In this article we examine the methods for detecting and predicting aging related process sensor failures by analyzing the noise of the sensor output signal. The study uses data from non-differential and differential pressure transmitters used in the pressure and water level measurements of the reactor pressure vessels of units 1 and 2 of the Olkiluoto nuclear power plant in Finland. The article contains a review of the current methods for detection of sensor failures. Additionally, we present a new method for detecting changes in the sensor output signal. The method creates fingerprints of the power spectra of the sensors by using Principal Component Analysis (PCA). The changes in these fingerprints together with the measurements of the redundant sensors can be used to detect indications of some of the impending sensor failures. In the experimental study we are able to produce stable fingerprints for both the non-differential and differential pressure transmitters. Also, a potential failure in one of the differential pressure transmitters in Olkiluoto unit 2 is detected by inspecting the fingerprints and analyzing the spectral changes of the transmitter output signal.

Deep convolutional neural networks (CNNs) have been immensely successful in many high-level computer vision tasks given large labelled datasets. However, for video semantic object segmentation, a domain where labels are scarce, effectively exploiting the representation power of CNN with limited training data remains a challenge. Simply borrowing the existing pre-trained CNN image recognition model for video segmentation task can severely hurt performance. We propose a semi-supervised approach to adapting CNN image recognition model trained from labelled image data to the target domain exploiting both semantic evidence learned from CNN, and the intrinsic structures of video data. By explicitly modelling and compensating for the domain shift from the source domain to the target domain, this proposed approach underpins a robust semantic object segmentation method against the changes in appearance, shape and occlusion in natural videos. We present extensive experiments on challenging datasets that demonstrate the superior performance of our approach compared with the state-ofthe- art methods.

The provision of a particle and power exhaust solution which is compatible with first-wall components and edge-plasma conditions is a key area of present-day fusion research and mandatory for a successful operation of ITER and DEMO. The work package plasma-facing components (WP PFC) within the European fusion programme complements with laboratory experiments, i.e. in linear plasma devices, electron and ion beam loading facilities, the studies performed in toroidally confined magnetic devices, such as JET, ASDEX Upgrade, WEST etc. The connection of both groups is done via common physics and engineering studies, including the qualification and specification of plasma-facing components, and by modelling codes that simulate edge-plasma conditions and the plasma–material interaction as well as the study of fundamental processes. WP PFC addresses these critical points in order to ensure reliable and efficient use of conventional, solid PFCs in ITER (Be and W) and DEMO (W and steel) with respect to heat-load capabilities (transient and steady-state heat and particle loads), lifetime estimates (erosion, material mixing and surface morphology), and safety aspects (fuel retention, fuel removal, material migration and dust formation) particularly for quasi-steady-state conditions. Alternative scenarios and concepts (liquid Sn or Li as PFCs) for DEMO are developed and tested in the event that the conventional solution turns out to not be functional. Here, we present an overview of the activities with an emphasis on a few key results: (i) the observed synergistic effects in particle and heat loading of ITER-grade W with the available set of exposition devices on material properties such as roughness, ductility and microstructure; (ii) the progress in understanding of fuel retention, diffusion and outgassing in different W-based materials, including the impact of damage and impurities like N; and (iii), the preferential sputtering of Fe in EUROFER steel providing an in situ W surface and a potential first-wall solution for DEMO.

According to a recent F-secure report, 97% of mobile malware is designed for the Android platform which has a growing number of consumers. In order to protect consumers from downloading malicious applications, there should be an effective system of malware classification, that can detect previously unseen viruses. In this paper, we present a scalable and highly accurate method for malware classification based on features extracted from Android application package (APK) files. We explored several techniques for tackling independence assumptions in Naive Bayes and proposed Normalized Naive Bayes that resulted in an improved class separation and higher accuracy. We conducted a set of experiments on an up-todate large dataset of APKs provided by F-Secure and achieved 0.1% false malicious rate with overall accuracy of 91%. The results indicate practical applicability of the method, high performance and ability to achieve low false malicious rates. Keywords-Malware Classification; Naive Bayes; Security in Android;

Bidirectional recurrent neural networks (RNN) are trained to predict both in the positive and negative time directions simultaneously. They have not been used commonly in unsupervised tasks, because a probabilistic interpretation of the model has been difficult. Recently, two different frameworks, GSN and NADE, provide a connection between reconstruction and probabilistic modeling, which makes the interpretation possible. As far as we know, neither GSN or NADE have been studied in the context of time series before. As an example of an unsupervised task, we study the problem of filling in gaps in high-dimensional time series with complex dynamics. Although unidirectional RNNs have recently been trained successfully to model such time series, inference in the negative time direction is non-trivial. We propose two probabilistic interpretations of bidirectional RNNs that can be used to reconstruct missing gaps efficiently. Our experiments on text data show that both proposed methods are much more accurate than unidirectional reconstructions, although a bit less accurate than a computationally complex bidirectional Bayesian inference on the unidirectional RNN. We also provide results on music data for which the Bayesian inference is computationally infeasible, demonstrating the scalability of the proposed methods.

Deep neural networks with several layers have recently become a highly successful and popular research topic in machine learning due to their excellent performance in many benchmark problems and applications. A key idea in deep learning is to learn not only the nonlinear mapping between the inputs and outputs but also the underlying structure of the data (input) vectors. In this chapter, we first consider problems with training deep networks using backpropagation-type algorithms. After this, we consider various structures used in deep learning, including restricted Boltzmann machines, deep belief networks, deep Boltzmann machines, and nonlinear autoencoders. In the latter part of this chapter, we discuss in more detail the recently developed neural autoregressive distribution estimator and its variants.

We consider an extension of ICA and BSS for separating mutually dependent and independent components from two related data sets. We propose a new method which first uses canonical correlation analysis for detecting subspaces of independent and dependent components. Different ICA and BSS methods can after this be used for final separation of these components. Our method has a sound theoretical basis, and it is straightforward to implement and computationally not demanding. Experimental results on synthetic and real-world fMRI data sets demonstrate its good performance.

Variational methods for approximate inference in machine learning often adapt a parametric probability distribution to optimize a given objective function. This view is especially useful when applying variational Bayes (VB) to models outside the conjugate-exponential family. For them, variational Bayesian expectation maximization (VB EM) algorithms are not easily available, and gradient-based methods are often used as alternatives.

Independent component analysis (ICA) and blind source separation (BSS) are usually applied to a single data set. Both these techniques are nowadays well understood, and several good methods based on somewhat varying assumptions on the data are available. In this paper, we consider an extension of ICA and BSS for separating mutually dependent and independent components from two different but related data sets. This problem is important in practice, because such data sets are common in real-world applications. We propose a new method which first uses canonical correlation analysis (CCA) for detecting subspaces of independent and dependent components. Standard ICA and BSS methods can after this be used for final separation of these components. The proposed method performs excellently for synthetic data sets for which the assumed data model holds exactly, and provides meaningful results for real-world robot grasping data. The method has a sound theoretical basis, and it is straightforward to implement and computationally not too demanding. Moreover, the proposed method has a very important by-product: its improves clearly the separation results provided by the FastICA and UniBSS methods that we have used in our experiments. Not only are the signal-to-noise ratios of the separated sources often clearly higher, but CCA preprocessing also helps FastICA to separate sources that it alone is not able to separate.

Variational Bayesian (VB) methods are typically only applied to models in the conjugate-exponential family using the variational Bayesian expectation maximisation (VB EM) algorithm or one of its variants. In this paper we present an efficient algorithm for applying VB to more general models. The method is based on specifying the functional form of the approximation, such as multivariate Gaussian. The parameters of the approximation are optimised using a conjugate gradient algorithm that utilises the Riemannian geometry of the space of the approximations. This leads to a very efficient algorithm for suitably structured approximations. It is shown empirically that the proposed method is comparable or superior in efficiency to the VB EM in a case where both are applicable. We also apply the algorithm to learning a nonlinear state-space model and a nonlinear factor analysis model for which the VB EM is not applicable. For these models, the proposed algorithm outperforms alternative gradient-based methods by a significant margin.

While variational Bayesian (VB) inference is typically done with the so called VB EM algorithm, there are models where it cannot be applied because either the E-step or the M-step cannot be solved analytically. In 2007, Honkela et al. introduced a recipe for a gradient-based algorithm for VB inference that does not have such a restriction. In this paper, we derive the algorithm in the case of the mixture of Gaussians model. For the first time, the algorithm is experimentally compared to VB EM and its variant with both artificial and real data. We conclude that the algorithms are approximately as fast depending on the problem.