Сборник трудов конференции "Информационные Технологии и Системы 2013"
We introduce a new compression scheme for high-dimensional vectors that approximates the vectors using sums of M codewords coming from M different codebooks. We show that the proposed scheme permits efficient distance and scalar product computations between compressed and uncompressed vectors. We further suggest vector encoding and codebook learning algorithms that can minimize the coding error within the proposed scheme. In the experiments, we demonstrate that the proposed compression can be used instead of or together with product quantization. Compared to product quantization and its optimized versions, the proposed compression approach leads to lower coding approximation errors, higher accuracy of approximate nearest neighbor search in the datasets of visual descriptors, and lower image classification error, whenever the classifiers are learned on or applied to compressed vectors.
The volume contains proceedings of the XIII International symposium on problems of redundancy in information and control systems.
A white binary noise information distortion model is proposed in previous our papers.
In this paperwork we analyze limiting behavior of CRC error probabilities with low values of p, which stands for distortion probability of a single bit that belongs to the information being transferred.
Using acquired theoretical results we assess values of error probabilities to claim transferred information packet undistorted having at least one distortion for protocols E1 and TTSI EN 302307 as examples.
The paper considers certain probability-theoretic models of packet mode-transferred information distortions. Attention is drawn mainly to distortions, including possible interferences influencing multiple transfer cycles. Distortions are modeled by a consequential impacts that are defined by dependent random variables. K-dimensioned values of CRC, respectively allow representation as a sum of k-dimensioned independent random variables.
In some cases it is possible to bring them to a sum of independent terms in a k-dimensioned vector space over a two-element field and, afterwards, apply to them existing limit theorems dealing with convergence to uniform distributions.
The paper discusses prospects for impacts stretching to m cycles of acquiring convergence conditions for CRC distribution as a sum of m-dependent terms or ones not interconnected to a non-homogeneous Markov chain.
Background Reliable and comparable data on causes of death are crucial for public health analysis, but the usefulness of these data can be markedly diminished when the approach to coding is not standardized across territories and/or over time. Because the Russian system of producing information on causes of death is highly decentralized, there may be discrepancies in the coding practices employed across the country. In this study, we evaluate the uniformity of cause-of-death coding practices across Russian regions using an indirect method. Methods Based on 2002–2012 mortality data, we estimate the prevalence of the major causes of death (70 causes) in the mortality structures of 52 Russian regions. For each region-cause combination we measured the degree to which the share of a certain cause in the mortality structure of a certain region deviates from the respective inter-regional average share. We use heat map visualization and a regression model to determine whether there is regularity in the causes and the regions that is more likely to deviate from the average level across all regions. In addition to analyzing the comparability of cause-specific mortality structures in a spatial dimension, we examine the regional cause-of-death time series to identify the causes with temporal trends that vary greatly across regions. Results A high level of consistency was found both across regions and over time for transport accidents, most of the neoplasms, congenital malformations, and perinatal conditions. However, a high degree of inconsistency was found for mental and behavioral disorders, diseases of the nervous system, endocrine disorders, ill-defined causes of death, and certain cardiovascular diseases. This finding suggests that the coding practices for these causes of death are not uniform across regions. The level of consistency improves when causes of death can be grouped into broader diagnostic categories. Conclusion This systematic analysis allows us to present a broader picture of the quality of cause-of-death coding at the regional level. For some causes of death, there is a high degree of variance across regions in the likelihood that these causes will be chosen as the underlying causes. In addition, for some causes of death the mortality statistics reflect the coding practices, rather than the real epidemiological situation.
We consider certain spaces of functions on the circle, which naturally appear in harmonic analysis, and superposition operators on these spaces. We study the following question: which functions have the property that each their superposition with a homeomorphism of the circle belongs to a given space? We also study the multidimensional case.
We consider the spaces of functions on the m-dimensional torus, whose Fourier transform is p -summable. We obtain estimates for the norms of the exponential functions deformed by a C1 -smooth phase. The results generalize to the multidimensional case the one-dimensional results obtained by the author earlier in “Quantitative estimates in the Beurling—Helson theorem”, Sbornik: Mathematics, 201:12 (2010), 1811 – 1836.
We consider the spaces of function on the circle whose Fourier transform is p-summable. We obtain estimates for the norms of exponential functions deformed by a C1 -smooth phase.