Recent increases in computing power, coupled with rapid growth in the availability and quantity of data have rekindled our interest in the theory and applications of artificial intelligence (AI). ...However, for AI to be confidently rolled out by industries and governments, users want greater transparency through explainable AI (XAI) systems. The author introduces XAI concepts, and gives an overview of areas in need of further exploration-such as type-2 fuzzy logic systems-to ensure such systems can be fully understood and analyzed by the lay user.
This paper presents a self-adaptive autonomous online learning through a general type-2 fuzzy system (GT2 FS) for the motor imagery (MI) decoding of a brain-machine interface (BMI) and navigation of ...a bipedal humanoid robot in a real experiment, using electroencephalography (EEG) brain recordings only. GT2 FSs are applied to BMI for the first time in this study. We also account for several constraints commonly associated with BMI in real practice: 1) the maximum number of EEG channels is limited and fixed; 2) no possibility of performing repeated user training sessions; and 3) desirable use of unsupervised and low-complexity feature extraction methods. The novel online learning method presented in this paper consists of a self-adaptive GT2 FS that can autonomously self-adapt both its parameters and structure via creation, fusion, and scaling of the fuzzy system rules in an online BMI experiment with a real robot. The structure identification is based on an online GT2 Gath-Geva algorithm where every MI decoding class can be represented by multiple fuzzy rules (models), which are learnt in a continous (trial-by-trial) non-iterative basis. The effectiveness of the proposed method is demonstrated in a detailed BMI experiment, in which 15 untrained users were able to accurately interface with a humanoid robot, in a single session, using signals from six EEG electrodes only.
Enhancers are non-coding regions of the genome that control the activity of target genes. Recent efforts to identify active enhancers experimentally and in silico have proven effective. While these ...tools can predict the locations of enhancers with a high degree of accuracy, the mechanisms underpinning the activity of enhancers are often unclear.
Using machine learning (ML) and a rule-based explainable artificial intelligence (XAI) model, we demonstrate that we can predict the location of known enhancers in Drosophila with a high degree of accuracy. Most importantly, we use the rules of the XAI model to provide insight into the underlying combinatorial histone modifications code of enhancers. In addition, we identified a large set of putative enhancers that display the same epigenetic signature as enhancers identified experimentally. These putative enhancers are enriched in nascent transcription, divergent transcription and have 3D contacts with promoters of transcribed genes. However, they display only intermediary enrichment of mediator and cohesin complexes compared to previously characterised active enhancers. We also found that 10-15% of the predicted enhancers display similar characteristics to super enhancers observed in other species.
Here, we applied an explainable AI model to predict enhancers with high accuracy. Most importantly, we identified that different combinations of epigenetic marks characterise different groups of enhancers. Finally, we discovered a large set of putative enhancers which display similar characteristics with previously characterised active enhancers.
Artificial intelligence (AI) is revolutionizing many real-world applications in various domains. In the field of genomics, multiple traditional machine-learning approaches have been used to ...understand the dynamics of genetic data. These approaches provided acceptable predictions; however, these approaches are based on opaque-box AI algorithms which are not able to provide the needed transparency to the community. Recently, the field of explainable artificial intelligence has emerged to overcome the interpretation problem of opaque box models by aiming to provide complete transparency of the model and its prediction to the users especially in sensitive areas such as healthcare, finance, or security. This paper highlights the need for eXplainable Artificial Intelligence (XAI) in the field of genomics and how the understanding of genomic regions, specifically the non-coding regulatory region of genomes (i.e., enhancers), can help uncover underlying molecular principles of disease states, in particular cancer in humans.
In the last decades, non-invasive and portable neuroimaging techniques, such as functional near infrared spectroscopy (fNIRS), have allowed researchers to study the mechanisms underlying the ...functional cognitive development of the human brain, thus furthering the potential of Developmental Cognitive Neuroscience (DCN). However, the traditional paradigms used for the analysis of infant fNIRS data are still quite limited. Here, we introduce a multivariate pattern analysis for fNIRS data, xMVPA, that is powered by eXplainable Artificial Intelligence (XAI). The proposed approach is exemplified in a study that investigates visual and auditory processing in six-month-old infants. xMVPA not only identified patterns of cortical interactions, which confirmed the existent literature; in the form of conceptual linguistic representations, it also provided evidence for brain networks engaged in the processing of visual and auditory stimuli that were previously overlooked by other methods, while demonstrating similar statistical performance.
Higher order fuzzy logic systems (FLSs), such as interval type-2 FLSs, have been shown to be very well suited to deal with the high levels of uncertainties present in the majority of real-world ...applications. General type-2 FLSs are expected to further extend this capability. However, the immense computational complexities associated with general type-2 FLSs have, until recently, prevented their application to real-world control problems. This paper aims to address this problem by the introduction of a complete representation framework, which is referred to as zSlices-based general type-2 fuzzy systems. The proposed approach will lead to a significant reduction in both the complexity and the computational requirements for general type-2 FLSs, while it offers the capability to represent complex general type-2 fuzzy sets. As a proof-of-concept application, we have implemented a zSlices-based general type-2 FLS for a two-wheeled mobile robot, which operates in a real-world outdoor environment. We have evaluated the computational performance of the zSlices-based general type-2 FLS, which is suitable for multiprocessor execution. Finally, we have compared the performance of the zSlices-based general type-2 FLS against type-1 and interval type-2 FLSs, and a series of results is presented which is related to the different levels of uncertainty handled by the different types of FLSs.
The interval type-2 fuzzy Proportional-Integral (PI) controller (IT2-FPI) might be able to handle high levels of uncertainties to produce a satisfactory control performance, which could be ...potentially due to the robust performance as a result of the smoother control surface around the steady state. However, the transient state and disturbance rejection performance of the IT2-FPI may degrade in comparison with the type-1 fuzzy PI (T1-FPI) counterpart. This drawback can be resolved via general type-2 fuzzy PI controllers which can provide a tradeoff between the robust control performance of the IT2-FPI and the acceptable transient and disturbance rejection performance of the type-1 PI controllers. In this paper, we will present a zSlices-based general type-2 fuzzy PI controller (zT2-FPI), where the secondary membership functions (SMFs) of the antecedent general type-2 fuzzy sets are adjusted in an online manner. We will examine the effect of the SMF on the closed-system control performance to investigate their induced performance improvements. This paper will focus on the case followed in conventional or self-tuning fuzzy controller design strategies, where the aim is to decrease the integral action sufficiently around the steady state to have robust system performance against noises and parameter variations. The zSlices approach will give the opportunity to construct the zT2-FPI controller as a collection of IT2-FPI and T1-FPI controllers. We will present a new way to design a zT2-FPI controller based on a single tuning parameter where the features of T1-FPI (speed) and IT2-FPI (robustness) are combined without increasing the computational complexity much when compared with the IT2-FPI structure. This will allow the proposed zT2-FPI controller to achieve the desired transient state response and provide an efficient disturbance rejection and robust control performance. We will present several simulation studies on benchmark systems, in addition to real-world experiments that were performed using the PIONEER 3-DX mobile robot that will act as a platform to evaluate the proposed systems. The results will show that the control performance of the self-tuning zT2-FPI control structure enhances both the transient state and disturbance rejection performances when compared with the type-1 and IT2-FPI counterparts. In addition, the self-tuning zT2-FPI is more robust to disturbances, noise, and uncertainties when compared with the type-1 and interval type-2 fuzzy counterparts.
Fuzzy logic control is a recognized approach for handling the faced uncertainties within control applications. However, type-1 fuzzy controllers using crisp type-1 fuzzy sets might not be able to ...fully handle the high levels of uncertainties and nonlinear dynamics associated with real world control applications. On the other hand, interval type-2 fuzzy controllers using Interval Type-2 Fuzzy Sets (IT2-FSs) might be able to handle such uncertainties to produce a better control performance. However, the systematic design of interval type-2 fuzzy controllers is still a challenging problem due to the difficulty in determining the parameters of the IT2-FSs. In this paper, we will present the novel application of Big Bang–Big Crunch optimization (BB–BC) approach to optimize the antecedent membership parameters of Interval Type-2 Fuzzy PID (IT2-FPID) controllers in a cascade control structure. Since the IT2-FPID control structure involves more design parameters compared to its type-1 counterpart, it is beneficial to employ the BB–BC method which has a low computational cost and a high convergence speed. The presented BB–BC based optimized IT2-FPID cascade structure will be compared with its Type-1 Fuzzy PID (T1-FPID) and conventional PID controller counterparts which were also optimized with the BB–BC optimization. In addition, the proposed IT2-FPID structure will be compared with a self-tuning T1-FPID control structure. We will then present the novel application of the cascade control architecture to solve the path tracking control problem of mobile robots which inherits large amounts of uncertainties caused by the internal dynamics and/or feedback sensors of the controlled system. Several experiments were performed in simulation and in real world using the PIONEER 3-DX mobile robot which will act as a platform to evaluate the proposed control systems in this paper. The results illustrated that the IT2-FPID structure enhanced significantly the control performance even in the presence of uncertainties and disturbances when compared to the PID, T1-FPID and self-tuning T1-FPID structures. Moreover, it has been shown that the reason for the superior control performance of the IT2-FPID under high levels of uncertainty and noise is not merely for its use of extra parameters, but rather its different way of dealing with the uncertainties and noise present in real world environments by comparing with a self-tuning T1-FPID structure.
In this paper, we will present a N-Non-Intersecting-Routing (NNIR) algorithm which is used to reduce the cost of resilient routing in telecommunications problems. Resilient Routing is the connections ...between two locations in a graph through the use of N completely independent routes. Resilient Routing is applicable in a wide variety of domains including telecommunications, logistics and embedded systems design. The proposed NNIR algorithm increase the cost of the primary route by taking a less optimal route, thus freeing a more optimal route for the resilient routes, in turn reducing the total cost of both routes. This is achieved through the use of a Genetic Algorithm, Dijkstra’s Algorithm and the repair operator. The proposed NNIR shows an average improvement of 34.2% when compared to Dijkstra’s Algorithm (one of the most widely used algorithm routing). Similarly, there is an average improvement of 34.2% when compared to A* (another popular shortest path algorithm). Additionally, there is an average improvement of 26.9% when compared to Simulated Annealing (a popular evolutionary technique used within routing problems). In this paper we show how NNIR performs within two different routing domains (telecommunications routing and road routing), and compares it against three other routing techniques to solve the resilient routing problem.
Our living environments are being gradually occupied with an abundant number of digital objects that have networking and computing capabilities. After these devices are plugged into a network, they ...initially advertise their presence and capabilities in the form of services so that they can be discovered and, if desired, exploited by the user or other networked devices. With the increasing number of these devices attached to the network, the complexity to configure and control them increases, which may lead to major processing and communication overhead. Hence, the devices are no longer expected to just act as primitive stand-alone appliances that only provide the facilities and services to the user they are designed for, but also offer complex services that emerge from unique combinations of devices. This creates the necessity for these devices to be equipped with some sort of intelligence and self-awareness to enable them to be self-configuring and self-programming. However, with this "smart evolution", the cognitive load to configure and control such spaces becomes immense. One way to relieve this load is by employing artificial intelligence (AI) techniques to create an intelligent "presence" where the system will be able to recognize the users and autonomously program the environment to be energy efficient and responsive to the user's needs and behaviours. These AI mechanisms should be embedded in the user's environments and should operate in a non-intrusive manner. This paper will show how computational intelligence (CI), which is an emerging domain of AI, could be employed and embedded in our living spaces to help such environments to be more energy efficient, intelligent, adaptive and convenient to the users.