2011

Projected Newton-type methods in machine learning

Schmidt, M., Kim, D., Sra, S.

In Optimization for Machine Learning, pages: 305-330, (Editors: Sra, S., Nowozin, S. and Wright, S. J.), MIT Press, Cambridge, MA, USA, December 2011 (inbook)

Abstract
We consider projected Newton-type methods for solving large-scale optimization problems arising in machine learning and related fields. We first introduce an algorithmic framework for projected Newton-type methods by reviewing a canonical projected (quasi-)Newton method. This method, while conceptually pleasing, has a high computation cost per iteration. Thus, we discuss two variants that are more scalable, namely, two-metric projection and inexact projection methods. Finally, we show how to apply the Newton-type framework to handle non-smooth objectives. Examples are provided throughout the chapter to illustrate machine learning applications of our framework.

ei

2011

Causal Inference on Discrete Data using Additive Noise Models
IEEE Transactions on Pattern Analysis and Machine Intelligence, 33(12):2436-2450, December 2011 (article)

Abstract
Inferring the causal structure of a set of random variables from a finite sample of the joint distribution is an important problem in science. The case of two random variables is particularly challenging since no (conditional) independences can be exploited. Recent methods that are based on additive noise models suggest the following principle: Whenever the joint distribution {\bf P}^{(X,Y)} admits such a model in one direction, e.g., Y=f(X)+N, N \perp\kern-6pt \perp X, but does not admit the reversed model X=g(Y)+\tilde{N}, \tilde{N} \perp\kern-6pt \perp Y, one infers the former direction to be causal (i.e., X\rightarrow Y). Up to now, these approaches only dealt with continuous variables. In many situations, however, the variables of interest are discrete or even have only finitely many states. In this work, we extend the notion of additive noise models to these cases. We prove that it almost never occurs that additive noise models can be fit in both directions. We further propose an efficient algorithm that is able to perform this way of causal inference on finite samples of discrete variables. We show that the algorithm works on both synthetic and real data sets.

ei

Spontaneous epigenetic variation in the Arabidopsis thaliana methylome

Becker, C., Hagmann, J., Müller, J., Koenig, D., Stegle, O., Borgwardt, K., Weigel, D.

Nature, 480(7376):245-249, December 2011 (article)

Abstract
Heritable epigenetic polymorphisms, such as differential cytosine methylation, can underlie phenotypic variation1, 2. Moreover, wild strains of the plant Arabidopsis thaliana differ in many epialleles3, 4, and these can influence the expression of nearby genes1, 2. However, to understand their role in evolution5, it is imperative to ascertain the emergence rate and stability of epialleles, including those that are not due to structural variation. We have compared genome-wide DNA methylation among 10 A. thaliana lines, derived 30 generations ago from a common ancestor6. Epimutations at individual positions were easily detected, and close to 30,000 cytosines in each strain were differentially methylated. In contrast, larger regions of contiguous methylation were much more stable, and the frequency of changes was in the same low range as that of DNA mutations7. Like individual positions, the same regions were often affected by differential methylation in independent lines, with evidence for recurrent cycles of forward and reverse mutations. Transposable elements and short interfering RNAs have been causally linked to DNA methylation8. In agreement, differentially methylated sites were farther from transposable elements and showed less association with short interfering RNA expression than invariant positions. The biased distribution and frequent reversion of epimutations have important implications for the potential contribution of sequence-independent epialleles to plant evolution.

ei

HHfrag: HMM-based fragment detection using HHpred

Kalev, I., Habeck, M.

Bioinformatics, 27(22):3110-3116, November 2011 (article)

Abstract
Motivation: Over the last decade, both static and dynamic fragment libraries for protein structure prediction have been introduced. The former are built from clusters in either sequence or structure space and aim to extract a universal structural alphabet. The latter are tailored for a particular query protein sequence and aim to provide local structural templates that need to be assembled in order to build the full-length structure. Results: Here, we introduce HHfrag, a dynamic HMM-based fragment search method built on the profile–profile comparison tool HHpred. We show that HHfrag provides advantages over existing fragment assignment methods in that it: (i) improves the precision of the fragments at the expense of a minor loss in sequence coverage; (ii) detects fragments of variable length (6–21 amino acid residues); (iii) allows for gapped fragments and (iv) does not assign fragments to regions where there is no clear sequence conservation. We illustrate the usefulness of fragments detected by HHfrag on targets from most recent CASP.

ei

Reward-Weighted Regression with Sample Reuse for Direct Policy Search in Reinforcement Learning

Hachiya, H., Peters, J., Sugiyama, M.

Neural Computation, 23(11):2798-2832, November 2011 (article)

Abstract
Direct policy search is a promising reinforcement learning framework, in particular for controlling continuous, high-dimensional systems. Policy search often requires a large number of samples for obtaining a stable policy update estimator, and this is prohibitive when the sampling cost is expensive. In this letter, we extend an expectation-maximization-based policy search method so that previously collected samples can be efficiently reused. The usefulness of the proposed method, reward-weighted regression with sample reuse (R), is demonstrated through robot learning experiments.

ei

Model Learning in Robotics: a Survey
Cognitive Processing, 12(4):319-340, November 2011 (article)

Abstract
Models are among the most essential tools in robotics, such as kinematics and dynamics models of the robot's own body and controllable external objects. It is widely believed that intelligent mammals also rely on internal models in order to generate their actions. However, while classical robotics relies on manually generated models that are based on human insights into physics, future autonomous, cognitive robots need to be able to automatically generate models that are based on information which is extracted from the data streams accessible to the robot. In this paper, we survey the progress in model learning with a strong focus on robot control on a kinematic as well as dynamical level. Here, a model describes essential information about the behavior of the environment and the in uence of an agent on this environment. In the context of model based learning control, we view the model from three di fferent perspectives. First, we need to study the di erent possible model learning architectures for robotics. Second, we discuss what kind of problems these architecture and the domain of robotics imply for the applicable learning methods. From this discussion, we deduce future directions of real-time learning algorithms. Third, we show where these scenarios have been used successfully in several case studies.

ei

FaST linear mixed models for genome-wide association studies

Lippert, C., Listgarten, J., Liu, Y., Kadie, CM., Davidson, RI., Heckerman, D.

Nature Methods, 8(10):833–835, October 2011 (article)

Abstract
We describe factored spectrally transformed linear mixed models (FaST-LMM), an algorithm for genome-wide association studies (GWAS) that scales linearly with cohort size in both run time and memory use. On Wellcome Trust data for 15,000 individuals, FaST-LMM ran an order of magnitude faster than current efficient algorithms. Our algorithm can analyze data for 120,000 individuals in just a few hours, whereas current algorithms fail on data for even 20,000 individuals (http://mscompbio.codeplex.com/).

ei

The effect of noise correlations in populations of diversely tuned neurons

Ecker, A., Berens, P., Tolias, A., Bethge, M.

Journal of Neuroscience, 31(40):14272-14283, October 2011 (article)

Abstract
The amount of information encoded by networks of neurons critically depends on the correlation structure of their activity. Neurons with similar stimulus preferences tend to have higher noise correlations than others. In homogeneous populations of neurons, this limited range correlation structure is highly detrimental to the accuracy of a population code. Therefore, reduced spike count correlations under attention, after adaptation, or after learning have been interpreted as evidence for a more efficient population code. Here, we analyze the role of limited range correlations in more realistic, heterogeneous population models. We use Fisher information and maximum-likelihood decoding to show that reduced correlations do not necessarily improve encoding accuracy. In fact, in populations with more than a few hundred neurons, increasing the level of limited range correlations can substantially improve encoding accuracy. We found that this improvement results from a decrease in noise entropy that is associated with increasing correlations if the marginal distributions are unchanged. Surprisingly, for constant noise entropy and in the limit of large populations, the encoding accuracy is independent of both structure and magnitude of noise correlations.

ei

Analysis of Fixed-Point and Coordinate Descent Algorithms for Regularized Kernel Methods
IEEE Transactions on Neural Networks, 22(10):1576-1587, October 2011 (article)

Abstract
In this paper, we analyze the convergence of two general classes of optimization algorithms for regularized kernel methods with convex loss function and quadratic norm regularization. The first methodology is a new class of algorithms based on fixed-point iterations that are well-suited for a parallel implementation and can be used with any convex loss function. The second methodology is based on coordinate descent, and generalizes some techniques previously proposed for linear support vector machines. It exploits the structure of additively separable loss functions to compute solutions of line searches in closed form. The two methodologies are both very easy to implement. In this paper, we also show how to remove non-differentiability of the objective functional by exactly reformulating a convex regularization problem as an unconstrained differentiable stabilization problem.

ei

A biomimetic approach to robot table tennis
Adaptive Behavior , 19(5):359-376 , October 2011 (article)

Abstract
Playing table tennis is a difficult motor task that requires fast movements, accurate control and adaptation to task parameters. Although human beings see and move slower than most robot systems, they significantly outperform all table tennis robots. One important reason for this higher performance is the human movement generation. In this paper, we study human movements during table tennis and present a robot system that mimics human striking behavior. Our focus lies on generating hitting motions capable of adapting to variations in environmental conditions, such as changes in ball speed and position. Therefore, we model the human movements involved in hitting a table tennis ball using discrete movement stages and the virtual hitting point hypothesis. The resulting model was evaluated both in a physically realistic simulation and on a real anthropomorphic seven degrees of freedom Barrett WAM™ robot arm.

ei

Whole-genome sequencing of multiple Arabidopsis thaliana populations

Cao, J., Schneeberger, K., Ossowski, S., Günther, T., Bender, S., Fitz, J., Koenig, D., Lanz, C., Stegle, O., Lippert, C., Wang, X., Ott, F., Müller, J., Alonso-Blanco, C., Borgwardt, K., Schmid, K., Weigel, D.

Nature Genetics, 43(10):956–963, October 2011 (article)

Abstract
The plant Arabidopsis thaliana occurs naturally in many different habitats throughout Eurasia. As a foundation for identifying genetic variation contributing to adaptation to diverse environments, a 1001 Genomes Project to sequence geographically diverse A. thaliana strains has been initiated. Here we present the first phase of this project, based on population-scale sequencing of 80 strains drawn from eight regions throughout the species' native range. We describe the majority of common small-scale polymorphisms as well as many larger insertions and deletions in the A. thaliana pan-genome, their effects on gene function, and the patterns of local and global linkage among these variants. The action of processes other than spontaneous mutation is identified by comparing the spectrum of mutations that have accumulated since A. thaliana diverged from its closest relative 10 million years ago with the spectrum observed in the laboratory. Recent species-wide selective sweeps are rare, and potentially deleterious mutations are more common in marginal populations.

ei

Multiple reference genomes and transcriptomes for Arabidopsis thaliana

Gan, X., Stegle, O., Behr, J., Steffen, J., Drewe, P., Hildebrand, K., Lyngsoe, R., Schultheiss, S., Osborne, E., Sreedharan, V., Kahles, A., Bohnert, R., Jean, G., Derwent, P., Kersey, P., Belfield, E., Harberd, N., Kemen, E., Toomajian, C., Kover, P., Clark, R., Rätsch, G., Mott, R.

Nature, 477(7365):419–423, September 2011 (article)

Abstract
Genetic differences between Arabidopsis thaliana accessions underlie the plant’s extensive phenotypic variation, and until now these have been interpreted largely in the context of the annotated reference accession Col-0. Here we report the sequencing, assembly and annotation of the genomes of 18 natural A. thaliana accessions, and their transcriptomes. When assessed on the basis of the reference annotation, one-third of protein-coding genes are predicted to be disrupted in at least one accession. However, re-annotation of each genome revealed that alternative gene models often restore coding potential. Gene expression in seedlings differed for nearly half of expressed genes and was frequently associated with cis variants within 5 kilobases, as were intron retention alternative splicing events. Sequence and expression variation is most pronounced in genes that respond to the biotic environment. Our data further promote evolutionary and functional studies in A. thaliana, especially the MAGIC genetic reference population descended from these accessions.

ei

Weisfeiler-Lehman Graph Kernels

Shervashidze, N., Schweitzer, P., van Leeuwen, E., Mehlhorn, K., Borgwardt, M.

Journal of Machine Learning Research, 12, pages: 2539-2561, September 2011 (article)

Abstract
In this article, we propose a family of efficient kernels for large graphs with discrete node labels. Key to our method is a rapid feature extraction scheme based on the Weisfeiler-Lehman test of isomorphism on graphs. It maps the original graph to a sequence of graphs, whose node attributes capture topological and label information. A family of kernels can be defined based on this Weisfeiler-Lehman sequence of graphs, including a highly efficient kernel comparing subtree-like patterns. Its runtime scales only linearly in the number of edges of the graphs and the length of the Weisfeiler-Lehman graph sequence. In our experimental evaluation, our kernels outperform state-of-the-art graph kernels on several graph classification benchmark data sets in terms of accuracy and runtime. Our kernels open the door to large-scale applications of graph kernels in various disciplines such as computational biology and social network analysis.

ei

What are the Causes of Performance Variation in Brain-Computer Interfacing?
International Journal of Bioelectromagnetism, 13(3):115-116, September 2011 (article)

Abstract
While research on brain-computer interfacing (BCI) has seen tremendous progress in recent years, performance still varies substantially between as well as within subjects, with roughly 10 - 20% of subjects being incapable of successfully operating a BCI system. In this short report, I argue that this variation in performance constitutes one of the major obstacles that impedes a successful commercialization of BCI systems. I review the current state of research on the neuro-physiological causes of performance variation in BCI, discuss recent progress and open problems, and delineate potential research programs for addressing this issue.

ei

Gravitational Lensing Accuracy Testing 2010 (GREAT10) Challenge Handbook

Kitching, T., Amara, A., Gill, M., Harmeling, S., Heymans, C., Massey, R., Rowe, B., Schrabback, T., Voigt, L., Balan, S., Bernstein, G., Bethge, M., Bridle, S., Courbin, F., Gentile, M., Heavens, A., Hirsch, M., Hosseini, R., Kiessling, A., Kirk, D., Kuijken, K., Mandelbaum, R., Moghaddam, B., Nurbaeva, G., Paulin-Henriksson, S., Rassat, A., Rhodes, J., Schölkopf, B., Shawe-Taylor, J., Shmakova, M., Taylor, A., Velander, M., van Waerbeke, L., Witherick, D., Wittman, D.

Annals of Applied Statistics, 5(3):2231-2263, September 2011 (article)

Abstract
GRavitational lEnsing Accuracy Testing 2010 (GREAT10) is a public image analysis challenge aimed at the development of algorithms to analyze astronomical images. Specifically, the challenge is to measure varying image distortions in the presence of a variable convolution kernel, pixelization and noise. This is the second in a series of challenges set to the astronomy, computer science and statistics communities, providing a structured environment in which methods can be improved and tested in preparation for planned astronomical surveys. GREAT10 extends upon previous work by introducing variable fields into the challenge. The “Galaxy Challenge” involves the precise measurement of galaxy shape distortions, quantified locally by two parameters called shear, in the presence of a known convolution kernel. Crucially, the convolution kernel and the simulated gravitational lensing shape distortion both now vary as a function of position within the images, as is the case for real data. In addition, we introduce the “Star Challenge” that concerns the reconstruction of a variable convolution kernel, similar to that in a typical astronomical observation. This document details the GREAT10 Challenge for potential participants. Continually updated information is also available from www.greatchallenges.info.

ei

MRI-Based Attenuation Correction for Whole-Body PET/MRI: Quantitative Evaluation of Segmentation- and Atlas-Based Methods

Hofmann, M., Bezrukov, I., Mantlik, F., Aschoff, P., Steinke, F., Beyer, T., Pichler, B., Schölkopf, B.

Journal of Nuclear Medicine, 52(9):1392-1399, September 2011 (article)

Abstract
PET/MRI is an emerging dual-modality imaging technology that requires new approaches to PET attenuation correction (AC). We assessed 2 algorithms for whole-body MRI-based AC (MRAC): a basic MR image segmentation algorithm and a method based on atlas registration and pattern recognition (AT&PR). METHODS: Eleven patients each underwent a whole-body PET/CT study and a separate multibed whole-body MRI study. The MR image segmentation algorithm uses a combination of image thresholds, Dixon fat-water segmentation, and component analysis to detect the lungs. MR images are segmented into 5 tissue classes (not including bone), and each class is assigned a default linear attenuation value. The AT&PR algorithm uses a database of previously aligned pairs of MRI/CT image volumes. For each patient, these pairs are registered to the patient MRI volume, and machine-learning techniques are used to predict attenuation values on a continuous scale. MRAC methods are compared via the quantitative analysis of AC PET images using volumes of interest in normal organs and on lesions. We assume the PET/CT values after CT-based AC to be the reference standard. RESULTS: In regions of normal physiologic uptake, the average error of the mean standardized uptake value was 14.1% ± 10.2% and 7.7% ± 8.4% for the segmentation and the AT&PR methods, respectively. Lesion-based errors were 7.5% ± 7.9% for the segmentation method and 5.7% ± 4.7% for the AT&PR method. CONCLUSION: The MRAC method using AT&PR provided better overall PET quantification accuracy than the basic MR image segmentation approach. This better quantification was due to the significantly reduced volume of errors made regarding volumes of interest within or near bones and the slightly reduced volume of errors made regarding areas outside the lungs.

ei

Semi-supervised kernel canonical correlation analysis with application to human fMRI

Blaschko, M., Shelton, J., Bartels, A., Lampert, C., Gretton, A.

Pattern Recognition Letters, 32(11):1572-1583 , August 2011 (article)

Abstract
Kernel canonical correlation analysis (KCCA) is a general technique for subspace learning that incorporates principal components analysis (PCA) and Fisher linear discriminant analysis (LDA) as special cases. By finding directions that maximize correlation, KCCA learns representations that are more closely tied to the underlying process that generates the data and can ignore high-variance noise directions. However, for data where acquisition in one or more modalities is expensive or otherwise limited, KCCA may suffer from small sample effects. We propose to use semi-supervised Laplacian regularization to utilize data that are present in only one modality. This approach is able to find highly correlated directions that also lie along the data manifold, resulting in a more robust estimate of correlated subspaces. Functional magnetic resonance imaging (fMRI) acquired data are naturally amenable to subspace techniques as data are well aligned. fMRI data of the human brain are a particularly interesting candidate. In this study we implemented various supervised and semi-supervised versions of KCCA on human fMRI data, with regression to single and multi-variate labels (corresponding to video content subjects viewed during the image acquisition). In each variate condition, the semi-supervised variants of KCCA performed better than the supervised variants, including a supervised variant with Laplacian regularization. We additionally analyze the weights learned by the regression in order to infer brain regions that are important to different types of visual processing.

ei

Multi-subject learning for common spatial patterns in motor-imagery BCI

Devlaminck, D., Wyns, B., Grosse-Wentrup, M., Otte, G., Santens, P.

Computational Intelligence and Neuroscience, 2011(217987):1-9, August 2011 (article)

Abstract
Motor-imagery-based brain-computer interfaces (BCIs) commonly use the common spatial pattern filter (CSP) as preprocessing step before feature extraction and classification. The CSP method is a supervised algorithm and therefore needs subject-specific training data for calibration, which is very time consuming to collect. In order to reduce the amount of calibration data that is needed for a new subject, one can apply multitask (from now on called multisubject) machine learning techniques to the preprocessing phase. Here, the goal of multisubject learning is to learn a spatial filter for a new subject based on its own data and that of other subjects. This paper outlines the details of the multitask CSP algorithm and shows results on two data sets. In certain subjects a clear improvement can be seen, especially when the number of training trials is relatively low.

ei

ccSVM: correcting Support Vector Machines for confounding factors in biological data classification

Li, L., Rakitsch, B., Borgwardt, K.

Bioinformatics, 27(13: ISMB/ECCB 2011):i342-i348, July 2011 (article)

Abstract
Motivation: Classifying biological data into different groups is a central task of bioinformatics: for instance, to predict the function of a gene or protein, the disease state of a patient or the phenotype of an individual based on its genotype. Support Vector Machines are a wide spread approach for classifying biological data, due to their high accuracy, their ability to deal with structured data such as strings, and the ease to integrate various types of data. However, it is unclear how to correct for confounding factors such as population structure, age or gender or experimental conditions in Support Vector Machine classification. Results: In this article, we present a Support Vector Machine classifier that can correct the prediction for observed confounding factors. This is achieved by minimizing the statistical dependence between the classifier and the confounding factors. We prove that this formulation can be transformed into a standard Support Vector Machine with rescaled input data. In our experiments, our confounder correcting SVM (ccSVM) improves tumor diagnosis based on samples from different labs, tuberculosis diagnosis in patients of varying age, ethnicity and gender, and phenotype prediction in the presence of population structure and outperforms state-of-the-art methods in terms of prediction accuracy.

ei

Policy Search for Motor Primitives in Robotics
Machine Learning, 84(1-2):171-203, July 2011 (article)

Abstract
Many motor skills in humanoid robotics can be learned using parametrized motor primitives. While successful applications to date have been achieved with imitation learning, most of the interesting motor learning problems are high-dimensional reinforcement learning problems. These problems are often beyond the reach of current reinforcement learning methods. In this paper, we study parametrized policy search methods and apply these to benchmark problems of motor primitive learning in robotics. We show that many well-known parametrized policy search methods can be derived from a general, common framework. This framework yields both policy gradient methods and expectation-maximization (EM) inspired algorithms. We introduce a novel EM-inspired algorithm for policy learning that is particularly well-suited for dynamical system motor primitives. We compare this algorithm, both in simulation and on a real robot, to several well-known parametrized policy search methods such as episodic REINFORCE, ‘Vanilla’ Policy Gradients with optimal baselines, episodic Natural Actor Critic, and episodic Reward-Weighted Regression. We show that the proposed method out-performs them on an empirical benchmark of learning dynamical system motor primitives both in simulation and on a real robot. We apply it in the context of motor learning and show that it can learn a complex Ball-in-a-Cup task on a real Barrett WAM™ robot arm.

ei

Epistasis detection on quantitative phenotypes by exhaustive enumeration using GPUs

Kam-Thong, T., Pütz, B., Karbalai, N., Müller-Myhsok, B., Borgwardt, K.

Bioinformatics, 27(13: ISMB/ECCB 2011):i214-i221, July 2011 (article)

Abstract
Motivation: In recent years, numerous genome-wide association studies have been conducted to identify genetic makeup that explains phenotypic differences observed in human population. Analytical tests on single loci are readily available and embedded in common genome analysis software toolset. The search for significant epistasis (gene–gene interactions) still poses as a computational challenge for modern day computing systems, due to the large number of hypotheses that have to be tested. Results: In this article, we present an approach to epistasis detection by exhaustive testing of all possible SNP pairs. The search strategy based on the Hilbert–Schmidt Independence Criterion can help delineate various forms of statistical dependence between the genetic markers and the phenotype. The actual implementation of this search is done on the highly parallelized architecture available on graphics processing units rendering the completion of the full search feasible within a day.

ei

Empirical Inference
International Journal of Materials Research, 2011(7):809-814, July 2011 (article)

Abstract
Empirical Inference is the process of drawing conclusions from observational data. For instance, the data can be measurements from an experiment, which are used by a researcher to infer a scientific law. Another kind of empirical inference is performed by living beings, continuously recording data from their environment and carrying out appropriate actions. Do these problems have anything in common, and are there underlying principles governing the extraction of regularities from data? What characterizes hard inference problems, and how can we solve them? Such questions are studied by a community of scientists from various fields, engaged in machine learning research. This short paper, which is based on the author’s lecture to the scientific council of the Max Planck Society in February 2010, will attempt to describe some of the main ideas and problems of machine learning. It will provide illustrative examples of real world machine learning applications, including the use of machine learning towards the design of intelligent systems.

ei

Online Multi-frame Blind Deconvolution with Super-resolution and Saturation Correction
Astronomy & Astrophysics, 531(A9):11, July 2011 (article)

Abstract
Astronomical images taken by ground-based telescopes suffer degradation due to atmospheric turbulence. This degradation can be tackled by costly hardware-based approaches such as adaptive optics, or by sophisticated software-based methods such as lucky imaging, speckle imaging, or multi-frame deconvolution. Software-based methods process a sequence of images to reconstruct a deblurred high-quality image. However, existing approaches are limited in one or several aspects: (i) they process all images in batch mode, which for thousands of images is prohibitive; (ii) they do not reconstruct a super-resolved image, even though an image sequence often contains enough information; (iii) they are unable to deal with saturated pixels; and (iv) they are usually non-blind, i.e., they assume the blur kernels to be known. In this paper we present a new method for multi-frame deconvolution called online blind deconvolution (OBD) that overcomes all these limitations simultaneously. Encouraging results on simulated and real astronomical images demonstrate that OBD yields deblurred images of comparable and often better quality than existing approaches.

ei

Closing the sensorimotor loop: haptic feedback facilitates decoding of motor imagery

Gomez Rodriguez, M., Peters, J., Hill, J., Schölkopf, B., Gharabaghi, A., Grosse-Wentrup, M.

Journal of Neural Engineering, 8(3):1-12, June 2011 (article)

Abstract
The combination of brain–computer interfaces (BCIs) with robot-assisted physical therapy constitutes a promising approach to neurorehabilitation of patients with severe hemiparetic syndromes caused by cerebrovascular brain damage (e.g. stroke) and other neurological conditions. In such a scenario, a key aspect is how to reestablish the disrupted sensorimotor feedback loop. However, to date it is an open question how artificially closing the sensorimotor feedback loop influences the decoding performance of a BCI. In this paper, we answer this issue by studying six healthy subjects and two stroke patients. We present empirical evidence that haptic feedback, provided by a seven degrees of freedom robotic arm, facilitates online decoding of arm movement intention. The results support the feasibility of future rehabilitative treatments based on the combination of robot-assisted physical therapy with BCIs.

ei

Greedy Learning of Binary Latent Trees

Harmeling, S., Williams, C.

IEEE Transactions on Pattern Analysis and Machine Intelligence, 33(6):1087-1097, June 2011 (article)

Abstract
Inferring latent structures from observations helps to model and possibly also understand underlying data generating processes. A rich class of latent structures are hierarchical latent class (HLC) models. Zhang (2004) proposed a search algorithm for learning such models that can find good solutions but is often computationally expensive. As an alternative we investigate two greedy procedures: the BIN-G algorithm determines both the structure of the tree and the cardinality of the latent variables in a bottom-up fashion. The BIN-A algorithm first determines the tree structure using agglomerative hierarchical clustering, and then determines the cardinality of the latent variables as for BIN-G. We show that even with restricting ourselves to binary trees we obtain HLC models of comparable quality to Zhang‘s solutions, while being faster to compute. This claim is validated by a comprehensive comparison on several datasets. Furthermore, we demonstrate that our methods are able to estimate int erpretable latent structures on real-world data with a large number of variables. By applying our method to a restricted version of the 20 newsgroups data these models turn out to be related to topic models, and on data from the PASCAL Visual Object Classes (VOC) 2007 challenge we show how such tree-structured models help us understand how objects co-occur in images.

ei

Learning Dynamic Tactile Sensing with Robust Vision-based Training

Kroemer, O., Lampert, C., Peters, J.

IEEE Transactions on Robotics, 27(3):545-557 , June 2011 (article)

Abstract
Dynamic tactile sensing is a fundamental ability to recognize materials and objects. However, while humans are born with partially developed dynamic tactile sensing and quickly master this skill, today's robots remain in their infancy. The development of such a sense requires not only better sensors but the right algorithms to deal with these sensors' data as well. For example, when classifying a material based on touch, the data are noisy, high-dimensional, and contain irrelevant signals as well as essential ones. Few classification methods from machine learning can deal with such problems. In this paper, we propose an efficient approach to infer suitable lower dimensional representations of the tactile data. In order to classify materials based on only the sense of touch, these representations are autonomously discovered using visual information of the surfaces during training. However, accurately pairing vision and tactile samples in real-robot applications is a difficult problem. The proposed approach, therefore, works with weak pairings between the modalities. Experiments show that the resulting approach is very robust and yields significantly higher classification performance based on only dynamic tactile sensing.

ei

Algebraic polynomials and moments of stochastic integrals
Statistics & Probability Letters, 81(6):627-631, June 2011 (article)

Abstract
We propose an algebraic method for proving estimates on moments of stochastic integrals. The method uses qualitative properties of roots of algebraic polynomials from certain general classes. As an application, we give a new proof of a variation of the Burkholder–Davis–Gundy inequality for the case of stochastic integrals with respect to real locally square integrable martingales. Further possible applications and extensions of the method are outlined.

ei

Inference for psychometric functions in the presence of nonstationary behavior

Fründ, I., Haenel, N., Wichmann, F.

Journal of Vision, 11(6):1-19, May 2011 (article)

Abstract
Measuring sensitivity is at the heart of psychophysics. Often, sensitivity is derived from estimates of the psychometric function. This function relates response probability to stimulus intensity. In estimating these response probabilities, most studies assume stationary observers: Responses are expected to be dependent only on the intensity of a presented stimulus and not on other factors such as stimulus sequence, duration of the experiment, or the responses on previous trials. Unfortunately, a number of factors such as learning, fatigue, or fluctuations in attention and motivation will typically result in violations of this assumption. The severity of these violations is yet unknown. We use Monte Carlo simulations to show that violations of these assumptions can result in underestimation of confidence intervals for parameters of the psychometric function. Even worse, collecting more trials does not eliminate this misestimation of confidence intervals. We present a simple adjustment of the confidence intervals that corrects for the underestimation almost independently of the number of trials and the particular type of violation.

ei

Transition from the locked in to the completely locked-in state: A physiological analysis

Ramos Murguialday, A., Hill, J., Bensch, M., Martens, S., Halder, S., Nijboer, F., Schölkopf, B., Birbaumer, N., Gharabaghi, A.

Clinical Neurophysiology, 122(5):925-933 , May 2011 (article)

Abstract
Objective To clarify the physiological and behavioral boundaries between locked-in (LIS) and the completely locked-in state (CLIS) (no voluntary eye movements, no communication possible) through electrophysiological data and to secure brain–computer-interface (BCI) communication. Methods Electromyography from facial muscles, external anal sphincter (EAS), electrooculography and electrocorticographic data during different psychophysiological tests were acquired to define electrophysiological differences in an amyotrophic lateral sclerosis (ALS) patient with an intracranially implanted grid of 112 electrodes for nine months while the patient passed from the LIS to the CLIS. Results At the very end of the LIS there was no facial muscle activity, nor external anal sphincter but eye control. Eye movements were slow and lasted for short periods only. During CLIS event related brain potentials (ERP) to passive limb movements and auditory stimuli were recorded, vibrotactile stimulation of different body parts resulted in no ERP response. Conclusions The results presented contradict the commonly accepted assumption that the EAS is the last remaining muscle under voluntary control and demonstrate complete loss of eye movements in CLIS. The eye muscle was shown to be the last muscle group under voluntary control. The findings suggest ALS as a multisystem disorder, even affecting afferent sensory pathways. Significance Auditory and proprioceptive brain–computer-interface (BCI) systems are the only remaining communication channels in CLIS.

ei

Incremental online sparsification for model learning in real-time robot control
Neurocomputing, 74(11):1859-1867, May 2011 (article)

Abstract
For many applications such as compliant, accurate robot tracking control, dynamics models learned from data can help to achieve both compliant control performance as well as high tracking quality. Online learning of these dynamics models allows the robot controller to adapt itself to changes in the dynamics (e.g., due to time-variant nonlinearities or unforeseen loads). However, online learning in real-time applications -- as required in control -- cannot be realized by straightforward usage of off-the-shelf machine learning methods such as Gaussian process regression or support vector regression. In this paper, we propose a framework for online, incremental sparsification with a fixed budget designed for fast real-time model learning. The proposed approach employs a sparsification method based on an independence measure. In combination with an incremental learning approach such as incremental Gaussian process regression, we obtain a model approximation method which is applicable in real-time online learning. It exhibits competitive learning accuracy when compared with standard regression techniques. Implementation on a real Barrett WAM robot demonstrates the applicability of the approach in real-time online model learning for real world systems.

ei

Causal Influence of Gamma Oscillations on the Sensorimotor Rhythm
NeuroImage, 56(2):837-842, May 2011 (article)

Abstract
Gamma oscillations of the electromagnetic field of the brain are known to be involved in a variety of cognitive processes, and are believed to be fundamental for information processing within the brain. While gamma oscillations have been shown to be correlated with brain rhythms at different frequencies, to date no empirical evidence has been presented that supports a causal influence of gamma oscillations on other brain rhythms. In this work, we study the relation of gamma oscillations and the sensorimotor rhythm (SMR) in healthy human subjects using electroencephalography. We first demonstrate that modulation of the SMR, induced by motor imagery of either the left or right hand, is positively correlated with the power of frontal and occipital gamma oscillations, and negatively correlated with the power of centro-parietal gamma oscillations. We then demonstrate that the most simple causal structure, capable of explaining the observed correlation of gamma oscillations and the SMR, entails a causal influence of gamma oscillations on the SMR. This finding supports the fundamental role attributed to gamma oscillations for information processing within the brain, and is of particular importance for brain–computer interfaces (BCIs). As modulation of the SMR is typically used in BCIs to infer a subject's intention, our findings entail that gamma oscillations have a causal influence on a subject's capability to utilize a BCI for means of communication.

ei

Statistical Learning Theory: Models, Concepts, and Results
In Handbook of the History of Logic, Vol. 10: Inductive Logic, 10, pages: 651-706, (Editors: Gabbay, D. M., Hartmann, S. and Woods, J. H.), Elsevier North Holland, Amsterdam, Netherlands, May 2011 (inbook)

Abstract
Statistical learning theory provides the theoretical basis for many of today's machine learning algorithms and is arguably one of the most beautifully developed branches of artificial intelligence in general. It originated in Russia in the 1960s and gained wide popularity in the 1990s following the development of the so-called Support Vector Machine (SVM), which has become a standard tool for pattern recognition in a variety of domains ranging from computer vision to computational biology. Providing the basis of new learning algorithms, however, was not the only motivation for developing statistical learning theory. It was just as much a philosophical one, attempting to answer the question of what it is that allows us to draw valid conclusions from empirical data. In this article we attempt to give a gentle, non-technical overview over the key ideas and insights of statistical learning theory. We do not assume that the reader has a deep background in mathematics, statistics, or computer science. Given the nature of the subject matter, however, some familiarity with mathematical concepts and notations and some intuitive understanding of basic probability is required. There exist many excellent references to more technical surveys of the mathematics of statistical learning theory: the monographs by one of the founders of statistical learning theory ([Vapnik, 1995], [Vapnik, 1998]), a brief overview over statistical learning theory in Section 5 of [Sch{\"o}lkopf and Smola, 2002], more technical overview papers such as [Bousquet et al., 2003], [Mendelson, 2003], [Boucheron et al., 2005], [Herbrich and Williamson, 2002], and the monograph [Devroye et al., 1996].

ei

The effect of patient positioning aids on PET quantification in PET/MR imaging

Mantlik, F., Hofmann, M., Werner, M., Sauter, A., Kupferschläger, J., Schölkopf, B., Pichler, B., Beyer, T.

European Journal of Nuclear Medicine and Molecular Imaging, 38(5):920-929, May 2011 (article)

Abstract
Objectives Clinical PET/MR requires the use of patient positioning aids to immobilize and support patients for the duration of the combined examination. Ancillary immobilization devices contribute to overall attenuation of the PET signal, but are not detected with conventional MR sequences and, hence, are ignored in standard MR-based attenuation correction (MR-AC). We report on the quantitative effect of not accounting for the attenuation of patient positioning aids in combined PET/MR imaging. Methods We used phantom and patient data acquired with positioning aids on a PET/CT scanner (Biograph 16, HI-REZ) to mimic PET/MR imaging conditions. Reference CT-based attenuation maps were generated from measured (original) CT transmission images (origCT-AC). We also created MR-like attenuation maps by following the same conversion procedure of the attenuation values except for the prior delineation and subtraction of the positioning aids from the CT images (modCT-AC). First, a uniform 68Ge cylinder was positioned centrally in the PET/CT scanner and fixed with a vacuum mattress (10 cm thick) and, in a repeat examination, with MR positioning foam pads. Second, 16 patient datasets were selected for subsequent processing. All patients were regionally immobilized with positioning aids: a vacuum mattress for head/neck imaging (nine patients) and a foam mattress for imaging of the lower extremities (seven patients). PET images were reconstructed following CT-based attenuation and scatter correction using the original and modified (MR-like) CT images: PETorigCT-AC and PETmodCT-AC, respectively. PET images following origCT-AC and modCT-AC were compared visually and in terms of mean differences of voxels with a standardized uptake value of at least 1.0. In addition, we report maximum activity concentration in lesions for selected patients. Results In the phantom study employing the vacuum mattress the average voxel activity in PETmodCT-AC was underestimated by 6.4% compared to PETorigCT-AC, with 3.4% of the PET voxels being underestimated by 10% or more. When the MR foam pads were not accounted for during AC, PETmodCT-AC was underestimated by 1.1% on average, with none of the PET voxels being underestimated by 10% or more. Evaluation of the head/neck patient data showed a decrease of 8.4% ([68Ga]DOTATOC) and 7.4% ([18F]FDG) when patient positioning aids were not accounted for during AC, while the corresponding decrease was insignificant for the lower extremities. Conclusion Depending on the size and density of the positioning aids used, a regionally variable underestimation of PET activity following AC is observed when positioning aids are not accounted for. This underestimation may become relevant in combined PET/MR imaging of patients with neuropsychiatric indications, but appears to be of no clinical relevance in imaging the extremities.

ei

Loose-limbed People: Estimating 3D Human Pose and Motion Using Non-parametric Belief Propagation

Sigal, L., Isard, M., Haussecker, H., Black, M. J.

International Journal of Computer Vision, 98(1):15-48, Springer Netherlands, May 2011 (article)

Abstract
We formulate the problem of 3D human pose estimation and tracking as one of inference in a graphical model. Unlike traditional kinematic tree representations, our model of the body is a collection of loosely-connected body-parts. In particular, we model the body using an undirected graphical model in which nodes correspond to parts and edges to kinematic, penetration, and temporal constraints imposed by the joints and the world. These constraints are encoded using pair-wise statistical distributions, that are learned from motion-capture training data. Human pose and motion estimation is formulated as inference in this graphical model and is solved using Particle Message Passing (PaMPas). PaMPas is a form of non-parametric belief propagation that uses a variation of particle filtering that can be applied over a general graphical model with loops. The loose-limbed model and decentralized graph structure allow us to incorporate information from "bottom-up" visual cues, such as limb and head detectors, into the inference process. These detectors enable automatic initialization and aid recovery from transient tracking failures. We illustrate the method by automatically tracking people in multi-view imagery using a set of calibrated cameras and present quantitative evaluation using the HumanEva dataset.

ps

Improving quantification of functional networks with EEG inverse problem: Evidence from a decoding point of view

Besserve, M., Martinerie, J., Garnero, L.

NeuroImage, 55(4):1536-1547, April 2011 (article)

Abstract
Decoding experimental conditions from single trial Electroencephalographic (EEG) signals is becoming a major challenge for the study of brain function and real-time applications such as Brain Computer Interface. EEG source reconstruction offers principled ways to estimate the cortical activities from EEG signals. But to what extent it can enhance informative brain signals in single trial has not been addressed in a general setting. We tested this using the minimum norm estimate solution (MNE) to estimate spectral power and coherence features at the cortical level. With a fast implementation, we computed a support vector machine (SVM) classifier output from these quantities in real-time, without prior on the relevant functional networks. We applied this approach to single trial decoding of ongoing mental imagery tasks using EEG data recorded in 5 subjects. Our results show that reconstructing the underlying cortical network dynamics significantly outperforms a usual electrode level approach in terms of information transfer and also reduces redundancy between coherence and power features, supporting a decrease of volume conduction effects. Additionally, the classifier coefficients reflect the most informative features of network activity, showing an important contribution of localized motor and sensory brain areas, and of coherence between areas up to 6 cm distance. This study provides a computationally efficient and interpretable strategy to extract information from functional networks at the cortical level in single trial. Moreover, this sets a general framework to evaluate the performance of EEG source reconstruction methods by their decoding abilities.

ei

Using brain–computer interfaces to induce neural plasticity and restore function

Grosse-Wentrup, M., Mattia, D., Oweiss, K.

Journal of Neural Engineering, 8(2):1-5, April 2011 (article)

Abstract
Analyzing neural signals and providing feedback in real-time is one of the core characteristics of a brain-computer interface (BCI). As this feature may be employed to induce neural plasticity, utilizing BCI-technology for therapeutic purposes is increasingly gaining popularity in the BCI-community. In this review, we discuss the state-of-the-art of research on this topic, address the principles of and challenges in inducing neural plasticity by means of a BCI, and delineate the problems of study design and outcome evaluation arising in this context. The review concludes with a list of open questions and recommendations for future research in this field.

ei

EPIBLASTER-fast exhaustive two-locus epistasis detection strategy using graphical processing units

Kam-Thong, T., Czamara, D., Tsuda, K., Borgwardt, K., Lewis, C., Erhardt-Lehmann, A., Hemmer, B., Rieckmann, P., Daake, M., Weber, F., Wolf, C., Ziegler, A., Pütz, B., Holsboer, F., Schölkopf, B., Müller-Myhsok, B.

European Journal of Human Genetics, 19(4):465-471, April 2011 (article)

Abstract
Detection of epistatic interaction between loci has been postulated to provide a more in-depth understanding of the complex biological and biochemical pathways underlying human diseases. Studying the interaction between two loci is the natural progression following traditional and well-established single locus analysis. However, the added costs and time duration required for the computation involved have thus far deterred researchers from pursuing a genome-wide analysis of epistasis. In this paper, we propose a method allowing such analysis to be conducted very rapidly. The method, dubbed EPIBLASTER, is applicable to casecontrol studies and consists of a two-step process in which the difference in Pearson‘s correlation coefficients is computed between controls and cases across all possible SNP pairs as an indication of significant interaction warranting further analysis. For the subset of interactions deemed potentially significant, a second-stage analysis is performed using the likelihood ratio test from the logistic regression to obtain the P-value for the estimated coefficients of the individual effects and the interaction term. The algorithm is implemented using the parallel computational capability of commercially available graphical processing units to greatly reduce the computation time involved. In the current setup and example data sets (211 cases, 222 controls, 299468 SNPs; and 601 cases, 825 controls, 291095 SNPs), this coefficient evaluation stage can be completed in roughly 1 day. Our method allows for exhaustive and rapid detection of significant SNP pair interactions without imposing significant marginal effects of the single loci involved in the pair.

ei

Model learning for robot control: a survey
Cognitive Processing, 12(4):319-340, April 2011 (article)

Abstract
Models are among the most essential tools in robotics, such as kinematics and dynamics models of the robot’s own body and controllable external objects. It is widely believed that intelligent mammals also rely on internal models in order to generate their actions. However, while classical robotics relies on manually generated models that are based on human insights into physics, future autonomous, cognitive robots need to be able to automatically generate models that are based on information which is extracted from the data streams accessible to the robot. In this paper, we survey the progress in model learning with a strong focus on robot control on a kinematic as well as dynamical level. Here, a model describes essential information about the behavior of the environment and the influence of an agent on this environment. In the context of model-based learning control, we view the model from three different perspectives. First, we need to study the different possible model learning architectures for robotics. Second, we discuss what kind of problems these architecture and the domain of robotics imply for the applicable learning methods. From this discussion, we deduce future directions of real-time learning algorithms. Third, we show where these scenarios have been used successfully in several case studies.

ei

Critical issues in state-of-the-art brain–computer interface signal processing

Krusienski, D., Grosse-Wentrup, M., Galan, F., Coyle, D., Miller, K., Forney, E., Anderson, C.

Journal of Neural Engineering, 8(2):1-8, April 2011 (article)

Abstract
This paper reviews several critical issues facing signal processing for brain–computer interfaces (BCIs) and suggests several recent approaches that should be further examined. The topics were selected based on discussions held during the 4th International BCI Meeting at a workshop organized to review and evaluate the current state of, and issues relevant to, feature extraction and translation of field potentials for BCIs. The topics presented in this paper include the relationship between electroencephalography and electrocorticography, novel features for performance prediction, time-embedded signal representations, phase information, signal non-stationarity, and unsupervised adaptation.

ei

Point-and-Click Cursor Control With an Intracortical Neural Interface System by Humans With Tetraplegia

Kim, S., Simeral, J. D., Hochberg, L. R., Donoghue, J. P., Friehs, G. M., Black, M. J.

IEEE Transactions on Neural Systems and Rehabilitation Engineering, 19(2):193-203, April 2011 (article)

Abstract
We present a point-and-click intracortical neural interface system (NIS) that enables humans with tetraplegia to volitionally move a 2D computer cursor in any desired direction on a computer screen, hold it still and click on the area of interest. This direct brain-computer interface extracts both discrete (click) and continuous (cursor velocity) signals from a single small population of neurons in human motor cortex. A key component of this system is a multi-state probabilistic decoding algorithm that simultaneously decodes neural spiking activity and outputs either a click signal or the velocity of the cursor. The algorithm combines a linear classifier, which determines whether the user is intending to click or move the cursor, with a Kalman filter that translates the neural population activity into cursor velocity. We present a paradigm for training the multi-state decoding algorithm using neural activity observed during imagined actions. Two human participants with tetraplegia (paralysis of the four limbs) performed a closed-loop radial target acquisition task using the point-and-click NIS over multiple sessions. We quantified point-and-click performance using various human-computer interaction measurements for pointing devices. We found that participants were able to control the cursor motion accurately and click on specified targets with a small error rate (< 3% in one participant). This study suggests that signals from a small ensemble of motor cortical neurons (~40) can be used for natural point-and-click 2D cursor control of a personal computer.

ps

A Blind Deconvolution Approach for Improving the Resolution of Cryo-EM Density Maps

Hirsch, M., Schölkopf, B., Habeck, M.

Journal of Computational Biology, 18(3):335-346, March 2011 (article)

Abstract
Cryo-electron microscopy (cryo-EM) plays an increasingly prominent role in structure elucidation of macromolecular assemblies. Advances in experimental instrumentation and computational power have spawned numerous cryo-EM studies of large biomolecular complexes resulting in the reconstruction of three-dimensional density maps at intermediate and low resolution. In this resolution range, identification and interpretation of structural elements and modeling of biomolecular structure with atomic detail becomes problematic. In this article, we present a novel algorithm that enhances the resolution of intermediate- and low-resolution density maps. Our underlying assumption is to model the low-resolution density map as a blurred and possibly noise-corrupted version of an unknown high-resolution map that we seek to recover by deconvolution. By exploiting the nonnegativity of both the high-resolution map and blur kernel, we derive multiplicative updates reminiscent of those used in nonnegative matrix factorization. Our framework allows for easy incorporation of additional prior knowledge such as smoothness and sparseness, on both the sharpened density map and the blur kernel. A probabilistic formulation enables us to derive updates for the hyperparameters; therefore, our approach has no parameter that needs adjustment. We apply the algorithm to simulated three-dimensional electron microscopic data. We show that our method provides better resolved density maps when compared with B-factor sharpening, especially in the presence of noise. Moreover, our method can use additional information provided by homologous structures, which helps to improve the resolution even further.

ei

Dynamics of excitable neural networks with heterogeneous connectivity

Chavez, M., Besserve, M., Le Van Quyen, M.

Progress in Biophysics and Molecular Biology, 105(1-2):29-33, March 2011 (article)

Abstract
A central issue of neuroscience is to understand how neural units integrates internal and external signals to create coherent states. Recently, it has been shown that the sensitivity and dynamic range of neural assemblies are optimal at a critical coupling among its elements. Complex architectures of connections seem to play a constructive role on the reliable coordination of neural units. Here we show that, the synchronizability and sensitivity of excitable neural networks can be tuned by diversity in the connections strengths. We illustrate our findings for weighted networks with regular, random and complex topologies. Additional comparisons of real brain networks support previous studies suggesting that heterogeneity in the connectivity may play a constructive role on information processing. These findings provide insights into the relationship between structure and function of neural circuits.

ei

Combining computational modeling with sparse and low-resolution data

Habeck, M., Nilges, M.

Journal of Structural Biology, 173(3):419, March 2011 (article)

Abstract
Structural biology is moving into a new era by shifting its focus from static structures of single proteins and protein domains to large and often fragile multi-component complexes. Over the past decade, structural genomics initiatives aimed to fill the voids in fold space and to provide a census of all protein structures. Completion of such an atlas of protein structures is still ongoing, but not sufficient for a mechanistic understanding of how living cells function. One of the great challenges is to bridge the gap between atomic resolution detail and the more fuzzy description of the molecular complexes that govern cellular processes or host–pathogen interactions. We want to move from cartoon-like representations of multi-component complexes to atomic resolution structures. To characterize the structures of the increasingly large and often flexible complexes, high resolution structure determination (as was possible for example for the ribosome) will likely stay the exception. Rather, data from many different methods providing information on the shape (X-ray crystallography, electron microscopy, SAXS, AFM, etc.) or on contacts between components (mass spectrometry, co-purification, or spectroscopic methods) need to be integrated with prior structural knowledge to build a consistent model of the complex. A particular difficulty is that the ratio between the number of conformational degrees of freedom and the number of measurements becomes unfavorable as we work with large complexes: data become increasingly sparse. Structural characterization of large molecular assemblies often involves a loss in resolution as well as in number and quality of data. We are good at solving structures of single proteins, but classical high-resolution structure determination by X-ray crystallography and NMR spectroscopy is often facing its limits as we move to higher molecular mass and increased flexibility. Therefore, structural studies on large complexes rely on new experimental techniques that complement the classical high resolution methods. But also computational approaches are becoming more important when it comes to integrating and analyzing structural information of often heterogeneous nature. Cryoelectron microscopy may serve as an example of how experimental methods can benefit from computation. Low-resolution data from cryo-EM show their true power when combined with modeling and bioinformatics methods such rigid docking and secondary structure hunting. Even in high resolution structure determination, molecular modeling is always necessary to calculate structures from data, to complement the missing information and to evaluate and score the obtained structures. With sparse data, all these three aspects become increasingly difficult, and the quality of the modeling approach becomes more important. With data alone, algorithms may not converge any more; scoring against data becomes meaningless; and the potential energy function becomes central not only as a help in making algorithms converge but also to score and evaluate the structures. In addition to the sparsity of the data, hybrid approaches bring the additional difficulty that the different sources of data may have rather different quality, and may be in the extreme case incompatible with each other. In addition to scoring the structures, modeling should also score in some way the data going into the calculation. This special issue brings together some of the numerous efforts to solve the problems that come from sparsity of data and from integrating data from different sources in hybrid approaches. The methods range from predominantly force-field based to mostly data based. Systems of very different sizes, ranging from single domains to multi-component complexes, are treated. We hope that you will enjoy reading the issue and find it a useful and inspiring resource.

ei

Batch-Mode Active-Learning Methods for the Interactive Classification of Remote Sensing Images

Demir, B., Persello, C., Bruzzone, L.

IEEE Transactions on Geoscience and Remote Sensing, 49(3):1014-1031, March 2011 (article)

Abstract
This paper investigates different batch-mode active-learning (AL) techniques for the classification of remote sensing (RS) images with support vector machines. This is done by generalizing to multiclass problem techniques defined for binary classifiers. The investigated techniques exploit different query functions, which are based on the evaluation of two criteria: uncertainty and diversity. The uncertainty criterion is associated to the confidence of the supervised algorithm in correctly classifying the considered sample, while the diversity criterion aims at selecting a set of unlabeled samples that are as more diverse (distant one another) as possible, thus reducing the redundancy among the selected samples. The combination of the two criteria results in the selection of the potentially most informative set of samples at each iteration of the AL process. Moreover, we propose a novel query function that is based on a kernel-clustering technique for assessing the diversity of samples and a new strategy for selecting the most informative representative sample from each cluster. The investigated and proposed techniques are theoretically and experimentally compared with state-of-the-art methods adopted for RS applications. This is accomplished by considering very high resolution multispectral and hyperspectral images. By this comparison, we observed that the proposed method resulted in better accuracy with respect to other investigated and state-of-the art methods on both the considered data sets. Furthermore, we derived some guidelines on the design of AL systems for the classification of different types of RS images.

ei

Statistical mechanics analysis of sparse data

Habeck, M.

Journal of Structural Biology, 173(3):541-548, March 2011 (article)

Abstract
Inferential structure determination uses Bayesian theory to combine experimental data with prior structural knowledge into a posterior probability distribution over protein conformational space. The posterior distribution encodes everything one can say objectively about the native structure in the light of the available data and additional prior assumptions and can be searched for structural representatives. Here an analogy is drawn between the posterior distribution and the canonical ensemble of statistical physics. A statistical mechanics analysis assesses the complexity of a structure calculation globally in terms of ensemble properties. Analogs of the free energy and density of states are introduced; partition functions evaluate the consistency of prior assumptions with data. Critical behavior is observed with dwindling restraint density, which impairs structure determination with too sparse data. However, prior distributions with improved realism ameliorate the situation by lowering the critical number of observations. An in-depth analysis of various experimentally accessible structural parameters and force field terms will facilitate a statistical approach to protein structure determination with sparse data that avoids bias as much as possible.

ei

Large Scale Bayesian Inference and Experimental Design for Sparse Linear Models
SIAM Journal on Imaging Sciences, 4(1):166-199, March 2011 (article)

Abstract
Many problems of low-level computer vision and image processing, such as denoising, deconvolution, tomographic reconstruction or super-resolution, can be addressed by maximizing the posterior distribution of a sparse linear model (SLM). We show how higher-order Bayesian decision-making problems, such as optimizing image acquisition in magnetic resonance scanners, can be addressed by querying the SLM posterior covariance, unrelated to the density‘s mode. We propose a scalable algorithmic framework, with which SLM posteriors over full, high-resolution images can be approximated for the first time, solving a variational optimization problem which is convex iff posterior mode finding is convex. These methods successfully drive the optimization of sampling trajectories for real-world magnetic resonance imaging through Bayesian experimental design, which has not been attempted before. Our methodology provides new insight into similarities and differences between sparse reconstruction and approximate Bayesian inference, and has important implications for compressive sensing of real-world images.

ei

Learning grasp affordance densities

Detry, R., Kraft, D., Kroemer, O., Peters, J., Krüger, N., Piater, J.

Paladyn: Journal of Behavioral Robotics, 2(1):1-17, March 2011 (article)

Abstract
We address the issue of learning and representing object grasp affordance models. We model grasp affordances with continuous probability density functions (grasp densities) which link object-relative grasp poses to their success probability. The underlying function representation is nonparametric and relies on kernel density estimation to provide a continuous model. Grasp densities are learned and refined from exploration, by letting a robot “play” with an object in a sequence of grasp-and-drop actions: the robot uses visual cues to generate a set of grasp hypotheses, which it then executes and records their outcomes. When a satisfactory amount of grasp data is available, an importance-sampling algorithm turns it into a grasp density. We evaluate our method in a largely autonomous learning experiment, run on three objects with distinct shapes. The experiment shows how learning increases success rates. It also measures the success rate of grasps chosen to maximize the probability of success, given reaching constraints.

ei

A Database and Evaluation Methodology for Optical Flow

Baker, S., Scharstein, D., Lewis, J. P., Roth, S., Black, M. J., Szeliski, R.

International Journal of Computer Vision, 92(1):1-31, March 2011 (article)

Abstract
The quantitative evaluation of optical flow algorithms by Barron et al. (1994) led to significant advances in performance. The challenges for optical flow algorithms today go beyond the datasets and evaluation methods proposed in that paper. Instead, they center on problems associated with complex natural scenes, including nonrigid motion, real sensor noise, and motion discontinuities. We propose a new set of benchmarks and evaluation methods for the next generation of optical flow algorithms. To that end, we contribute four types of data to test different aspects of optical flow algorithms: (1) sequences with nonrigid motion where the ground-truth flow is determined by tracking hidden fluorescent texture, (2) realistic synthetic sequences, (3) high frame-rate video used to study interpolation error, and (4) modified stereo sequences of static scenes. In addition to the average angular error used by Barron et al., we compute the absolute flow endpoint error, measures for frame interpolation error, improved statistics, and results at motion discontinuities and in textureless regions. In October 2007, we published the performance of several well-known methods on a preliminary version of our data to establish the current state of the art. We also made the data freely available on the web at http://vision.middlebury.edu/flow/ . Subsequently a number of researchers have uploaded their results to our website and published papers using the data. A significant improvement in performance has already been achieved. In this paper we analyze the results obtained to date and draw a large number of conclusions from them.

ps

Client–Server Multitask Learning From Distributed Datasets

Dinuzzo, F., Pillonetto, G., De Nicolao, G.

IEEE Transactions on Neural Networks, 22(2):290-303, February 2011 (article)

Abstract
A client-server architecture to simultaneously solve multiple learning tasks from distributed datasets is described. In such architecture, each client corresponds to an individual learning task and the associated dataset of examples. The goal of the architecture is to perform information fusion from multiple datasets while preserving privacy of individual data. The role of the server is to collect data in real time from the clients and codify the information in a common database. Such information can be used by all the clients to solve their individual learning task, so that each client can exploit the information content of all the datasets without actually having access to private data of others. The proposed algorithmic framework, based on regularization and kernel methods, uses a suitable class of “mixed effect” kernels. The methodology is illustrated through a simulated recommendation system, as well as an experiment involving pharmacological data coming from a multicentric clinical trial.

ei

Extraction of functional information from ongoing brain electrical activity: Extraction en temps-réel d’informations fonctionnelles à partir de l’activité électrique cérébrale

Besserve, M., Martinerie, J.

IRBM, 32(1):27-34, February 2011 (article)

Abstract
The modern analysis of multivariate electrical brain signals requires advanced statistical tools to automatically extract and quantify their information content. These tools include machine learning techniques and information theory. They are currently used both in basic neuroscience and challenging applications such as brain computer interfaces. We review here how these methods have been used at the Laboratoire d’Électroencéphalographie et de Neurophysiologie Appliquée (LENA) to develop a general tool for the real time analysis of functional brain signals. We then give some perspectives on how these tools can help understanding the biological mechanisms of information processing.

ei