About
132
Publications
4,848
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
487
Citations
Introduction
Skills and Expertise
Current institution
Additional affiliations
September 1976 - July 1977
May 1982 - June 2014
Publications
Publications (132)
The paper proposes a protocol of a multiple-time, two-round digital friend-or-foe authentication for a group of users. The protocol is based on the following scheme: in each authentication session, group members sign individual w-blocks of a message provided by verifier using the Winternitz signature scheme. The verifier verifies the validity of th...
Paraphrase generation is a fundamental problem in natural language processing. Due to the significant success of transfer learning, the “pre-training → fine-tuning” approach has become the standard. However, popular general pre-training methods typically require extensive datasets and great computational resources, and the available pre-trained mod...
This paper studies binary reverse multi-delimiter (RMD) data compression codes. RMD codes have a range of useful properties, such as unique decodability, completeness, universality, synchronizability, recognition using a finite automaton, and the ability for rapid data retrieval within an encoded file. The authors have constructed a simple monotoni...
We study a class of binary reverse multi-delimiter (RMD) data compression codes in application to natural language text compression. The RMD-codewords start with delimiters, i.e., prefixes of the form that cannot occur in other places of the codeword. The position of the delimiter in an RMD codeword differs from its position in “direct” multi-delim...
SPECIALIZED PRE-TRAINING OF NEURAL NETWORKS ON SYNTHETIC DATA FOR IMPROVING PARAPHRASE GENERATION Abstract. Generating paraphrases is a fundamental problem in natural language processing. In light of the significant success of transfer learning technology, the “pre-training fine-tuning” approach has become the standard. However, popular general-pur...
An overview of PARCS (Parallel Asynchronous Recursive Control Space) technology developments is provided. The concept of the control space, i.e., a model apparatus, based on which the logical structure of the investigated problem (system) is described, and dynamic changes in it are reflected, is considered. The PARCS model is proposed whose applica...
A solution to the problem of low computational efficiency of the relational data model is investigated. An extension of the relational algebra with the help of operations over data sets, i.e., the basic construction of CODASYL DBTG proposals, is suggested. A user is able to choose a method for implementing data links depending on the following requ...
Introduction. We study the problem of constructing a locally optimal directed graph when each edge is assigned a fixed numerical weight. Due to the broad interpretation of the numerical weights of edges, the problem of finding the least-cost paths has many applications and is one of the most common and studied problems in the applied theory of algo...
The article describes an evolutionary model for artificial intelligence intended for the design and development of intelligent systems. The key element of the proposed model is the so-called ALF that is an intelligent agent with the ability for self-learning, communication, joint actions, and self-organization among similar agents. The development...
We describe a fast three-round mutual authentication protocol for parties A and B belonging to the same coalition group. Parties A and B keep their own independent long-term private keys that are used in the process of authentication and can be used for other purposes. The scheme assumes an initial setup with a trusted third party T. This party ini...
We describe a fast three-round mutual authentication protocol for parties A and B belonging to the same coalition group. Parties A and B keep their own independent long-term private keys that are used in the process of authentication and can be used for other purposes. The scheme assumes an initial setup with a trusted third party T. This party ini...
The (2,3)-representation of integers utilizes the mixed numeration base of the radix-2 and auxilary radix-3. This representation yields a universal prefix-free binary encoding of all natural numbers with a variety of useful properties: robustness (self-synchronization), local error corrections, statistic regularities of code parameters, etc. The pa...
An enhanced version of a recently introduced family of variable length binary codes with multiple pattern delimiters is presented and discussed. These codes are complete, universal, synchronizable, they have monotonic indexing and allow a standard search in compressed
�les. Comparing the compression rate on natural language texts demonstrates that...
The paper considers details of implementation of the PARCS-WCF system. The PARCS-WCF system provides creation and operation of asynchronous-recursive processes in solving parallel computing tasks both on a single PC and on a computer network. It allows different modules to exchange information providing data exchange in systems with distributed and...
An evolutionary model of constructing artificial intelligence is presented, which is destined for designing and developing intelligent systems. The model allows describing a variety of subject areas with constructing knowledge bases. It has universal means to formally describe tasks and environments for implementing computational processes to solve...
The article describes the method of creation, structure and operation of the method for determining the direction to the sound source. The method relies on a simple mathematical apparatus, which allows it to be implemented on equipment of minimal power, for example, on Arduino microprocessors. The key elements of the method and conditions of use af...
This paper describes algorithms for non-negative factorization of sparse matrices and tensors, which is a popular technology in artificial intelligence in general and in computer linguistics in particular. It is proposed to use the latent Dirichlet distribution to reduce matrices and tensors to block-diagonal form for parallelizing computations and...
A solution to the integrity verification problem is proposed for arithmetic programs with branching and looping statements executed on a remote computational resource. The solution is to replace arithmetic operations such as multiplication and division by corresponding procedures of the addition machine introduced by R. Floyd and D. Knuth. Instruct...
A unified encoding of ordered binary trees with integer-valued labels at their vertices is proposed using linear forms of neighboring members of linear recursive sequences of the form Pn + 2 = an + 2 Pn + 1 + P, where P1 = P2 = 1; a3 and a4 … are natural numbers. Encoding and decoding procedures are simply implemented and use the recursive techniqu...
A solution to the integrity problem for arithmetic programs executing on a remote computing resource is proposed. The solution is to replace arithmetic operations such as multiplication and division by procedures of the addition machine introduced by R. Floyd and D. Knuth. Computations and sequence order of instructions (as well as current meanings...
A new effective algorithm and a system for paraphrase identification have been developed using a machine learning approach. The system architecture has the form of a multilayer classifier. According to their strategies, sub-classifiers of the lower level make decisions about the presence of paraphrase in sentences, while a super-classifier of the u...
A new effective system for identification and verification of text authorship has been developed. The system is created on the basis of machine learning. The originality of the model is caused by a suggested unique profile of the author’s style features. Together with the use of the Support Vector Machine method, this allows us to achieve the high...
Let m1, m2, . . . . , mt be a fixed set of natural integers given in ascending order. A multi-delimiter code Dm1, . . . . ,mt consists of t words of the form 1mi0 and all other binary words with suffixes having one of the forms 01mi0 that cannot occur in other places of a word, i = 1, . . . . ,t. Completeness, universality and density characteristi...
Variable-length splittable codes are derived from encoding sequences of
ordered integer pairs, where one of the pair's components is upper bounded by
some constant, and the other one is any positive integer. Each pair is encoded
by the concatenation of two fixed independent prefix encoding functions applied
to the corresponding components of a pair...
This paper proposes parallel methods of non-negative large sparse matrix factorization – a very popular technique in computational linguistics. Memory usage and data transmitting necessity of factorization algorithm was analysed and optimized. The described effective GPU-based and distributed algorithms were implemented, tested and compared by mean...
A method for developing a structural model of natural language syntax and semantics is proposed. Syntactic and semantic relations between parts of a sentence are presented in the form of a recursive structure called a control space. Numerical characteristics of these data are stored in multidimensional arrays. After factorization, the arrays serve...
A method of developing a structural model of natural language syntax and semantics is proposed. Factorization of lexical combinability arrays obtained from text corpora generates linguistic databases that used for natural language semantic and syntactic analyses.
A new class of forward error correcting codes is introduced. The main idea of the code construction is to utilize special arithmetic properties of input words and codewords considered as whole numbers in the process of encoding and decoding. The numbers are represented in the two-base numeration system with the main radix 2 and the auxiliary radix...
This paper describes a method for automatic detection of semantic relations between concept nodes of a networked ontological knowledge base by analyzing matrices of semantic-syntactic valences of words. These matrices are obtained by means of nonnegative factorization of tensors of syntactic compatibility of words. Such tensors are generated in the...
Binary error-correction encoding using the two-base numeration system with the radices 2 and 3 is proposed. Properties and error-correcting capabilities of such codes for one and two errors are analyzed. A special kind of the (2, 3) code, namely, the lower (2, 3) code is introduced to provide error-correcting properties.
This paper deals with the process of developing a lexical semantic database for Ukrainian language – UkrWordNet. The architecture of the developed system is described in detail. The data storing structure and mechanisms of access to knowledge are reviewed along with the internal logic of the system and some key software modules. The article is also...
Two-base numeration systems are introduced in the paper. They are generalizations of standard numeration systems, which are based on powers of a given radix. The results are reviewed and applications for encoding numbers and trees are given. Further research fields are outlined. A new parallel algorithm for modular exponentiation and new classes of...
The modified system of algorithmic algebras was analysed and its benefits were considered. The set of Floyd-Warshall algorithm schemes in terms of this mathematical apparatus was created, and a generalized parallel scheme of the algorithm was developed. It is valid for a number of target parallel computer architectures. A more detailed scheme was c...
A new family of universal self-synchronizable variable-length codes is introduced. This family is not a generalization or improvement of the existing prefix codes, but is based on a new method of integer representation in a mixed base using the radix-2 and the auxiliary radix-3. Upper length bounds for such codes are obtained. The asymptotic estima...
An implementation of the non-negative matrix factorization algorithm for the purpose of text mining on graphics processing units is presented. Performance gains of more than one order of magnitude are obtained.
This paper develops methods for calculating the semantic similarity (closeness)-relatedness of natural language words. The
concept of semantic relatedness allows one to construct algorithmic models for the context-linguistic analysis with a view
to solving problems such as word sense disambiguation, named entity recognition, natural language text a...
This paper's main result is presenting a new conception in the geometric modeling and visualization - a generalized efficient parallel-and-recursive algorithm with optimal bound complexity O(log2N). Voronoi diagram is one of key elements the algorithm. The algorithm solves in unified manner the variety of interrelated geometrical problems for the c...
An efficient approach to the accurate computer modeling of phenomena and processes is discussed. The “divide-and-conquer”
technique is used to develop a generalized parallel-recursive algorithm for simultaneous solution of a collection of interrelated
problems that use a common unified data structure (weighted concatenable queue) at the merge stage...
The mixed-base numeration system based on series of powers of the numbers 2 and 3 is introduced and investigated. Representation
of numbers in this numeration system is a special case of recursive number decomposition into linear forms in the basis of
orthogonal number sequences. A new universal robust prefix coding of integers with the help of thi...
Theoretical results are reviewed that are concerned with the construction of speed-optimal parallel-pipeline algorithms for
mass calculations in solving filtering problems. The optimality is proved in the corresponding classes of algorithms equivalent
in terms of information graphs. The effectiveness of using the developed algorithmic constructions...
A computational model of estimation of the time complexity of logical circuits constructed from elements of an optical element
base is investigated. A fast parallel multiplier is constructed.
A universal heuristic algorithm is considered that constructs the vector of properties of an object that are distributed among
heterogeneous groups. This algorithm is based on a feed-forward neural network. An automatic system for objects classification
is described. The use of the automatic classification system for a literary authorship attributi...
We offer novel algorithms for efficient encoding/decoding of variable-to-fixed length codes, requiring at most quadratic amount of space: O(L 2), where L is the depth of a coding tree. This is a major improvement compared to exponential O(2 L) usage of space by conventional techniques using complete representations of coding trees in computer's mem...
In this paper, the approach to the development of deductive assistance based on the current state of the System for Automated Deduction, SAD, is described briefly. The SAD may be considered as a three-dimensional "shell" for deductive assistance in a formal natural language environment when solving a number of scientific and industrial tasks requir...
Questions of construction of time-efficient parallel logical circuits of basic arithmetic operations from optical elements are considered. Multiplication and adding circuits are proposed that make it possible to use distinctive features of optical switches to reduce their operating time.
Existing methods of simulation of motion of rigid bodies are considered that take into account collisions and joints of such
bodies. A new fast universal fixed-step-size algorithm of speed constraints is proposed. This algorithm combines all the advantages
of well-known algorithms and is deprived of their drawbacks. Using a modified Dantzig method,...
The system PARCS-JAVA provides software tools for solution of problems on computer networks. It can be installed on heterogeneous computer networks and allows users of small computers to use parallel data processing.
Algorithms are developed that determine canonic forms and grammar categories from individual words and the vocabulary of a text. Algorithms are proved to be highly efficient. The accuracy of determination directly depends on the length of the text.
We show how a digital tree (or trie) structure can be used for both parsing and encoding (in a Variable-Length to Block (VB) or Variable-Length to Variable-Length (VV) fashion) of sequences of symbols from a stochastic source. As an example, we construct a simple VB code based on a fixed database adaptation model, and derive an asymptotic expressio...
This paper proposes to code integers and binary trees by linear forms of aun-1+bun-type, where un-1 and un-1 are adjacent members of some integer sequence. New prefix codes of integers are obtained. Such codes have better characteristics than the well-known Levenshtein code. Data compression by means of coding by linear forms is considered. Linear...
Summary form only given. The average redundancy rate of Krichevsky's sample-based universal block codes was analyzed, in a situation wherein the samples and block codes for compression were produced from two different memoryless sources. It was proven that the average redundancy rate of adaptive block codes O/sub /spl lscr/,T/, constructed using sa...
A coalition is understood to be a group of participants (coalitionists) who can collaborate in order to achieve common objectives. The basic principle of a coalition is the absence of a threat to communication flows within the coalition from its participants. In this paper, two new RSA coalition protocols are considered. According to the first prot...
A quantum Turing machine is considered. A review of basic methodological principles and achievements in the field of quantum
computations is given. Some problems of construction of correct quantum computations and their complexity are considered.
The result of P. Shor concerning the solution of the problems of taking discrete logarithms in polynomi...
A new general algorithm is proposed for computation of the multiprecision modular reduction x mod m. This algorithm has better
time estimates than the well-known Montgomery method and much better application characteristics. The algorithm does not require
changes in the initial and final values of the arguments. The precomputation time is no more t...
A new topological structure called the Fibonacci hypercube, which is obtained recursively using formulas similar to the relations of Fibonacci numbers is introduced. It is proved that this structure posses many properties analogous to the classic Boolean hypercube but has some specific peculiarities connected with vertex valence heterogeneity.
The evaluation of modular reduction x mod n for large numbers is a central problem for most open-key data security systems. In this article the author shows how some accelerating procedures for modular reduction are obtained from a standard linear relationship for the GCD of two numbers.
Researchers are actively searching for discrete topological structures that can be used as a prototype of communication architectures for multiprocessor systems. Structures of this kind have to satisfy specific requirements that allow increasing computer performance through simple combination of structures and ensure ease of message transmission. L...
Applications of special representation of natural numbers as linear forms of the type x F
t−1+y F
1, where F
t−1 and F
t
are adjacent Fibonacci numbers, for constructing effective parallel algorithms of modular exponentiation and factorization are considered in this report. These operations over large numbers are ones of the main in design and anal...
This article analyzes some ideas and proposals advanced by Viktor Mikhailovich Glushkov and his students in the late 1970s and early 1980s. Subsequently, these ideas provided the theoretical foundation for some domestic and foreign projects.
Some peculiarities of PARUS (Parallel Asynchronous Recursively Controlled Systems) technology designed for processing and implementation of recursively parallel algorithms are discussed. Comparative characteristic of PARUS-technology versus other well - known technologies as OCCAM, ADA, MANIFOLD is presented.
This article describes the basic characteristics of a system for parallel programming that supports recursive parallel programming. 8 refs., 8 figs.
Proposed PARCS -tools (PARCS - 'Parallel Asynchronous Recursive Controlled Systems') are intended for programming parallel systems with dynamically changed structures and designed according to the PARCS - technology of programming. Programming system PARCS supports algorithms design and realization for parallel information processing and is based o...
The family of transputer-implemented parallel languages (Parallel C, Parallel FORTRAN, etc.) is augmented with dynamic-parallelism software tools. The proposed tools are implemented on a transputer ring with an arbitrary number of transputers. Resource allocation and scheduling problems are solved automatically by a program switch. The capabilities...
A family of parallel languages realized on transputer systems (Parallel C, Parallel Fortran etc) is extended by software tools of dynamic parallelism. The suggested tools are realized on a transputer ring with an arbitrary quantity of transputers. Problems of the resource distribution and routing are solved automatically by means of the program com...
The article describes the basic characteristics of a system for parallel programming that supports recursive parallel programming.
A formal model of the structure of natural-language sentences is considered, based on the concept of control space. The proposed model allows for recursion of predicative and attributive relations in the language and generalizes the linguistic models of domination trees and constituent systems. It is designed for efficient implementation in paralle...
An algorithm based on parallel programming technology is proposed for solving coordination problems in decentralized local economic models. Examples of decomposition methods for linear distributed systems are considered. The software tools for the solution of these problems are supplied by the PARUS programming system.
Parallel algorithms for dictionary operations (search, insert, delete) in dynamic structures are reviewed. The algorithms considered in the paper are applicable to asynchronous processes on shared memory and to systolic structures of various architectures.
Specially interpreted Petri nets are proposed for use in describing and studying the life cycle of a complex system. This life cycle is the design, fabrication, and application process. The topics of implementing these nets in a program and the methodology of using the proposed apparatus to model the life cycles of complex processes and their aggre...
A recursive nonlinear program scheme is described allowing efficient iterative representation. Examples are cited of the application of the proposed method of recursion elimination to the towers of Hanoi problems, merge sorting, and tree traversal.
