The Inverse Epsilon Distribution as an Alternative to Inverse Exponential Distribution with a Survival Times Data Example

This paper is devoted to a new flexible two-parameter lower-truncated distribution, which is based on the inversion of the so-called epsilon distribution. It is called the inverse epsilon distribution. In some senses, it can be viewed as an alternative to the inverse exponential distribution, which has many applications in reliability theory and biology. Diverse properties of the new lower-truncated distribution are derived including relations with existing distributions, hazard and reliability functions, survival and reverse hazard rate functions, stochastic ordering, quantile function with related skewness and kurtosis measures, and moments. A demonstrative survival times data example is used to show the applicability of the new model.

Domain Semirings United

Domain operations on semirings have been axiomatised in two different ways: by a map from an additively idempotent semiring into a boolean subalgebra of the semiring bounded by the additive and multiplicative unit of the semiring, or by an endofunction on a semiring that induces a distributive lattice bounded by the two units as its image. This note presents classes of semirings where these approaches coincide.

Computing Different Realizations of Linear Dynamical Systems with Embedding Eigenvalue Assignment

In this paper we investigate realizability of discrete time linear dynamical systems (LDSs) in fixed state space dimension. We examine whether there exist different Θ = (A,B,C,D) state space realizations of a given Markov parameter sequence Y with fixed B, C and D state space realization matrices. Full observation is assumed in terms of the invertibility of output mapping matrix C. We prove that the set of feasible state transition matrices associated to a Markov parameter sequence Y is convex, provided that the state space realization matrices B, C and D are known and fixed. Under the same conditions we also show that the set of feasible Metzler-type state transition matrices forms a convex subset. Regarding the set of Metzler-type state transition matrices we prove the existence of a structurally unique realization having maximal number of non-zero off-diagonal entries. Using an eigenvalue assignment procedure we propose linear programming based algorithms capable of computing different state space realizations. By using the convexity of the feasible set of Metzler-type state transition matrices and results from the theory of non-negative polynomial systems, we provide algorithms to determine structurally different realization. Computational examples are provided to illustrate structural non-uniqueness of network-based LDSs.

Traquest Model

Atomicity, consistency, isolation and durability are essential properties of many distributed systems. They are often abbreviated as the ACID properties. Ensuring ACID comes with a price: it requires extra computing and network capacity to ensure that the atomic operations are done perfectly, or they are rolled back. When we have higher requirements on performance, we need to give up the ACID properties entirely or settle for eventual consistency. Since the ambiguity of the order of the events, such algorithms can get very complicated since they have to be prepared for any possible contingencies. Traquest model is an attempt for creating a general concurrency model that can bring the ACID properties without sacrificing a too significant amount of performance.

Access Control of EHR Records in a Heterogeneous Cloud Infrastructure

Since the advent of smartphones, IoT and cloud computing, we have seen an industry-wide requirement to integrate different healthcare applications with each other and with the cloud, connecting multiple institutions or even countries. But despite these trends, the domain of access control and security of sensitive healthcare data still raises a serious challenge for multiple developers and lacks the necessary definitions to create a general security framework addressing these issues. Taking into account newer, more special cases, such as the popular heterogeneous infrastructures with a combination of public and private clouds, fog computing, Internet of Things, the area becomes more and more complicated. In this paper we will introduce a categorization of these required policies, describe an infrastructure as a possible solution to these security challenges, and finally evaluate it with a set of policies based on real-world requirements.

Energy-Efficient Routing in Wireless Sensor Networks

Efficient data collection is the core concept of implementing Industry4.0 on IoT platforms. This requires energy aware communication protocols for Wireless Sensor Networks (WSNs) where different functions, like sensing and processing on the IoT nodes must be supported by local battery power. Thus, energy aware network protocols, such as routing, became one of fundamental challenges in IoT data collection schemes.In our research, we have developed novel routing algorithms which guarantee minimum energy consumption data transfer which is achieved subject to pre-defined reliability constraints. We assume that data is transmitted in the form of packets and the routing algorithm identifies the paths over which the packets can reach the Base Station (BS) with minimum transmission energy, while the probability of successful packet transmission still exceeds a pre-defined reliability parameter. In this way, the longevity and the information throughput of the network is maximized and the low energy transmissions will considerably extend the lifetime of the IoT nodes. In this paper we propose a solution that maximizes the lifetime of the nodes.

Using the Fisher Vector Approach for Cold Identification

In this paper, we present a computational paralinguistic method for assessing whether a person has an upper respiratory tract infection (i.e. cold) using their speech. Having a system that can accurately assess a cold can be helpful for predicting its propagation. For this purpose, we utilize Mel-frequency Cepstral Coefficients (MFCC) as audio-signal representations, extracted from the utterances, which allowed us to fit a generative Gaussian Mixture Model (GMM) that serves to produce an encoding based on the Fisher Vector (FV) approach. Here, we use the URTIC dataset provided by the organizers of the ComParE Challenge 2017 of the Interspeech Conference. The classification is done by a linear kernel Support Vector Machines (SVM); owing to the high imbalance of classes on the training dataset, we opt for undersampling the majority class, that is, to reduce the number of samples to those of the minority class. We find that applying Power Normalization (PN) and Principal Component Analysis (PCA) on the Fisher vector features is an effective strategy for the classification performance. We get better performance than that of the Bag-of-Audio-Words approach reported in the paper of the challenge.

Toolset for Supporting the Research of Lattice Based Number Expansions

The world of generalized number systems contains many challenging areas. Computer experiments often support the theoretical research. In this paper we introduce a toolset that helps to analyze some properties of lattice based number expansions. The toolset is able to (1) analyze the expansions, (2) decide the number system property, (3) classify and visualize the periodic points. The toolset is implemented in Python, published alongside with a database that stores plenty of special expansions, and is able to store the custom properties like signature, operator eigenvalues, etc. Researchers can connect to the server and request/upload data, or perform experiments on them. We present an introductory usage of the toolset and detail some results that has been observed by the toolset. The toolset can be downloaded from http://numsys.info domain.

Zero Initialized Active Learning with Spectral Clustering using Hungarian Method

Supervised machine learning tasks often require a large number of labeled training data to set up a model, and then prediction - for example the classification - is carried out based on this model. Nowadays tremendous amount of data is available on the web or in data warehouses, although only a portion of those data is annotated and the labeling process can be tedious, expensive and time consuming. Active learning tries to overcome this problem by reducing the labeling cost through allowing the learning system to iteratively select the data from which it learns. In special case of active learning, the process starts from zero initialized scenario, where the labeled training dataset is empty, and therefore only unsupervised methods can be performed. In this paper a novel query strategy framework is presented for this problem, called Clustering Based Balanced Sampling Framework (CBBSF), which is not only select the initial labeled training dataset, but uniformly selects the items among the categories to get a balanced labeled training dataset. The framework includes an assignment technique to implicitly determine the class membership probabilities. Assignment solution is updated during CBBSF iterations, hence it simulates supervised machine learning more accurately as the process progresses. The proposed Spectral Clustering Based Sampling (SCBS) query startegy realizes the CBBSF framework, and therefore it is applicable in the special zero initialized situation. This selection approach uses ClusterGAN (Clustering using Generative Adversarial Networks) integrated in the spectral clustering algorithm and then it selects an unlabeled instance depending on the class membership probabilities. Global and local versions of SCBS were developed, furthermore, most confident and minimal entropy measures were calculated, thus four different SCBS variants were examined in total. Experimental evaluation was conducted on the MNIST dataset, and the results showed that SCBS outperforms the state-of-the-art zero initialized active learning query strategies.

Speech De-identification with Deep Neural Networks

Cloud-based speech services are powerful practical tools but the privacy of the speakers raises important legal concerns when exposed to the Internet. We propose a deep neural network solution that removes personal characteristics from human speech by converting it to the voice of a Text-to-Speech (TTS) system before sending the utterance to the cloud. The network learns to transcode sequences of vocoder parameters, delta and delta-delta features of human speech to those of the TTS engine. We evaluated several TTS systems, vocoders and audio alignment techniques. We measured the performance of our method by (i) comparing the result of speech recognition on the de-identified utterances with the original texts, (ii) computing the Mel-Cepstral Distortion of the aligned TTS and the transcoded sequences, and (iii) questioning human participants in A-not-B, 2AFC and 6AFC tasks. Our approach achieves the level required by diverse applications.