In this paper, we ensue on the development of a taxonomy aimed at categorizing distractions in the P300b domain. Explicitly, we investigate the effect that auditory distractions, distinctively that of ambient noise (AN), passive talking (PT), and active listening (AL) have on the signal of a visual P300 Speller in terms of accuracy, amplitude, latency, user preference, signal morphology, and overall signal quality. This work is part of a larger EEG based project and is based on the P300 speller BCI (oddball) paradigm and the xDAWN algorithm, with eight healthy subjects; while using a non-invasive Brain-Computer Interface based on low fidelity electroencephalographic (EEG) equipment. Our results show that the accuracy was best for the Lab (LC) at 100%, followed by AN at 92.5%, PT at 90% and last AL at 87.5%, which results were in identical order to the subjects’ preferences. In addition, the amplitude and latency did not show any statistical significance in all settings. This paper provides additional results that impart insight into the practicability of the aforementioned P300 speller methodology and low-cost equipment to be used in real-world applications.
Aside from their physical exterior/interior design, the energy demands of industrial buildings are strongly related to how they are used. It follows that the behaviour of the occupants contributes and is related to the energy consumption of a building. In a hospital, this could mean equipment usage, heating, water and so on. This implies that energy consumption in a specific area may be measured as a function of occupancy, making occupancy monitoring an important part of an organisation's energy management plan. This paper presents the design, implementation and testing of an Occupancy Monitoring Unit (OMU), based on thermal imaging technology, to provide occupancy data on individual selected wards/areas of Medway NHS Foundation Trust (MWNFT). The paper also presents successful tests performed to assess the functionality of the OMU in counting people and identifying the direction of motion. The results show the potential for the OMU to count individual people as well as groups.
In the modern era where the tourism industry continues to grow and contribute both socially and economically at global scale, a demanding need to boost the tourism sector through IT arises. This need along with recent, technology driven, behavior patterns where users, with the use of their smartphones capture a wide number of images daily has led to the proposal of a new concept namely “Moments of Interest” (MOIs). This novel concept is realized on an innovative mobile application that aims at offering personalized recommendations to tourists based on special “moments” residing in user captured images. In order to achieve this, image labeling through machine learning is utilized and a novel cloud-based “Memories Database” is created.
This paper presents an automated procedure for analyzing SeismoCardioGraphic (SCG) traces, i.e. chest vibrations induced by heart activity. Such signals are acquired by means of an inexpensive, compact accelerometer device, placed over the subject’s sternum and held in place by a light strap. The methodology for beat detection and annotation features a preliminary coarse beat detection phase, followed by a self-managed calibration, that is subsequently leveraged to carry out actual SCG annotation and beat localization. The performance of such process was measured and compared against the ECG gold standard, used only to provide ground-truth beat-to-beat intervals (i.e. R-peak landmarks). Results show high average values of sensitivity and precision (99.1% and 97.9%, respectively). The coefficient of determination R2 reaches a 0.984 value, which shows that the algorithm is able to temporally localize SCG complexes in a very consistent way, compared to the ECG gold standard; indeed, the variability between the ECG and SCG beat to beat measures is found to be very limited (σdiff ≈ 8.5 ms, i.e. less than one sample interval).
Soft errors are one of the significant design technology challenges at smaller technology nodes and especially in radiation enviro nments. This paper presents a particular class of approaches to provide reliability against radiation-induced soft errors. The paper provides a review of the lockstep mechanism across different levels of design abstraction: processor design, architectural level, and the software level. This work explores techniques providing modifications in the processor pipeline, techniques allied with FPGA dynamic reconfiguration strategies and different types of spatial redundancy.
In recent years, the significant increase in transmission control protocol and internet protocol (TCP/IP) network traffic has urged the importance of faster network systems. This study proposes a simultaneous TCP multi-streaming system to improve data transfer rate by integrating TCP streams over multiple IP domains. The system is implemented using a virtual network switch that processes the packet from multiple network interface cards (NICs); it also presents a possibility in software defined network, which attracts attention as a backbone of future networks as 5G networks. We utilized actual equipment for performing experiments and evaluations on a local network to measure the data transfer rate. Results confirmed an improvement in the communication rate.
Great advancements have been achieved in the field of robotics, however, main challenges remain, including building robots with an adaptive Theory of Mind (ToM). In the present paper, seven current robotic architectures for human-robot interactions were described as well as four main functional advantages of equipping robots with an adaptive ToM. The aim of the present paper was to determine in which way and how often ToM features are integrated in the architectures analyzed, and if they provide robots with the associated functional advantages. Our assessment shows that different methods are used to implement ToM features in robotic architectures. Furthermore, while a ToM for false-belief understanding and tracking is often built in social robotic architectures, a ToM for proactivity, active perception and learning is less common. Nonetheless, progresses towards better adaptive ToM features in robots are warranted to provide them with full access to the advantages of having a ToM resembling that of humans.
It is typical for a machine learning system to have numerous hyperparameters that affect its learning rate and prediction quality. Finding a good combination of the hyperparameters is, however, a challenging job. This is mainly because evaluation of each combination is extremely expensive computationally; indeed, training a machine learning system on real data with just a single combination of hyperparameters usually takes hours or even days. In this paper, we address this challenge by trying to predict the performance of the machine learning system with a given combination of hyperparameters without completing the expensive learning process. Instead, we terminate the training process at an early stage, collect the model performance data and use it to predict which of the combinations of hyperparameters is most promising. Our preliminary experiments show that such a prediction improves the performance of the commonly used random search approach.
The Generalised Travelling Salesman Problem (GTSP) is a well-known problem that, among other applications, arises in warehouse order picking, where each stock is distributed between several locations a typical approach in large modern warehouses. However, the instances commonly used in the literature have a completely different structure, and the methods are designed with those instances in mind. In this paper, we give a new pseudo-random instance generator that reflects the warehouse order picking and publish new benchmark testbeds. We also use the Conditional Markov Chain Search framework to automatically generate new GTSP metaheuristics trained specifically for warehouse order picking. Finally, we report the computational results of our metaheuristics to enable further competition between solvers.
Mobile-based gaming applications can motivate and facilitate in educating adolescents in health awareness and further can prevent obesity. The purpose of this paper was to observe and assess the initial acceptance of a prototype mobile game application in promoting a healthy Body Mass Index among adolescents. Based on the observations, the developed game was found to be interesting and has positive feedback. Positive user experiences were being expressed on the game's style and aesthetics, operability, learnability, autonomous learning, enjoyment, and social interaction offered. All of these added the excitements of the players and most of them were looking forward to continuing playing in the future. However, some challenges were reported, related to the player's confidence, challenges offered and winning opportunity. The positive findings and concerns discovered during the testing session are reported in this paper. Good health promotion via game-based learning using mobiles demonstrates the potential to offer enjoyment and feasible for adolescents although in our case, further usability testing with larger-scale of participants and a longer session is needed.
Energy Harvesting Systems seek to remove the batteries from electronic devices and replace them with devices that generate directly from the environment around them. This paper presents an intelligent algorithm to manage the charging and discharging of a supercapacitor for use in sensor nodes which has considerable efficiency advantages. The ultralow power microcontrollers used make intelligent decisions whether to sleep or wake according to previously received and stored energy. Using an adaptive strategy of this kind the amount of work can be precisely matched to the resources available to achieve maximum utilization allowing the satisfactory completion of a stated set of tasks.
General Video Game Level Generation provides a platform to develop level generators that work for general games within a certain domain. In this paper, we present our n-gram and constraint-based level generator. The generator generates levels based on the player logs. The aesthetics are handled by n-gram and playability is ensured by using precise constraints. The generated levels are also evaluated using relative performance profile and user study. The experimental results show that the generated levels are of adequate quality comparative to the levels generated by sample level generators.
functional Near-infrared Spectroscopy (fNIRS) is fast becoming an alternative optical neuroimaging modality to functional Magnetic Resonance Imaging (fMRI) owing to its portable, non-constrained environment. With the advent of fNIRS numerous studies, and key populations once considered indecipherable because of multitude operational challenges associated with fMRI such as restriction to stay in a supine position motionless surrounded by huge magnets, can now be read into. In this work, brain activity of participants, who are gamers of varying expertise levels, watch images or videos of a game, League of Legends, is recorded using fNIRS. The participant's fNIRS data is then analysed to distinguish between expertise levels of gamers using a support vector machine classifier with a radial basis function kernel. Our results demonstrate the adequacy of using fNIRS data to distinguish between the expertise levels of participants. In particular, the classification accuracy is optimum for novice vs. intermediate participants followed by novices vs. experts for all experiments. The least accurate classification results obtained are for intermediates vs. experts. An attempt is also made to read from different dimensions of hemoglobin to establish which biomarker best respresents the neural activity in the brain. Since the methods employed are independent of the study, we believe this work has strong implications for a professional's objective assessment which is paramount for those occupations especially associated with greater risks e.g. surgeons, pilots.
With the increasing usage of electronic emails, the ratio of spam is increasing day by day. Thus, spam emails have become a major threat that lowers the usage of electronic emails as a way for communication. There are several machine learning techniques that provide email spam filtering methods, such as Naive Bayes (NB), K-Nearest Neighbor (KNN), Support Vector Machine (SVM), Artificial Neural Network (ANN) and Decision tree (DT). This paper considers different machine learning techniques to filter spam emails, specifically Adaboost and Stochastic Gradient Descent (SGD). R tool was used for the pre-processing stage. Adaboost and SGD were implemented in Orange software for building the classifiers. Using Orange tool, the experimental results showed that the algorithms Adaboost and stochastic gradient descent (SGD) provided true positive value of 100 % and 98.1% respectively and false positive rates of 0.0% and 1.9% respectively. The good accuracy of these algorithms and the favorable results put them among the best choices of spam filtering methods.
Autonomous/unmanned driving has the capability to provide numerous benefits such as better traffic management, increased safety, reduced emission, and enhanced transportation network. Once autonomous ground vehicles (AGVs) are deployed, they will have to interact with other such vehicles. Interaction between multiple AGVs is an important area of research where analysis on the performance of algorithms/control schemes of AGVs is carried out. Performing real-world experiments with teams of autonomous vehicles is a challenging task due to cost and complexity. On the other hand, a simulation can emulate reality and provide an inexpensive and less time-consuming development process compared to the real world robots' testing. Therefore, a simulation tool is developed for multi-robot navigation. This simulator is based on open-source Robot Operating Systems (ROS) and natively supported robotics simulator Gazebo.
Planners can be used to manage game content such as interactive narratives and agent behaviour. However, planners require a semantic representation of games objects and events to reason about them. This paper describes a framework to create a symbolic representation of the semantic content of a real-time game for use with a story generation planner. Asynchronous events from the game are marshalled by an event handler into messages that are passed to a domain-specific message-handler. This message handler interprets and converts these messages to first-order predicates for use in a domain-independent planner. This approach aims to limit the amount of semantic data stored within the game that is unnecessary for gameplay logic.
Deep Convolutional Generative Adversarial Networks (DCGANs) are a machine learning approach that can learn to mimic any distribution of data. DCGANs consist of a generator and discriminator where generator generates new content and discriminator finds whether the generated content is real or fake. Procedural Content Generation (PCG) for level generation could potentially benefit from such models, particularly for a game where there is an existing level to emulate. In this paper, DCGANs are trained on existing levels generated through a random generator. Three different games (Freeway, Zelda, and Colourescape) are selected from the General Video Game Artificial Intelligence (GVG-AI) framework for level generation. The proposed approach successfully generates various levels which mimic the training levels. Generated levels are also evaluated using agent-based testing to ensure their play-ability.
This paper investigates the dynamics of information spread across social network services (SNSs) such as Twitter using the susceptible-infected-recovered (SIR) model. In most practical applications, an exact analytic solution is not available for the SIR model, so previous studies have largely been based on the assumption that the probability of a node having the target information is independent of whether or not its neighbors have that information. In contrast, we herein propose a different approach based on an assumption called the “strong correlation assumption”, in which the probability of a node having the target information is strongly correlated with whether its neighboring nodes have that information. We then analyze information spread for the SIR model and show that the use of the strong correlation assumption makes it possible to analyze the spread of information with far greater accuracy than the node independence assumption.
Cognitive load theory (CLT) is concerned with the design of interfaces that allow users maximize their working memory when problem solving. This includes eliminating any subjective mental load that may be imposed by the instruction interface. Augmented reality (AR) is fast gaining application in areas of education, military, business and medicine with new AR applications developed daily. Despite its wide application little has been done in the area of design guidelines for AR user interface and the evaluation of its effect. This is necessary so as to ensure that users are not burdened by the format and amount of information presented in the augmented view. This work - in - progress aims to propose a framework for designing AR user interfaces that imposes reduced subjective mental workload on users. Particular attention is given to highly cognitive tasks such as simultaneous interpretation. This research work will concentrate on presenting the novel guidelines for the user interface design.
In this paper, we propose a social negotiation system in which agents can communicate and interact with each other socially throughout a Sheriff of Nottingham game. We address issues with the number of options available while negotiating, particularly when bluffing is involved. Experiments are proposed that would allow us to validate how closely this framework mirrors real social interaction in the game, and the possibility of generalising multi-agent negotiation beyond this framework is raised.
