, may randomly appear. Therefore, it is better to use a multi-robot system (MRS) consisting of heterogeneous robots than designing a single robot capable of doing all tasks. An efficient task allocation algorithm capable of handling the dynamic nature of the environment, the heterogeneity of robots and tasks, and the prioritisation of tasks is required to reap the benefits of introducing an MRS. This paper proposes Consensus Based Parallel Auction and Execution (CBPAE), a distributed algorithm for task allocation in a system of multiple heterogeneous autonomous robots deployed in a healthcare facility, based on auction and consensus principles. Unlike many of the existing market based task allocation algorithms, which use a time extended allocation of tasks before the actual execution is initialised, the proposed algorithm uses a parallel auction and execution framework, and is thus suitable for highly dynamic real world environments. The robots continuously resolve any conflicts in the bids on tasks using inter-robot communication and a consensus process in each robot before a task is assigned to a robot. We demonstrate the effectiveness of the CBPAE by comparing its simulation results with those of an existing market based distributed multi-robot task allocation algorithm and through experiments on real robots.

A novel online learning algorithm for Spiking Neural Networks (SNNs) with dynamically adaptive structure is presented. The main contribution of this work lies in the fact that the proposed adaptive SNN is able to classify spike-based spatio-temporal inputs after just one presentation of the training set, i.e. in one pass only, and does not require the entire training set to be available at once. Both the structure and weights of the SNN are learned dynamically through a combination of unsupervised and supervised learning paradigms. The proposed feed-forward SNN consists of three layers of spiking neurons: an input layer which temporally encodes real valued features into spike-based spatio-temporal patterns, a hidden layer of dynamically grown and pruned neurons which perform spatio-temporal clustering, and an output layer for classification. An unsupervised spiking-based clustering algorithm is implemented by the hidden layer whose spiking neurons are trained to compute a temporal Radial Basis Function (RBF) where incoming inputs will selectively activate hidden neurons based on how close the inputs are to the preferred inputs of the hidden neurons. The centre of each hidden RBF spiking neuron is represented by its time to first spike. In addition, a growing and pruning strategy is proposed to adjust the structure of the hidden layer 'on-the-fly' as inputs are presented to the SNN. Both the weights and the centres of the hidden RBF neurons are learned in an unsupervised way and classification at the output layer is achieved through supervised learning where the learning windows proposed for STDP and anti-STDP are used to adjust the weights of the output neurons afferent connections. Competition at both the hidden and the output layers is achieved through the use of lateral inhibitory connections between the neurons of each layer. The proposed online learning algorithm is validated on several benchmark datasets. The evaluation results demonstrate that SNNs trained with the proposed approach require only one pass through the training set in order to classify the inputs with comparable accuracies to existing SNN-based approaches as well as traditional representative classifiers.

It has been found that dysregulation in the orexin/hypocretin (Ox/HCRT) neuropeptide system in the lateral hypothalamus (LHA) is known to affect sleep disorder, depression and motor activities. However, to date there is no common agreement regarding the resulting specific changes induced in the Ox system. In this study, we inject corticosterone to produce stress-induced depressed mice and investigate the Ox neuronal and corresponding behavioural changes. Different doses (10, 20, 50mg/kgbw) of corticosterone were injected in adult mice, and then were tested in the open field test, forced swim test, tail suspension test, elevated plus maze test and motor activity measurements to validate the depressed animal model. Significant dose-dependent behavioural changes were observed in correlation with the doses of corticosterone. The effect is most significant and robust in the high 50mg/kgbw dose group five weeks after injection. Interestingly, we found on average a reduction in motor activity during the 12-hour dark phase (awake) of the depressed mice and no significant change during the light phase (asleep). Finally, using confocal microscopy, immunofluorescence (IF) analysis shows a significant increase (∼20%) in the number of Ox neurons in the LHA of the depressed mice as compared to the age-matched controls. This study suggests that an increase in Ox neuronal signaling may be functionally linked to high and prolonged external stress-induced depression.

The creation of a predictive system that correctly forecasts future changes of a stock price is crucial for investment management and algorithmic trading. The use of technical analysis for financial forecasting has been successfully employed by many researchers. Input window length is a time frame parameter required to be set when calculating many technical indicators. This study explores how the performance of the predictive system depends on a combination of a forecast horizon and an input window length for forecasting variable horizons. Technical indicators are used as input features for machine learning algorithms to forecast future directions of stock price movements. The dataset consists of ten years daily price time series for fifty stocks. The highest prediction performance is observed when the input window length is approximately equal to the forecast horizon. This novel pattern is studied using multiple performance metrics: prediction accuracy, winning rate, return per trade and Sharpe ratio.

A fundamental problem in cooperative HumanRobot Interaction is object handover. Existing works in this area assume the human can reliably grasp the object from the robot hand. However, in some situations the human can produce perturbing forces in the object that are not meant to end in a handover. These perturbations can result in the object being dropped or the robot hand being damaged. This paper addresses this problem and presents a mechanism for reliable robot to human object handover implemented in a Shadow Robot hand endowed with tactile sensing. Given a stable grasping configuration, using BioTAC sensors we are able to estimate the contact forces applied to the object, and provide a feedback signal to a joint effort controller to maintain grasp forces despite perturbations. Our system is able to identify between object pulling forces which should result in an object handover, and other disturbances. Experimental results show that the hand releases the object only when the object is pulled, validating the proposed algorithm.

One of the urgent challenges in the automated analysis and interpretation of electrical brain activity is the effective handling of uncertainties associated with the complexity and variability of brain dynamics, reflected in the nonstationary nature of brain signals such as electroencephalogram (EEG). This poses a severe problem for existing approaches to the classification task within brain–computer interface (BCI) systems. Recently emerged type-2 fuzzy logic (T2FL) methodology has shown a remarkable potential in dealing with uncertain information given limited insight into the nature of the data generating mechanism. The objective of this work is thus to examine the applicability of T2FL approach to the problem of EEG pattern recognition. In particular, the focus is two-fold: i) the design methodology for the interval T2FL system (IT2FLS) that can robustly deal with inter-session as well as within-session manifestations of nonstationary spectral EEG correlates of motor imagery (MI), and ii) the comprehensive examination of the proposed fuzzy classifier in both off-line and on-line EEG classification case studies. The on-line evaluation of the IT2FLS-controlled real-time neurofeedback over multiple recording sessions holds special importance for EEG-based BCI technology. In addition, a retrospective comparative analysis accounting for other popular BCI classifiers such as linear discriminant analysis (LDA), kernel Fisher discriminant (KFD) and support vector machines (SVMs) as well as a conventional type-1 FLS (T1FLS), simulated off-line on the recorded EEGs, has demonstrated the enhanced potential of the proposed IT2FLS approach to robustly handle uncertainty effects in BCI classification.

A major challenge in two-class brain-computer interface (BCI) systems is the low bandwidth of the communication channel, especially while communicating and controlling assistive devices, such as a smart wheelchair or a telepresence mobile robot, which requires multiple motion command options in the form of forward, left, right, backward, and start/stop. To address this, an adaptive user-centric graphical user interface referred to as the intelligent adaptive user interface (iAUI) based on an adaptive shared control mechanism is proposed. The iAUI offers multiple degrees-of-freedom control of a robotic device by providing a continuously updated prioritized list of all the options for selection to the BCI user, thereby improving the information transfer rate. Results have been verified with multiple participants controlling a simulated as well as physical pioneer robot.

A wash trade refers to the illegal activities of traders who utilize carefully designed limit orders to manually increase the trading volumes for creating a false impression of an active market. As one of the primary formats of market abuse, a wash trade can be extremely damaging to the proper functioning and integrity of capital markets. The existing work focuses on collusive clique detections based on certain assumptions of trading behaviors. Effective approaches for analyzing and detecting wash trade in a real-life market have yet to be developed. This paper analyzes and conceptualizes the basic structures of the trading collusion in a wash trade by using a directed graph of traders. A novel method is then proposed to detect the potential wash trade activities involved in a financial instrument by first recognizing the suspiciously matched orders and then further identifying the collusions among the traders who submit such orders. Both steps are formulated as a simplified form of the knapsack problem, which can be solved by dynamic programming approaches. The proposed approach is evaluated on seven stock data sets from the NASDAQ and the London Stock Exchange. The experimental results show that the proposed approach can effectively detect all primary wash trade scenarios across the selected data sets.

Tactile sensing has been used in robotics for object identification, grasping, and material recognition. Most material recognition approaches use vibration signals from a tactile exploration, typically above one second long, to identify the material. This work proposes a tactile multi-modal (vibration and thermal) material identification approach based on recursive Bayesian estimation. Through the frequency response of the vibration induced by the material and thermal features, like an estimate of the thermal power loss of the finger, we show that it is possible to identify materials in less than half a second. Moreover, a comparison between vibration only and multi-modal identification shows that both recognition time and classification errors are reduced by adding thermal information.

A context-aware cognitive system is a prime requirement for a sensor rich smart home environment. In this paper, we discuss the development and evaluation of a self-sustaining cognitive architecture for the RUBICON (Robotic UBIquitous COgnitive Network) system which builds its knowledge as per the environmental situations. The proposed cognitive architecture consists of a reasoning module, a decision module, and a supporting memory module. An online sliding-window based self-organising fuzzy neural network (SOFNN), which explores relationships between the event inputs and desired reasoning outputs, is developed for the reasoning module. We also propose a prediction model based on event information to support the reasoning module for continuous training in the absence of external training data. The decision module generates control goals for the robots according to the status outputs from the reasoning module. We develop a MySQL based database for the memory module which supports the overall system by storing processed information about the states of the environment and providing historical information for enhanced understanding. The architecture is trained and tested with environmentally realistic synthesized data to show its adaptation capabilities. The results demonstrate that the proposed system can learn activities and track them within a smart home environment. This initial implementation also highlights the potential of the architecture and will serve as a very important test-bed for future work. We envisage that the proposed combination of the prediction model and the reasoning module will eventually result in a general purpose, self-sustaining, self-organising cognitive architecture for different applications and thus the proposed architecture enters into the sphere of the biologically inspired cognitive architecture (BICA) challenge.

Based on the information processing functionalities of spiking neurons, hierarchical spiking neural networks are proposed to simulate visual attention. Using spiking neural networks inspired by the visual system, an image can be decomposed into multiple visual image components. Based on specific visual image components and image features, a visual attention system is proposed to extract attention areas according to top–down volition-controlled signals. The hierarchical spiking neural networks are constructed with a conductance-based integrate-and-fire neuron model and a set of specific receptive fields in different levels. The simulation algorithm and properties of the networks are detailed in this paper. Simulation results show that the attention system is able to perform visual attention of objects based on specific image components or features, and a demonstration shows how the attention system can detect a house in a visual image. Using the proposed saliency index, attention areas of interest can be extracted from spike rate maps of multiple visual pathways, such as ON/OFF colour pathways. According to this visual attention principle, the visual image processing system can quickly focus on specific areas while ignoring other areas.

Existing models for document summarization mostly use the similarity between sentences in the document to extract the most salient sentences. The documents as well as the sentences are indexed using traditional term indexing measures, which do not take the context into consideration. Therefore, the sentence similarity values remain independent of the context. In this paper, we propose a context sensitive document indexing model based on the Bernoulli model of randomness. The Bernoulli model of randomness has been used to find the probability of the cooccurrences of two terms in a large corpus. A new approach using the lexical association between terms to give a context sensitive weight to the document terms has been proposed. The resulting indexing weights are used to compute the sentence similarity matrix. The proposed sentence similarity measure has been used with the baseline graph-based ranking models for sentence extraction. Experiments have been conducted over the benchmark DUC data sets and it has been shown that the proposed Bernoulli-based sentence similarity model provides consistent improvements over the baseline IntraLink and UniformLink methods [1].

A user query for information retrieval (IR) applications may not contain the most appropriate terms (words) as actually intended by the user. This is usually referred to as the term mismatch problem and is a crucial research issue in IR. Using the notion of relevance, we provide a comprehensive theoretical analysis of a parametric query vector, which is assumed to represent the information needs of the user. A lexical association function has been derived analytically using the system relevance criteria. The derivation is further justified using an empirical evidence from the user relevance criteria. Such analytical derivation as presented in this paper provides a proper mathematical framework to the query expansion techniques, which have largely been heuristic in the existing literature. By using the generalized retrieval framework, the proposed query representation model is equally applicable to the vector space model (VSM), Okapi best matching 25 (Okapi BM25), and Language Model (LM). Experiments over various data sets from TREC show that the proposed query representation gives statistically significant improvements over the baseline Okapi BM25 and LM as well as other well-known global query expansion techniques. Empirical results along with the theoretical foundations of the query representation confirm that the proposed model extends the state of the art in global query expansion.

We present a least trimmed square (LTS) robust regression method to combine different runs/subjects for second/high level effective connectivity analysis. The basic idea of this method is to treat the extreme nonlinear model variability as outliers if they exceed a certain threshold. A bootstrap method for the LTS estimation is employed to detect model outliers. We compared the LTS robust method with a non-robust method using simulated and real datasets. The difference between LTS and the non-robust method for second level effective connectivity analysis is significant, suggesting the conventional non-robust method is easily affected by the model variability from the first level analysis. In addition, after these outliers are detected and excluded for the high level analysis, the model coefficients of the second level are combined within the framework of a mixed model. The variance of the mixed model is estimated using the Newton-Raphson (NR) type Levenberg-Marquardt algorithm. Three sets of real data are adopted to compare conventional methods which do not include random effects in the analysis with a mixed model for second level effective connectivity analysis. The results show that the conventional method is significantly different from the mixed model when greater model variability exists, suggesting there is a strong random effect, and the mixed model should be employed for the second level effective connectivity analysis.

Inflammatory bowel disease (IBD) is a group of inflammatory diseases of the human colon and small intestine. IBD symptoms are non-specific; diagnosis can be delayed because an invasive colonoscopy is required for confirmation. Delayed diagnosis is linked to poor growth in children. Imbalances in the human intestinal microbiome - the community of microorganisms that reside in the human gut - are thought to contribute to the development of IBD. Work done to date in classifying host health statuses from patterns in human microbiomes with supervised learning algorithms has focused on modelling what is present in the gut (i.e. a bacterial census) with the random forest algorithm. Metagenomic shotgun sequencing is required to understand what is occurring in the gut (i.e. gene functions) and is often cost prohibitive for hundreds of samples. However, gene functions can be predicted with the Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PiCRUSt) software package, which could represent a valuable source of new features. In this paper we investigate feature relevance across the feature set with the Boruta algorithm. We find that the majority of relevant features are from the predicted metagenome. Support vector machines (SVM) and multilayer perceptrons (MLP) are rarely used with microbiomic datasets but offer several theoretical advantages. To determine if the new features and alternative algorithms are appropriate, we experiment with a range of machine learning and computational intelligence algorithms. With the best performing algorithms we also implement a conditional multiple classifier system that can identify IBD presence, IBD subtype, and IBD activity from a non-invasive stool sample.

Tinnitus is the phantom perception of a sound heard in or around the head in the absence of an identifiable source affecting 10–15% worldwide. The majority of tinnitus sufferers have some form of hearing loss. The multiple pathologies that generate and sustain tinnitus in a diverse tinnitus population make it challenging to establish a homogeneous cohort for experimental studies. People with no hearing loss or previous experience of tinnitus also begin to perceive phantom sounds when situated in a sound proof room for five minutes or less. This is consistently observed across multiple studies. Studies that induced tinnitus through acoustic deprivation in healthy subjects provide a more controlled environment to observe tinnitus. Although experimental work shows what is happening it does not explain how the tinnitus related activity is generated. Computational modelling of tinnitus following hearing loss has shown that underlying mechanisms, such as adaptive gain, can generate hyperactivity in the regions of hearing loss. These models do not account for the generation of tinnitus in people with no hearing loss. In this work we model the development of tinnitus related activity in cases of no hearing loss and induced acoustic deprivation. The tinnitus related activity disappears once the model is returned to normal ambient noise.

Understanding how simple biological entities perform complex tasks (such as moving away from predators, reproduction and seeking for food) is one of the main goals of researchers across the world. The C. elegans nematode is one of the simplest and well characterised biological nervous systems (BNS) and the Si elegans EU FP7 project aims to develop a framework to mimic its functionality and better understand basic functions. This paper presents a small scale emulation environment composed of 8 neurons running on field programmable gate arrays (FPGAs), where each neuron FPGA is interconnected to all of the other neuron FPGAs via an innovative optical synaptic connectivity strategy. This paper evaluates the fidelity of synaptic communication for transmission of spikes using both wired and optical methods.