Jukka Kohonen

Balanced Data Gathering in Energy-Constrained Sensor Networks

We consider the problem of gathering data from a wireless multi-hop network of energy-constrained sensor nodes to a common base station. Specifically, we aim to balance the total amount of data received from the sensor network during its lifetime against a requirement of sufficient coverage for all the sensor locations surveyed. Our main contribution lies in formulating this balanced data gathering task and in studying the effects of balancing. We give an LP network flow formulation and present experimental results on optimal data routing designs also with impenetrable obstacles between the nodes. We then proceed to consider the effect of augmenting the basic sensor network with a small number of auxiliary relay nodes with less stringent energy constraints. We present an algorithm for finding approximately optimal placements for the relay nodes, given a system of basic sensor locations, and compare it with a straightforward grid arrangement of the relays.

Exact and approximate balanced data gathering in energy-constrained sensor networks

We consider the problem of gathering data from a wireless multi-hop network of energy-constrained sensor nodes to a common base station. Specifically, we aim to balance the total amount of data received from the sensor network during its lifetime against a requirement of sufficient coverage for all the sensor locations surveyed. Our main contribution lies in formulating this balanced data gathering task, studying the effects of balancing, and proposing an approximation algorithm for the problem. Based on an LP network flow formulation, we present experimental results on both optimal and approximate data routing designs, in open transmission ranges and with impenetrable obstacles between the nodes.

Transcriptome analysis reveals signature of adaptation to landscape fragmentation.

We characterize allelic and gene expression variation between populations of the Glanville fritillary butterfly (Melitaea cinxia) from two fragmented and two continuous landscapes in northern Europe. The populations exhibit significant differences in their life history traits, e.g. butterflies from fragmented landscapes have higher flight metabolic rate and dispersal rate in the field, and higher larval growth rate, than butterflies from continuous landscapes. In fragmented landscapes, local populations are small and have a high risk of local extinction, and hence the long-term persistence at the landscape level is based on frequent re-colonization of vacant habitat patches, which is predicted to select for increased dispersal rate. Using RNA-seq data and a common garden experiment, we found that a large number of genes (1,841) were differentially expressed between the landscape types. Hexamerin genes, the expression of which has previously been shown to have high heritability and which correlate strongly with larval development time in the Glanville fritillary, had higher expression in fragmented than continuous landscapes. Genes that were more highly expressed in butterflies from newly-established than old local populations within a fragmented landscape were also more highly expressed, at the landscape level, in fragmented than continuous landscapes. This result suggests that recurrent extinctions and re-colonizations in fragmented landscapes select a for specific expression profile. Genes that were significantly up-regulated following an experimental flight treatment had higher basal expression in fragmented landscapes, indicating that these butterflies are genetically primed for frequent flight. Active flight causes oxidative stress, but butterflies from fragmented landscapes were more tolerant of hypoxia. We conclude that differences in gene expression between the landscape types reflect genomic adaptations to landscape fragmentation.

A faster subquadratic algorithm for finding outlier correlations

We study the problem of detecting outlier pairs of strongly correlated variables among a collection of n variables with otherwise weak pairwise correlations. After normalization, this task amounts to the geometric task where we are given as input a set of n vectors with unit Euclidean norm and dimension d, and for some constants 0<τ < ρ < 1, we are asked to find all the outlier pairs of vectors whose inner product is at least ρ in absolute value, subject to the promise that all but at most q pairs of vectors have inner product at most τ in absolute value. Improving on an algorithm of Valiant [FOCS 2012; J. ACM 2015], we present a randomized algorithm that for Boolean inputs ({ −1,1}-valued data normalized to unit Euclidean length) runs in time Õ((nmax,{ 1−γ +M(Δ γ ,γ),M(1−γ ,2 Δ γ)}+qdn2γ), where 0<γ < 1 is a constant tradeoff parameter and M(μ, ν) is the exponent to multiply an ⌊ nμ ⌋ × ⌊ nν ⌋ matrix with an ⌊ nν ⌋ × ⌊ nμ ⌋ matrix and Δ =1/(1−logτ ρ). As corollaries we obtain randomized algorithms that run in time Õ( (n2/ω 3−logτ ρ + qdn2/(1−logτ ρ)3=logττ ρ) and in time õ( (n4 / 2+α (1−logτ ρ)+qdn2/α (1−logτ ρ)2+α (1−logτ ρ)>), where 2≤ ω <2.38 is the exponent for square matrix multiplication and 0.3<α ≤ 1 is the exponent for rectangular matrix multiplication. The notation Õ(ṡ) hides polylogarithmic factors in n and d whose degree may depend on ρ and τ. We present further corollaries for the light bulb problem and for learning sparse Boolean functions.

A Naive Bayes classifier for protein function prediction.

A Naive Bayes classifier tool is presented for annotating proteins on the basis of amino acid motifs, cellular localization and protein-protein interactions. Annotations take the form of posterior probabilities within the Molecular Function hierarchy of the Gene Ontology (GO). Experiments with the data available for yeast, Saccharomyces cerevisiae, show that our prediction method can yield a relatively high level of accuracy. Several apparent challenges and possibilities for future developments are also discussed. A common approach to functional characterization is to use sequence similarities at varying levels, by utilizing several existing databases and local alignment/identification algorithms. Such an approach is typically quite labor-intensive when performed by an expert in a manual fashion. Integration of several sources of information is in this context generally considered as the only possibility to obtain valuable predictions with practical implications. However, some improvements in the prediction accuracy of the molecular functions, and thereby also savings in the computational effort, can be achieved by restricting attention to only those data sources that involve a higher degree of specificity. We employ here a Naive Bayes model in order to provide probabilistic predictions, and to enable a computationally efficient approach to data integration.

Lessons learned in the challenge: making predictions and scoring them

In this paper we present lessons learned in the Evaluating Predictive Uncertainty Challenge. We describe the methods we used in regression challenges, including our winning method for the Outaouais data set. We then turn our attention to the more general problem of scoring in probabilistic machine learning challenges. It is widely accepted that scoring rules should be proper in the sense that the true generative distribution has the best expected score; we note that while this is useful, it does not guarantee finding the best methods for practical machine learning tasks. We point out some problems in local scoring rules such as the negative logarithm of predictive density (NLPD), and illustrate with examples that many of these problems can be avoided by a distance-sensitive rule such as the continuous ranked probability score (CRPS).

Explicit Correlation Amplifiers for Finding Outlier Correlations in Deterministic Subquadratic Time

We derandomize G. Valiants [J. ACM 62 (2015) Art. 13] subquadratic-time algorithm for finding outlier correlations in binary data. Our derandomized algorithm gives deterministic subquadratic scaling essentially for the same parameter range as Valiants randomized algorithm, but the precise constants we save over quadratic scaling are more modest. Our main technical tool for derandomization is an explicit family of correlation amplifiers built via a family of zigzag-product expanders in Reingold, Vadhan, and Wigderson [Ann. of Math. 155 (2002) 157--187]. We say that a function

Computing exact clustering posteriors with subset convolution

f:\{-1,1\}^d\rightarrow\{-1,1\}^D

Exact Bayesian learning of Partition Directed Acyclic Graphs

is a correlation amplifier with threshold

An Adaptive Prefix-Assignment Technique for Symmetry Reduction

0\leq\tau\leq 1