Showing papers by "Helsinki University of Technology published in 2017"

The Genetic Landscape of Renal Complications in Type 1 Diabetes.

Abstract: Diabetes is the leading cause of ESRD. Despite evidence for a substantial heritability of diabetic kidney disease, efforts to identify genetic susceptibility variants have had limited success. We extended previous efforts in three dimensions, examining a more comprehensive set of genetic variants in larger numbers of subjects with type 1 diabetes characterized for a wider range of cross-sectional diabetic kidney disease phenotypes. In 2843 subjects, we estimated that the heritability of diabetic kidney disease was 35% (P=6.4×10-3). Genome-wide association analysis and replication in 12,540 individuals identified no single variants reaching stringent levels of significance and, despite excellent power, provided little independent confirmation of previously published associated variants. Whole-exome sequencing in 997 subjects failed to identify any large-effect coding alleles of lower frequency influencing the risk of diabetic kidney disease. However, sets of alleles increasing body mass index (P=2.2×10-5) and the risk of type 2 diabetes (P=6.1×10-4) associated with the risk of diabetic kidney disease. We also found genome-wide genetic correlation between diabetic kidney disease and failure at smoking cessation (P=1.1×10-4). Pathway analysis implicated ascorbate and aldarate metabolism (P=9.0×10-6), and pentose and glucuronate interconversions (P=3.0×10-6) in pathogenesis of diabetic kidney disease. These data provide further evidence for the role of genetic factors influencing diabetic kidney disease in those with type 1 diabetes and highlight some key pathways that may be responsible. Altogether these results reveal important biology behind the major cause of kidney disease.

Residual Connections Encourage Iterative Inference

Abstract: Residual networks (Resnets) have become a prominent architecture in deep learning. However, a comprehensive understanding of Resnets is still a topic of ongoing research. A recent view argues that Resnets perform iterative refinement of features. We attempt to further expose properties of this aspect. To this end, we study Resnets both analytically and empirically. We formalize the notion of iterative refinement in Resnets by showing that residual connections naturally encourage features of residual blocks to move along the negative gradient of loss as we go from one block to the next. In addition, our empirical analysis suggests that Resnets are able to perform both representation learning and iterative refinement. In general, a Resnet block tends to concentrate representation learning behavior in the first few layers while higher layers perform iterative refinement of features. Finally we observe that sharing residual layers naively leads to representation explosion and counterintuitively, overfitting, and we show that simple existing strategies can help alleviating this problem.

The Genetic Landscape of Renal Complications in Type 1 Diabetes

Abstract: Diabetes is the leading cause of ESRD. Despite evidence for a substantial heritability of diabetic kidney disease, efforts to identify genetic susceptibility variants have had limited success. We extended previous efforts in three dimensions, examining a more comprehensive set of genetic variants in larger numbers of subjects with type 1 diabetes characterized for a wider range of cross-sectional diabetic kidney disease phenotypes. In 2843 subjects, we estimated that the heritability of diabetic kidney disease was 35% (P=6.4×10-3). Genome-wide association analysis and replication in 12,540 individuals identified no single variants reaching stringent levels of significance and, despite excellent power, provided little independent confirmation of previously published associated variants. Whole-exome sequencing in 997 subjects failed to identify any large-effect coding alleles of lower frequency influencing the risk of diabetic kidney disease. However, sets of alleles increasing body mass index (P=2.2×10-5) and the risk of type 2 diabetes (P=6.1×10-4) associated with the risk of diabetic kidney disease. We also found genome-wide genetic correlation between diabetic kidney disease and failure at smoking cessation (P=1.1×10-4). Pathway analysis implicated ascorbate and aldarate metabolism (P=9.0×10-6), and pentose and glucuronate interconversions (P=3.0×10-6) in pathogenesis of diabetic kidney disease. These data provide further evidence for the role of genetic factors influencing diabetic kidney disease in those with type 1 diabetes and highlight some key pathways that may be responsible. Altogether these results reveal important biology behind the major cause of kidney disease.

The Fulde-Ferrell-Larkin-Ovchinnikov state for ultracold fermions in lattice and harmonic potentials: a review.

Abstract: We review the concepts and the present state of theoretical studies of spin-imbalanced superfluidity, in particular the elusive Fulde-Ferrell-Larkin-Ovchinnikov (FFLO) state, in the context of ultracold quantum gases. The comprehensive presentation of the theoretical basis for the FFLO state that we provide is useful also for research on the interplay between magnetism and superconductivity in other physical systems. We focus on settings that have been predicted to be favourable for the FFLO state, such as optical lattices in various dimensions and spin-orbit coupled systems. These are also the most likely systems for near-future experimental observation of the FFLO state. Theoretical bounds, such as Bloch's and Luttinger's theorems, and experimentally important limitations, such as finite-size effects and trapping potentials, are considered. In addition, we provide a comprehensive review of the various ideas presented for the observation of the FFLO state. We conclude our review with an analysis of the open questions related to the FFLO state, such as its stability, superfluid density, collective modes and extending the FFLO superfluid concept to new types of lattice systems.

Experimental Evaluation of Simulcast for WebRTC

Abstract: One of the important use cases for WebRTC is multi-party audio/video conferencing, and there are multiple services that offer that functionality. Two aspects that the maintainers of such services often try to optimize are the maximum size of their conferences, and the cost of their infrastructure. Simulcast is one approach to addressing these problems in conference systems that build upon video routers. It consists of the simultaneous transmission of multiple versions of a participant's video stream, which use different resolutions, frame rates and, most importantly, different bit rates. In this article we present the analysis and experimental evaluation of a conference system based on WebRTC using simulcast with the VP8 codec. We examine the computational and bandwidth requirements of both the server and the clients, as well as the end-to-end video quality. Our results show significant gains for the receivers and the infrastructure, with minor penalties to the image quality and resource usage for the senders.

Residual Connections Encourage Iterative Inference

Abstract: Residual networks (Resnets) have become a prominent architecture in deep learning. However, a comprehensive understanding of Resnets is still a topic of ongoing research. A recent view argues that Resnets perform iterative refinement of features. We attempt to further expose properties of this aspect. To this end, we study Resnets both analytically and empirically. We formalize the notion of iterative refinement in Resnets by showing that residual connections naturally encourage features of residual blocks to move along the negative gradient of loss as we go from one block to the next. In addition, our empirical analysis suggests that Resnets are able to perform both representation learning and iterative refinement. In general, a Resnet block tends to concentrate representation learning behavior in the first few layers while higher layers perform iterative refinement of features. Finally we observe that sharing residual layers naively leads to representation explosion and counterintuitively, overfitting, and we show that simple existing strategies can help alleviating this problem.

Improved Distributed Degree Splitting and Edge Coloring

Abstract: The degree splitting problem requires coloring the edges of a graph red or blue such that each node has almost the same number of edges in each color, up to a small additive discrepancy. The directed variant of the problem requires orienting the edges such that each node has almost the same number of incoming and outgoing edges, again up to a small additive discrepancy. We present deterministic distributed algorithms for both variants, which improve on their counterparts presented by Ghaffari and Su [SODA'17]: our algorithms are significantly simpler and faster, and have a much smaller discrepancy. This also leads to a faster and simpler deterministic algorithm for $(2+o(1))\Delta$-edge-coloring, improving on that of Ghaffari and Su.

Boundary correlations in planar LERW and UST

Abstract: We find explicit formulas for the probabilities of general boundary visit events for planar loop-erased random walks, as well as connectivity events for branches in the uniform spanning tree. We show that both probabilities, when suitably renormalized, converge in the scaling limit to conformally covariant functions which satisfy partial differential equations of second and third order, as predicted by conformal field theory. The scaling limit connectivity probabilities also provide formulas for the pure partition functions of multiple $\mathrm{SLE}_\kappa$ at $\kappa=2$.

Level algebras and $\boldsymbol{s}$-lecture hall polytopes

Abstract: Given a family of lattice polytopes, a common endeavor in Ehrhart theory is the classification of those polytopes in the family that are Gorenstein, or more generally level. In this article, we consider these questions for $\boldsymbol{s}$-lecture hall polytopes, which are a family of simplices arising from $\boldsymbol{s}$-lecture hall partitions. In particular, we provide concrete classifications for both of these properties purely in terms of $\boldsymbol{s}$-inversion sequences. Moreover, for a large subfamily of $\boldsymbol{s}$-lecture hall polytopes, we provide a more geometric classification of the Gorenstein property in terms of its tangent cones. We then show how one can use the classification of level $\boldsymbol{s}$-lecture hall polytopes to construct infinite families of level $\boldsymbol{s}$-lecture hall polytopes, and to describe level $\boldsymbol{s}$-lecture hall polytopes in small dimensions.

Feature Selection in Learning Using Privileged Information

Abstract: The paper considers the problem of feature selection in learning using privileged information (LUPI), where some of the features (referred to as privileged ones) are only available for training, while being absent for test data. In the latest implementation of LUPI, these privileged features are approximated using regressions constructed on standard data features, but this approach could lead to polluting the data with poorly constructed and/or noisy features. This paper proposes a privileged feature selection method that addresses some of these issues. Since not many LUPI datasets are currently available in open access, while calibration of parameters of the proposed method requires testing it on a wide variety of datasets, a modified version of the method for traditional machine learning paradigm (i.e., without privileged features) was also studied. This lead to a novel mechanism of error rate reduction by constructing and selecting additional regression-based features capturing mutual relationships among standard features. The results on calibration datasets demonstrate the efficacy of the proposed feature selection method both for standard classification problems (tested on multiple calibration datasets) and for LUPI (for several datasets described in the literature).

Multi-Focus Image Fusion Using Sparse Representation and Coupled Dictionary Learning

Abstract: We address the multi-focus image fusion problem, where multiple images captured with different focal settings are to be fused into an all-in-focus image of higher quality. Algorithms for this problem necessarily admit the source image characteristics along with focused and blurred features. However, most sparsity-based approaches use a single dictionary in focused feature space to describe multi-focus images, and ignore the representations in blurred feature space. We propose a multi-focus image fusion approach based on sparse representation using a coupled dictionary. It exploits the observations that the patches from a given training set can be sparsely represented by a couple of overcomplete dictionaries related to the focused and blurred categories of images and that a sparse approximation based on such coupled dictionary leads to a more flexible and therefore better fusion strategy than the one based on just selecting the sparsest representation in the original image estimate. In addition, to improve the fusion performance, we employ a coupled dictionary learning approach that enforces pairwise correlation between atoms of dictionaries learned to represent the focused and blurred feature spaces. We also discuss the advantages of the fusion approach based on coupled dictionary learning, and present efficient algorithms for fusion based on coupled dictionary learning. Extensive experimental comparisons with state-of-the-art multi-focus image fusion algorithms validate the effectiveness of the proposed approach.

Joint Data and Key Distribution of Simple, Multiple, and Multidimensional Linear Cryptanalysis Test Statistic and Its Impact to Data Complexity

Abstract: The power of a statistical attack is inversely proportional to the number of plaintexts needed to recover information on the encryption key. By analyzing the distribution of the random variables involved in the attack, cryptographers aim to provide a good estimate of the data complexity of the attack. In this paper, we analyze the hypotheses made in simple, multiple, and multidimensional linear attacks that use either non-zero or zero correlations, and provide more accurate estimates of the data complexity of these attacks. This is achieved by taking, for the first time, into consideration the key variance of the statistic for both the right and wrong keys. For the family of linear attacks considered in this paper, we differentiate between the attacks which are performed in the known-plaintext and those in the distinct-known-plaintext model.

On modeling weakly stationary processes

Abstract: In this article, we show that a general class of weakly stationary time series can be modeled applying Gaussian subordinated processes. We show that, for any given weakly stationary time series $(z_t)_{z\in\mathbb{N}}$ with given equal one-dimensional marginal distribution, one can always construct a function $f$ and a Gaussian process $(X_t)_{t\in\mathbb{N}}$ such that $\left(f(X_t)\right)_{t\in\mathbb{N}}$ has the same marginal distributions and, asymptotically, the same autocovariance function as $(z_t)_{t\in\mathbb{N}}$. Consequently, we obtain asymptotic distributions for the mean and autocovariance estimators by using the rich theory on limit theorems for Gaussian subordinated processes. This highlights the role of Gaussian subordinated processes in modeling general weakly stationary time series. We compare our approach to standard linear models, and show that our model is more flexible and requires weaker assumptions.