Boris Ryabko

Followers

Following

Co-authors

Public Views

Robert Lickliter

Florida International University

Mark Rowlands

University of Miami

NANCY SOLOMON

Miami University

David Seamon

Kansas State University

T. Florian Jaeger

University of Rochester

Adam M Croom

Case Western Reserve University

Eros Carvalho

Universidade Federal do Rio Grande do Sul

Samuel LÉZÉ

École Normale Supérieure de Lyon

Larissa Swedell

Queens College of the City University of New York

Ibrahim Cahit

Near East University

Interests

Uploads

Papers by Boris Ryabko

Information-theoretical analysis of two Shannon's ciphers

We describe generalized running key ciphers and apply them for analysis of two Shannon's methods.... more

Download

Information-Theoretic Approach to Estimating the Capacity of Distributed Memory Systems

Problems of Information Transmission, Apr 1, 2018

Systems with cash memory (or more generally, with distributed memory) are very widely used in inf... more Systems with cash memory (or more generally, with distributed memory) are very widely used in information technologies. Such are content delivery networks (CDN) of various types, which deliver digital movies, books, and similar content; peer-to-peer (P2P) networks, where millions of members exchange various information; and many other systems and devices of this kind. We introduce the notions of capacity and entropy efficiency for distributed memory systems, propose methods for estimating these quantities, and give an example of their application.

The experimental distinguishing attack on RC4

IACR Cryptology ePrint Archive, 2006

The output of RC4 was analyzed using the "book stack" test for randomness from . It is experiment... more

Download

Information-Theoretic method for classification of texts

Problems of Information Transmission, May 13, 2017

We consider a method for automatic (i.e., unmanned) text classification based on methods of unive... more We consider a method for automatic (i.e., unmanned) text classification based on methods of universal source coding (or "data compression"). We show that under certain restrictions the proposed method is consistent, i.e., the classification error tends to zero with increasing text lengths. As an example of practical use of the method we consider the classification problem for scientific texts (research papers, books, etc.). The proposed method is experimentally shown to be highly efficient.

Download

Reduction of the secret key length in the perfect cipher by data compression and randomisation

arXiv (Cornell University), Jul 18, 2023

Perfect ciphers have been a very attractive cryptographic tool ever since C. Shannon described th... more Perfect ciphers have been a very attractive cryptographic tool ever since C. Shannon described them. Note that, by definition, if a perfect cipher is used, no one can get any information about the encrypted message without knowing the secret key. We consider the problem of reducing the key length of perfect ciphers, because in many applications the length of the secret key is a crucial parameter. This paper describes a simple method of key length reduction. This method gives a perfect cipher and is based on the use of data compression and randomisation, and the average key length can be made close to Shannon entropy (which is the key length limit). It should be noted that the method can effectively use readily available data compressors (archivers).

Download

Unconditionally secure ciphers with a short key for a source with unknown statistics

arXiv (Cornell University), Mar 27, 2023

We consider the problem of constructing an unconditionally secure cipher with a short key for the... more We consider the problem of constructing an unconditionally secure cipher with a short key for the case where the probability distribution of encrypted messages is unknown. Note that unconditional security means that an adversary with no computational constraints can obtain only a negligible amount of information ("leakage") about an encrypted message (without knowing the key). Here we consider the case of a priori (partially) unknown message source statistics. More specifically, the message source probability distribution belongs to a given family of distributions. We propose an unconditionally secure cipher for this case. As an example, one can consider constructing a single cipher for texts written in any of the languages of the European Union. That is, the message to be encrypted could be written in any of these languages.

Download

Development of an information-theoretical method of attribution of literary texts

We propose an information-theoretical method of attribution of literary texts, developed within t... more We propose an information-theoretical method of attribution of literary texts, developed within the framework of information theory and mathematical statistics. Using the proposed method, the following two problems of disputed authorship in Russian and Soviet literature were investigated: i) the problem of false attribution of some novels to Nekrasov and ii) the problem of dubious attribution of two novels to Bulgakov. The research has shown the high efficiency of the data-compression method for attribution of literary texts.

A Pseudo-Random Generator Whose Output is a Normal Sequence

International Journal of Foundations of Computer Science, Sep 22, 2021

Pseudo-random number generators (PRNGs) are widely used in computer simulation, cryptography, and... more Pseudo-random number generators (PRNGs) are widely used in computer simulation, cryptography, and many other fields. In this paper, we describe a PRNG class, which, firstly, has been successfully tested using the most powerful modern test batteries, and secondly, is proved to consist of generators that generate normal sequences. The latter property means that, for any generated sequence [Formula: see text] and any binary word [Formula: see text], we have [Formula: see text] where [Formula: see text] is the number of occurrences of [Formula: see text] in the sequence [Formula: see text], [Formula: see text].

The time-adaptive statistical testing for random number generators

International Symposium on Information Theory and its Applications, Oct 24, 2020

The problem of constructing effective statistical tests for random number generators (RNG) is con... more The problem of constructing effective statistical tests for random number generators (RNG) is considered. Currently, there are hundreds of RNG statistical tests that are often combined into so-called batteries, each containing from a dozen to more than one hundred tests. When a battery test is used, it is applied to a sequence generated by the RNG, and the calculation time is determined by the length of the sequence and the number of tests. Generally speaking, the longer the sequence, the smaller deviations from randomness can be found by a specific test. So, when a battery is applied, on the one hand, the "better" tests are in the battery, the more chances to reject a "bad" RNG. On the other hand, the larger the battery, the less time can be spent on each test and, therefore, the shorter the test sequence. In turn, this reduces the ability to find small deviations from randomness. To reduce this trade-off, we propose an adaptive way to use batteries (and other sets) of tests, which requires less time but, in a certain sense, preserves the power of the original battery. We call this method time-adaptive battery of tests. The suggested method is based on the theorem which describes asymptotic properties of the so-called p-values of tests. Namely, the theorem claims that if the RNG can be modelled by a stationary ergodic source, the value -log π(x 1 x 2 ...x n )/n goes to 1 -h when n grows, where x 1 x 2 ... is the sequence, π( ) is the p-value of the most powerful test, h is the limit Shannon entropy of the stationary ergodic source. Keywords statistical test • randomness testing • random number generators • adaptive statistical test • battery of tests

Download

Computer Capacity As a Tool For the Processors Development Analysis

In this paper, we analyze the development of processors in recent decades in order to find the fa... more In this paper, we analyze the development of processors in recent decades in order to find the factors that most strongly influence it. We show how to use the analytically calculated characteristic Computer Capacity to identify these factors and quantify their influence. Presented work describes the factors that have the greatest impact on the processors development and shows which of them have the potential for further development, and which have reached saturation.

Studying hunting behaviour in the striped field mouse using data compression

Acta Ethologica, Apr 13, 2017

We compare predatory behaviour towards a mobile insect in three species of small mammals: the gra... more We compare predatory behaviour towards a mobile insect in three species of small mammals: the granivorous striped field mouse, the insectivorous common shrew and the Norway rat (a generalist). The striped field mouse displays a surprisingly efficient hunting stereotype. We apply the data compression method (Ryabko et al. Theory Comput Syst 52:133-147, 2013) to compare the complexity of hunting behavioural patterns and to evaluate the flexibility of stereotypes and their succinctness. Norway rats demonstrated the highest level of complexity of hunting behaviour, with the highest proportion of 'auxiliary' and 'noise' elements and relatively low proportion of 'key' elements in their behaviours. The predominance of 'key' elements resulted in similarly low levels of complexity of hunting stereotypes in striped field mice and shrews. The similarity between hunting stereotypes of the insectivorous shrew and the granivorous striped field mouse enables us to argue about evolutionary roots of hunting behaviour in small mammals. We show that this method is a useful tool for comparing ethograms as 'biological texts'.

Download

Classification by compression: Application of information-theory methods for the identification of themes of scientific texts

Automatic Documentation and Mathematical Linguistics, Jun 1, 2017

Abstract⎯A method for automatic classification of scientific texts based on data compression is p... more Abstract⎯A method for automatic classification of scientific texts based on data compression is proposed. The method is implemented and investigated based on the data from an archive of scientific texts (arXiv.org) and in the CyberLeninka scientific electronic library (CyberLeninka.ru). Experiments showed that the method correctly identified the themes of scientific texts with a probability of 75-95%; its accuracy depends on the quality of the original data.

Download

Application of the distinguishing attack to lightweight block ciphers

By definition, a distinguishing attack is any test that allows distinguishing the encrypted data ... more By definition, a distinguishing attack is any test that allows distinguishing the encrypted data from random data and modern symmetric-key ciphers must be immune to such an attack. In other words, modern block ciphers are designed to have ciphertext indistinguishability. If an algorithm that can distinguish the cipher output from random faster than a brute force search is found, then that is considered a break of the cipher. We applied the distinguishing attack to modern lightweight block ciphers Speck, Simon, Simeck, HIGHT and LEA using statistical test “Book Stack”. We experimentally found a minimal number of rounds where encrypted data looked like random bit-sequence. It is shown, that this number of rounds varies significantly for different ciphers. In particular, our experiments and estimations show that cipher Simeck48/96 is not immune to distinguishing attack.

Fast adaptive arithmetic code for large alphabet sources with asymmetrical distributions

We address the problem of constructing an adaptive arithmetic code in the case where the source a... more We address the problem of constructing an adaptive arithmetic code in the case where the source alphabet is large and there are lots of different symbols with equal counts of occurrence. For an alphabet of N symbols and r distinct symbol weights we describe a code for which the number of operations needed for encoding and decoding is equal to

Application of data compression techniques to time series forecasting

arXiv (Cornell University), Apr 7, 2019

It is known in Information Theory that the problems of data compression and time series forecasti... more It is known in Information Theory that the problems of data compression and time series forecasting are closely related. More precisely, in the case of stationary processes, a data compressor which gives an asymptotically minimal length of a compressed sequence can be used as a time-series predictor whose prediction error (per symbol) is asymptotically minimal, see details in the book (Ryabko, Astola, Malyutov, Compression-Based Methods of Statistical Analysis and Prediction of Time Series. Springer, 2016). In this study we show that standard well-known file compression programs (zlib, bzip2, etc.) are able to forecast real-world time series data well. The strength of our approach is its ability to use a set of data compression algorithms and "automatically" choose the best one of them during the process of forecasting. Besides, modern data-compressors are able to find many kinds of latent regularities using some methods of artificial intelligence (for example, some data-compressors are based on finding the smallest formal grammar that describes the time series). Thus, our approach makes it possible to apply some particular methods of artificial intelligence for time-series forecasting. As examples of the application of the proposed method, we made forecasts for the monthly T-index and the Kp-index time series using standard compressors. In both cases, we used the Mean Absolute Error (MAE) as an accuracy measure. For the monthly T-index time series, we made 18 forecasts beyond the available data for each month since January 2011 to July 2017. We show that, in comparison with the forecasts made by the Australian Bureau of Meteorology, our method more accurately predicts one value ahead. The Kp-index time series consists of 3-hour values ranging from 0 to 9. For each day from February 4, 2018 to March 28, 2018, we made forecasts for 24 values ahead. We compared our forecasts with the forecasts made by the Space Weather Prediction Center (SWPC). The results showed that the accuracy of our method is similar to the accuracy of the SWPC's method. As in the previous case, we also obtained more accurate one-step forecasts.

Download

On asymptotically optimal methods of prediction and adaptive coding for Markov sources with unknown memory

The problem of predicting a sequence x 1 , x 2 , ... generated by a discrete source with unknown ... more The problem of predicting a sequence x 1 , x 2 , ... generated by a discrete source with unknown statistics is considered. Each letter x t+1 is predicted using information on the word x 1 x 2 • • • x t only. In fact, this problem is a classical problem which has received much attention. Its history can be traced back to Laplace. To estimate the efficiency of a method of prediction, three quantities are considered: the precision as given by the Kullback-Leibler divergence, the memory size of the program needed to implement the method on a computer, and the time required, measured by the number of binary operations needed at each time instant. A method is presented for which the memory size and the average time are close to the minimum. The results can readily be translated to results about adaptive coding.

Download

A new statistical testing for random numbers and its application to some cryptographic problems

We address the problem of testing the hypothesis H0 that the letters from some alphabet A = {a1, ... more We address the problem of testing the hypothesis H0 that the letters from some alphabet A = {a1, a2, . . . , aK } are distributed uniformly (i.e. p(a1) = p(a2) = . . . = p(a k ) = 1/K) against the alternative hypothesis H1 that the true distribution is not uniform, in the case when K is large. First, let us explain the main idea of the new test. Let there be given a sample which can be used for testing. The sample is divided into two parts, which are called the training sample and the testing sample. The training sample is used for estimation of frequencies of the letter occurrences. After that the letters of the alphabet A are combined into subsets A1, A2, . . . , As, s ≥ 2, in such a way that, first, one subset contains letters with close (or even equal) frequencies of occurrence and, second, s is much less than K (say, K = 2 20 , s = 2). Then, the set of subsets {A1, A2, . . . , As} is considered as a new alphabet and the new hypotheses Ĥ0 : p(A1) = |A1|/K, p(A2) = |A2|/K, . . . , p(As) = |As|/K and Ĥ1 = ¬ Ĥ0 are tested based on the second ('testing') part of the sample. Obviously, if H0 is true, then Ĥ0 is also true and, if Ĥ1 is true, then H1 is true. That is why this new test can be used for testing the initial H0 and H1. The idea of such a scheme is quite simple. If H1 is true, then there are letters with relatively large and relatively small probabilities. Generally speaking, the high-probable letters will have relatively large frequencies of occurrence and will be accumulated in some subsets Ai whereas low-probable letters will be accumulated in the other subsets. That is why this difference can be found based on the testing sample. It should be pointed out that a decrease in the number of categories from large K to small s can essentially increase the power of the test and, therefore, can essentially decrease the required sample size. For example, let the number of categories K is even and let H1 is as follows: , where δ ∈ (0, 1), {i1, . . . , i K/2 } {i (K/2)+1 , . . . , iK } = {1, . . . , K}. It turns out that the new test can be successfully applied when the total sample size is O( √ K) instead of const K for usual χ 2 test. Claim 1. Let the new test be applied for testing H0 and H1. Then, for each δ ∈ (0, 1) and α ∈ (0, 1) there exist such a training sample size m(δ) and a testing sample size n(δ) that i) (m(δ) + n(δ)) = O( √ K), ii) the level of significance of the test is α and the Type II error is less than 1/2. The natural question is why the case of large alphabet size K can be interesting. It turns out that there exist such random bit sequences x1x2 . . . which, from the one hand, are very far from truly random and, on the other hand, the distribution of the words x1 . .

Download

On asymptotically optimal methods of prediction and adaptive coding

Download

Linear hash-functions and their applications to error detection and correction

arXiv (Cornell University), Aug 20, 2020

We describe and explore so-called linear hash functions and show how they can be used to build er... more We describe and explore so-called linear hash functions and show how they can be used to build error detection and correction codes. The method can be applied for different types of errors (for example, burst errors). When the method is applied to a model where number of distorted letters is limited, the obtained estimate of its performance is slightly better than the known Varshamov-Gilbert bound. We also describe random code whose performance is close to the same boundary, but its construction is much simpler. In some cases the obtained methods are simpler and more flexible than the known ones. In particular, the complexity of the obtained error detection code and the well-known CRC code is close, but the proposed code, unlike CRC, can detect with certainty errors whose number does not exceed a predetermined limit.

Download

Using Information Theory to Study the Efficiency and Capacity of Computers and Similar Devices

arXiv (Cornell University), Mar 18, 2010

We address the problems of estimating the computer efficiency and the computer capacity. We defin... more We address the problems of estimating the computer efficiency and the computer capacity. We define the computer efficiency and capacity and suggest a method for their estimation, based on the analysis of processor instructions and kinds of accessible memory. It is shown how the suggested method can be applied to estimate the computer capacity. In particular, this consideration gives a new look at the organization of the memory of a computer. Obtained results can be of some interest for practical applications.

Download

Information-theoretical analysis of two Shannon's ciphers

We describe generalized running key ciphers and apply them for analysis of two Shannon's methods.... more

Download

Information-Theoretic Approach to Estimating the Capacity of Distributed Memory Systems

Problems of Information Transmission, Apr 1, 2018