Local correlation models that meet the requirements of a theoretical model chemistry are discussed. Two types of models are considered. The first class uses a valence active space that associates one correlating orbital for each occupied valence orbital. In these models the fundamental quantity is the electron pair; even the simplest local approximation (perfect pairing) exactly treats an isolated electron pair in this active space. The second class of models uses no active space. In this case we argue that the most appropriate fundamental quantity is the atom; even the simplest local approximation (atoms-in-molecules) exactly treats an isolated atom.

With the exponential growth in Internet-of-Things (IoT) devices, security and privacy issues have emerged as critical challenges that can potentially compromise their successful deployment in many data-sensitive applications. Hence, there is a pressing need to address these challenges, given that IoT systems suffer from different limitations, and IoT devices are constrained in terms of energy and computational power, which renders them extremely vulnerable to attacks. Traditional cryptographic algorithms use a static structure that requires several rounds of computations, which leads to significant overhead in terms of execution time and computational resources. Moreover, the problem is compounded when dealing with multimedia contents, since the associated algorithms have stringent QoS requirements. In this paper, we propose a lightweight cipher algorithm based on a dynamic structure with a single round that consists of simple operations, and that targets multimedia IoT. In this algorithm, a dynamic key is generated and then used to build two robust substitution tables, a dynamic permutation table, and two pseudo-random matrices. This dynamic cipher structure minimizes the number of rounds to a single one, while maintaining a high level of randomness and security. Moreover, the proposed cipher scheme is flexible as the dimensions of the input matrix can be selected to match the devices’ memory capacity. Extensive security tests demonstrated the robustness of the cipher against various kinds of attacks. The speed, simplicity and high-security level, in addition to low error propagation, make of this approach a good encryption candidate for multimedia IoT devices.

Matrix factorization, when the matrix has missing values, has become one of the leading techniques for recommender systems. To handle web-scale datasets with millions of users and billions of ratings, scalability becomes an important issue. Alternating least squares (ALS) and stochastic gradient descent (SGD) are two popular approaches to compute matrix factorization, and there has been a recent flurry of activity to parallelize these algorithms. However, due to the cubic time complexity in the target rank, ALS is not scalable to large-scale datasets. On the other hand, SGD conducts efficient updates but usually suffers from slow convergence that is sensitive to the parameters. Coordinate descent, a classical optimization approach, has been used for many other large-scale problems, but its application to matrix factorization for recommender systems has not been thoroughly explored. In this paper, we show that coordinate descent-based methods have a more efficient update rule compared to ALS and have faster and more stable convergence than SGD. We study different update sequences and propose the CCD++ algorithm, which updates rank-one factors one by one. In addition, CCD++ can be easily parallelized on both multi-core and distributed systems. We empirically show that CCD++ is much faster than ALS and SGD in both settings. As an example, with a synthetic dataset containing 14.6 billion ratings, on a distributed memory cluster with 64 processors, to deliver the desired test RMSE, CCD++ is 49 times faster than SGD and 20 times faster than ALS. When the number of processors is increased to 256, CCD++ takes only 16 s and is still 40 times faster than SGD and 20 times faster than ALS.

Two algorithms for the solution of a large sparse linear system of equations are proposed. The first is a modification of Lanczos' method and the second is based on one of Brezinski's methods. Both the latter methods are iterative and they can break down. In practical situations, serious numerical error is far more likely to occur because an ill-conditioned pair of polynomials is (implicitly) used in the calculation rather than complete breakdown arising because a large square block of exactly defective polynomials is encountered. The algorithms proposed use a method based on selecting well-conditioned pairs of neighbouring polynomials (in the associated Padé table), and the method is equivalent to going round the blocks instead of going across them, as is done in the well-known look-ahead methods.

Engineering is the use of science and technology to build useful artifacts. Those who design computer systems are clearly acting as engineers. However, there are deep differences between the way that computer systems are designed and the way that engineers in other areas work. Mechanical, electrical, and civil engineers make extensive use of mathematics to provide precise descriptions of their products. In contrast, computer systems are usually described, quite inaccurately, using anthropomorphic analogies and intuitive language. If an engineer produces a system from smaller components, he realises the necessity of precise specification of each of the components. Computer systems engineers, particularly programmers rarely write such specifications. They usually rely on a “cut and try” approach in which substantial redesign must be done after “integration” is begun.

Adaptive algebraic level-set segmentation algorithm of financial time series is presented in this paper. The proposed algorithm is based on the algebraic one step-forward predictor with internal smoothing, which is used to identify a near optimal algebraic model. Particle swarm optimization algorithm is exploited for the detection of a base algebraic fragment of the time series. A combinatorial algorithm is used to detect intervals where predictions are lower than a predefined level. Moreover, the combinatorial algorithm does assess the simplicity of the identified near optimal algebraic model. Automatic adaptive identification of quasi-stationary segments can be employed for complex financial time series.

We investigate design-level structural transformations that aim at easier subsequent verification of real-time systems with shared data variables, modelled as networks of extended timed automata (ETA). Our contributions to this end are the following: (1) we first equip ETA with an operator for layered composition, intermediate between parallel and sequential composition. Under certain non-interference and/or precedence conditions imposed on the structure of the ETA networks, the communication closed layer (CCL) laws and associated partial-order (po-) and (layered) reachability equivalences are shown to hold. (2) Next, we investigate (under certain cycle conditions on the ETA) the (reachability preserving) transformations of separation and flattening aimed at reducing the number of cycles of the ETA. (3) We then show that our separation and flattening in (2) may be applied together with the CCL laws in (1), in order to restructure ETA networks such that the verification of layered reachability properties is rendered easier. This interplay of the three structural transformations (separation, flattening, and layering) is demonstrated on an enhanced version of Fischer’s real-time mutual exclusion protocol for access to multiple critical sections.

In this demonstration paper we present an application to compare and evaluate machine learning methods used for natural language processing within a content analysis framework. Our aim is to provide an example set of possible machine learning results for different inputs to increase the acceptance of using machine learning in settings that originally rely on manual treatment. We will demonstrate the possibility to compare machine learning algorithms regarding the outcome of the implemented approaches. The application allows the user to evaluate the benefit of using machine learning algorithms for content analysis by a visual comparison of their results.

This paper presents a method for synthesising sound and complete tableau calculi. Given a specification of the formal semantics of a logic, the method generates a set of tableau inference rules which can then be used to reason within the logic. The method guarantees that the generated rules form a calculus which is sound and constructively complete. If the logic can be shown to admit finite filtration with respect to a well-defined first-order semantics then adding a general blocking mechanism produces a terminating tableau calculus. The process of generating tableau rules can be completely automated and produces, together with the blocking mechanism, an automated procedure for generating tableau decision procedures. For illustration we show the workability of the approach for propositional intuitionistic logic.