Informatica

We tested the ability of humans and machines (data mining techniques) to assign stress to Slovene words. This is a challenging comparison for machines since humans accomplish the task outstandingly even on unknown words without any context. The goal of finding good machine-made models for stress assignment was set by applying new methods and by making use of a known theory about rules for stress assignment in Slovene. The upgraded data mining methods outperformed expert-defined rules on practically all subtasks, thus showing that data mining can more than compete with humans when constructing formal knowledge about stress assignment is concerned. Unfortunately, compared to humans directly, the data mining methods still failed to achieve as good results as humans on assigning stress to unknown words.

Advanced Encryption Standard (AES) block cipher system is widely used in cryptographic applications. A nonlinear substitution operation is the main factor of the AES cipher system strength. The purpose of the proposed approach is to generate the random S-boxes changing for every change of the secret key. The fact that the S-boxes are randomly key-dependent and unknown is the main strength of the new approach, since both linear and differential cryptanalysis require known S-boxes. In the paper, we briefly analyze the AES algorithm, substitution S-boxes, linear and differential cryptanalysis, and describe a randomly key-dependent S-box and inverse S-box generation algorithm. After that, we introduce the independency measure of the S-box elements, and experimentally investigate the quality of the generated S-boxes.

In the previous papers (Pupeikis, 2000; Genov et al., 2006), a direct approach for estimating the parameters of a discrete-time linear time-invariant (LTI) dynamic system, acting in a closed-loop in the case of additive correlated noise with contaminating outliers uniformly spread in it, is presented. It is assumed here that the parameters of the LQG (Linear Quadratic Gaussian Control) controller are known beforehand. The aim of the given paper is development of a parametric identification approach for a closed-loop system when the parameters of an LTI system as well as that of LQG controller are not known and ought to be estimated. The recursive techniques based on an the M- and GM- estimator algorithms are applied here in the calculation of the system as well as noise filter parameters. Afterwards, the recursive parameter estimates are used in each current iteration to determine unknown parameters of the LQG-controller, too. The results of numerical simulation by computer are discussed.

The aim of the given paper is development of a recursive approach for calculating the statistics of decimated realizations of a basic discrete-time signal, obtained by sampling a continuous-time one. The statistics values are calculated without storing decimated realizations in the memory of a computer. In such a case, only all samples of the basic discrete-time signal are required. The summing calculation amount has been significantly reduced here by applying recursive expressions that use, in the current operation, the respective statistic obtained by processing the samples of a previous decimated realization. We prove the corollaries referring to recursive statistics calculation and present here an example. The results of calculation of the total quantity of computer operations for a different amount of decimated signals are also given and compared to that, determined by using the ordinary formulas.

A transitive distributed knowledge logic is considered. The considered logic S4nD is obtained from multi-modal logic S4n by adding transitive distributed knowledge operator. For a fragment of this logic loop-check-free sequent calculus is proposed. The considered fragment is such that it can be applied for specification and verification of safety properties of knowledge-based distributed systems. By relying on the constructed loop-check-free sequent calculus a PSPACE procedure to determine a termination of backward derivation in considered fragment of the logic S4nD is presented.

A new computational methodology for executing calculations with infinite and infinitesimal quantities is described in this paper. It is based on the principle ‘The part is less than the whole’ introduced by Ancient Greeks and applied to all numbers (finite, infinite, and infinitesimal) and to all sets and processes (finite and infinite). It is shown that it becomes possible to write down finite, infinite, and infinitesimal numbers by a finite number of symbols as particular cases of a unique framework. The new methodology has allowed us to introduce the Infinity Computer working with such numbers (its simulator has already been realized). Examples dealing with divergent series, infinite sets, and limits are given.

This paper presents a novel robust digital image watermarking scheme using subsampling and nonnegative matrix factorization. Firstly, subsampling is used to construct a subimage sequence. Then, based on the column similarity of the subimage sequence, nonnegative matrix factorization (NMF) is applied to decompose the sequence. A Gaussian pseudo-random watermark sequence is embedded in the factorized decomposition coefficients. Because of the high similarity of subimages and meaningful factorization for NMF, the proposed scheme can achieve good robustness, especially to common permutation attacks. Numerical experiment results demonstrate the good performance of the proposed scheme.

This paper examins approaches for translation between English and morphology-rich languages. Experiment with English–Russian and English–Lithuanian revels that “pure” statistical approaches on 10 million word corpus gives unsatisfactory translation. Then, several Web-available linguistic resources are suggested for translation. Syntax parsers, bilingual and semantic dictionaries, bilingual parallel corpus and monolingualWeb-based corpus are integrated in one comprehensive statistical model. Multi-abstraction language representation is used for statistical induction of syntactic and semantic transformation rules called multi-alignment templates. The decodingmodel is described using the feature functions, a log-linear modeling approach and A* search algorithm. An evaluation of this approach is performed on the English–Lithuanian language pair. Presented experimental results demonstrates that the multi-abstraction approach and hybridization of learning methods can improve quality of translation.

In this paper, we present a novel fine-grained access control system for applications where the information flow is critical; the confidentiality of the data is essential and there are a huge number of users who access different portions of an XML document as in military applications. We combine MAC and RBACK models for XML for use in the mentioned type of applications. In accordance with the peculiarities of the target applications, the access control model is structured in such a way that the implementation can be done efficiently for large number of users. In the system presented, instead of using access control lists, we use a security labeling approach in defining the grant rules. By combining the advantages of role-based and mandatory access control schemes, the access control system presented provides a fine-grained, flexible and effective access for applications where the confidentiality of data is crucial. The system is implemented and tested for correctness. Performance analysis is also given.

The present work is concerned with speech recognition using a small or medium size vocabulary. The possibility to use the English speech recognizer for the recognition of Lithuanian was investigated. Two methods were used to deal with such problems: the expert-driven (knowledge-based) method and the data-driven one. Phonological systems of English and Lithuanian were compared on the basis of the knowledge of phonology, and relations between certain Lithuanian and English phonemes were established. Situations in which correspondences between the phonemes were to be established experimentally (i.e., using the data-driven method) and the English phonemes that best matched the Lithuanian sounds or their combinations (e.g., diphthongs) in such situations were identified. The results obtained were used for creating transcriptions of the Lithuanian names and surnames that were used in recognition experiments. The experiments without transcriptions, with a single transcription and with many transcriptions were carried on. The method that allowed finding a small number of best transcriptions was proposed. The recognition rate achieved was as follows: 84.2% with the vocabulary containing 500 word pairs.