Header logo is ei


2006


no image
A Kernel Method for the Two-Sample-Problem

Gretton, A., Borgwardt, K., Rasch, M., Schölkopf, B., Smola, A.

20th Annual Conference on Neural Information Processing Systems (NIPS), December 2006 (talk)

Abstract
We propose two statistical tests to determine if two samples are from different distributions. Our test statistic is in both cases the distance between the means of the two samples mapped into a reproducing kernel Hilbert space (RKHS). The first test is based on a large deviation bound for the test statistic, while the second is based on the asymptotic distribution of this statistic. We show that the test statistic can be computed in $O(m^2)$ time. We apply our approach to a variety of problems, including attribute matching for databases using the Hungarian marriage method, where our test performs strongly. We also demonstrate excellent performance when comparing distributions over graphs, for which no alternative tests currently exist.

PDF [BibTex]

2006

PDF [BibTex]


no image
Ab-initio gene finding using machine learning

Schweikert, G., Zeller, G., Zien, A., Ong, C., de Bona, F., Sonnenburg, S., Phillips, P., Rätsch, G.

NIPS Workshop on New Problems and Methods in Computational Biology, December 2006 (talk)

Web [BibTex]

Web [BibTex]


no image
Graph boosting for molecular QSAR analysis

Saigo, H., Kadowaki, T., Kudo, T., Tsuda, K.

NIPS Workshop on New Problems and Methods in Computational Biology, December 2006 (talk)

Abstract
We propose a new boosting method that systematically combines graph mining and mathematical programming-based machine learning. Informative and interpretable subgraph features are greedily found by a series of graph mining calls. Due to our mathematical programming formulation, subgraph features and pre-calculated real-valued features are seemlessly integrated. We tested our algorithm on a quantitative structure-activity relationship (QSAR) problem, which is basically a regression problem when given a set of chemical compounds. In benchmark experiments, the prediction accuracy of our method favorably compared with the best results reported on each dataset.

Web [BibTex]

Web [BibTex]


no image
Inferring Causal Directions by Evaluating the Complexity of Conditional Distributions

Sun, X., Janzing, D., Schölkopf, B.

NIPS Workshop on Causality and Feature Selection, December 2006 (talk)

Abstract
We propose a new approach to infer the causal structure that has generated the observed statistical dependences among n random variables. The idea is that the factorization of the joint measure of cause and effect into P(cause)P(effect|cause) leads typically to simpler conditionals than non-causal factorizations. To evaluate the complexity of the conditionals we have tried two methods. First, we have compared them to those which maximize the conditional entropy subject to the observed first and second moments since we consider the latter as the simplest conditionals. Second, we have fitted the data with conditional probability measures being exponents of functions in an RKHS space and defined the complexity by a Hilbert-space semi-norm. Such a complexity measure has several properties that are useful for our purpose. We describe some encouraging results with both methods applied to real-world data. Moreover, we have combined constraint-based approaches to causal discovery (i.e., methods using only information on conditional statistical dependences) with our method in order to distinguish between causal hypotheses which are equivalent with respect to the imposed independences. Furthermore, we compare the performance to Bayesian approaches to causal inference.

Web [BibTex]


no image
Minimal Logical Constraint Covering Sets

Sinz, F., Schölkopf, B.

(155), Max Planck Institute for Biological Cybernetics, Tübingen, December 2006 (techreport)

Abstract
We propose a general framework for computing minimal set covers under class of certain logical constraints. The underlying idea is to transform the problem into a mathematical programm under linear constraints. In this sense it can be seen as a natural extension of the vector quantization algorithm proposed by Tipping and Schoelkopf. We show which class of logical constraints can be cast and relaxed into linear constraints and give an algorithm for the transformation.

PDF [BibTex]

PDF [BibTex]


no image
Learning Optimal EEG Features Across Time, Frequency and Space

Farquhar, J., Hill, J., Schölkopf, B.

NIPS Workshop on Current Trends in Brain-Computer Interfacing, December 2006 (talk)

PDF Web [BibTex]

PDF Web [BibTex]


no image
Semi-Supervised Learning

Zien, A.

Advanced Methods in Sequence Analysis Lectures, November 2006 (talk)

Web [BibTex]

Web [BibTex]


no image
New Methods for the P300 Visual Speller

Biessmann, F.

(1), (Editors: Hill, J. ), Max-Planck Institute for Biological Cybernetics, Tübingen, Germany, November 2006 (techreport)

PDF [BibTex]

PDF [BibTex]


no image
A Machine Learning Approach for Determining the PET Attenuation Map from Magnetic Resonance Images

Hofmann, M., Steinke, F., Judenhofer, M., Claussen, C., Schölkopf, B., Pichler, B.

IEEE Medical Imaging Conference, November 2006 (talk)

Abstract
A promising new combination in multimodality imaging is MR-PET, where the high soft tissue contrast of Magnetic Resonance Imaging (MRI) and the functional information of Positron Emission Tomography (PET) are combined. Although many technical problems have recently been solved, it is still an open problem to determine the attenuation map from the available MR scan, as the MR intensities are not directly related to the attenuation values. One standard approach is an atlas registration where the atlas MR image is aligned with the patient MR thus also yielding an attenuation image for the patient. We also propose another approach, which to our knowledge has not been tried before: Using Support Vector Machines we predict the attenuation value directly from the local image information. We train this well-established machine learning algorithm using small image patches. Although both approaches sometimes yielded acceptable results, they also showed their specific shortcomings: The registration often fails with large deformations whereas the prediction approach is problematic when the local image structure is not characteristic enough. However, the failures often do not coincide and integration of both information sources is promising. We therefore developed a combination method extending Support Vector Machines to use not only local image structure but also atlas registered coordinates. We demonstrate the strength of this combination approach on a number of examples.

[BibTex]

[BibTex]


no image
Geometric Analysis of Hilbert Schmidt Independence criterion based ICA contrast function

Shen, H., Jegelka, S., Gretton, A.

(PA006080), National ICT Australia, Canberra, Australia, October 2006 (techreport)

Web [BibTex]

Web [BibTex]


no image
Semi-Supervised Support Vector Machines and Application to Spam Filtering

Zien, A.

ECML Discovery Challenge Workshop, September 2006 (talk)

Abstract
After introducing the semi-supervised support vector machine (aka TSVM for "transductive SVM"), a few popular training strategies are briefly presented. Then the assumptions underlying semi-supervised learning are reviewed. Finally, two modern TSVM optimization techniques are applied to the spam filtering data sets of the workshop; it is shown that they can achieve excellent results, if the problem of the data being non-iid can be handled properly.

PDF Web [BibTex]


no image
Inferential Structure Determination: Probabilistic determination and validation of NMR structures

Habeck, M.

Gordon Research Conference on Computational Aspects of Biomolecular NMR, September 2006 (talk)

Web [BibTex]

Web [BibTex]


no image
A tutorial on spectral clustering

von Luxburg, U.

(149), Max Planck Institute for Biological Cybernetics, Tübingen, August 2006 (techreport)

Abstract
In recent years, spectral clustering has become one of the most popular modern clustering algorithms. It is simple to implement, can be solved efficiently by standard linear algebra software, and very often outperforms traditional clustering algorithms such as the k-means algorithm. Nevertheless, on the first glance spectral clustering looks a bit mysterious, and it is not obvious to see why it works at all and what it really does. This article is a tutorial introduction to spectral clustering. We describe different graph Laplacians and their basic properties, present the most common spectral clustering algorithms, and derive those algorithms from scratch by several different approaches. Advantages and disadvantages of the different spectral clustering algorithms are discussed.

PDF [BibTex]

PDF [BibTex]


no image
Machine Learning Algorithms for Polymorphism Detection

Schweikert, G., Zeller, G., Clark, R., Ossowski, S., Warthmann, N., Shinn, P., Frazer, K., Ecker, J., Huson, D., Weigel, D., Schölkopf, B., Rätsch, G.

2nd ISCB Student Council Symposium, August 2006 (talk)

Abstract
Analyzing resequencing array data using machine learning, we obtain a genome-wide inventory of polymorphisms in 20 wild strains of Arabidopsis thaliana, including 750,000 single nucleotide poly- morphisms (SNPs) and thousands of highly polymorphic regions and deletions. We thus provide an unprecedented resource for the study of natural variation in plants.

Web [BibTex]

Web [BibTex]


no image
Towards the Inference of Graphs on Ordered Vertexes

Zien, A., Raetsch, G., Ong, C.

(150), Max Planck Institute for Biological Cybernetics, Tübingen, August 2006 (techreport)

Abstract
We propose novel methods for machine learning of structured output spaces. Specifically, we consider outputs which are graphs with vertices that have a natural order. We consider the usual adjacency matrix representation of graphs, as well as two other representations for such a graph: (a) decomposing the graph into a set of paths, (b) converting the graph into a single sequence of nodes with labeled edges. For each of the three representations, we propose an encoding and decoding scheme. We also propose an evaluation measure for comparing two graphs.

PDF [BibTex]

PDF [BibTex]


no image
Inferential structure determination: Overview and new developments

Habeck, M.

Sixth CCPN Annual Conference: Efficient and Rapid Structure Determination by NMR, July 2006 (talk)

Web [BibTex]

Web [BibTex]


no image
MCMC inference in (Conditionally) Conjugate Dirichlet Process Gaussian Mixture Models

Rasmussen, C., Görür, D.

ICML Workshop on Learning with Nonparametric Bayesian Methods, June 2006 (talk)

Abstract
We compare the predictive accuracy of the Dirichlet Process Gaussian mixture models using conjugate and conditionally conjugate priors and show that better density models result from using the wider class of priors. We explore several MCMC schemes exploiting conditional conjugacy and show their computational merits on several multidimensional density estimation problems.

Web [BibTex]

Web [BibTex]


no image
Sampling for non-conjugate infinite latent feature models

Görür, D., Rasmussen, C.

(Editors: Bernardo, J. M.), 8th Valencia International Meeting on Bayesian Statistics (ISBA), June 2006 (talk)

Abstract
Latent variable models are powerful tools to model the underlying structure in data. Infinite latent variable models can be defined using Bayesian nonparametrics. Dirichlet process (DP) models constitute an example of infinite latent class models in which each object is assumed to belong to one of the, mutually exclusive, infinitely many classes. Recently, the Indian buffet process (IBP) has been defined as an extension of the DP. IBP is a distribution over sparse binary matrices with infinitely many columns which can be used as a distribution for non-exclusive features. Inference using Markov chain Monte Carlo (MCMC) in conjugate IBP models has been previously described, however requiring conjugacy restricts the use of IBP. We describe an MCMC algorithm for non-conjugate IBP models. Modelling the choice behaviour is an important topic in psychology, economics and related fields. Elimination by Aspects (EBA) is a choice model that assumes each alternative has latent features with associated weights that lead to the observed choice outcomes. We formulate a non-parametric version of EBA by using IBP as the prior over the latent binary features. We infer the features of objects that lead to the choice data by using our sampling scheme for inference.

PDF [BibTex]

PDF [BibTex]


no image
An Automated Combination of Sequence Motif Kernels for Predicting Protein Subcellular Localization

Zien, A., Ong, C.

(146), Max Planck Institute for Biological Cybernetics, Tübingen, April 2006 (techreport)

Abstract
Protein subcellular localization is a crucial ingredient to many important inferences about cellular processes, including prediction of protein function and protein interactions. While many predictive computational tools have been proposed, they tend to have complicated architectures and require many design decisions from the developer. We propose an elegant and fully automated approach to building a prediction system for protein subcellular localization. We propose a new class of protein sequence kernels which considers all motifs including motifs with gaps. This class of kernels allows the inclusion of pairwise amino acid distances into their computation. We further propose a multiclass support vector machine method which directly solves protein subcellular localization without resorting to the common approach of splitting the problem into several binary classification problems. To automatically search over families of possible amino acid motifs, we generalize our method to optimize over multiple kernels at the same time. We compare our automated approach to four other predictors on three different datasets.

PDF Web [BibTex]

PDF Web [BibTex]


no image
Training a Support Vector Machine in the Primal

Chapelle, O.

(147), Max Planck Institute for Biological Cybernetics, Tübingen, April 2006, The version in the "Large Scale Kernel Machines" book is more up to date. (techreport)

Abstract
Most literature on Support Vector Machines (SVMs) concentrate on the dual optimization problem. In this paper, we would like to point out that the primal problem can also be solved efficiently, both for linear and non-linear SVMs, and there is no reason for ignoring it. Moreover, from the primal point of view, new families of algorithms for large scale SVM training can be investigated.

PDF [BibTex]

PDF [BibTex]


no image
An Inventory of Sequence Polymorphisms For Arabidopsis

Clark, R., Ossowski, S., Schweikert, G., Rätsch, G., Shinn, P., Zeller, G., Warthmann, N., Fu, G., Hinds, D., Chen, H., Frazer, K., Huson, D., Schölkopf, B., Nordborg, M., Ecker, J., Weigel, D.

17th International Conference on Arabidopsis Research, April 2006 (talk)

Abstract
We have used high-density oligonucleotide arrays to characterize common sequence variation in 20 wild strains of Arabidopsis thaliana that were chosen for maximal genetic diversity. Both strands of each possible SNP of the 119 Mb reference genome were represented on the arrays, which were hybridized with whole genome, isothermally amplified DNA to minimize ascertainment biases. Using two complementary approaches, a model based algorithm, and a newly developed machine learning method, we identified over 550,000 SNPs with a false discovery rate of ~ 0.03 (average of 1 SNP for every 216 bp of the genome). A heuristic algorithm predicted in addition ~700 highly polymorphic or deleted regions per accession. Over 700 predicted polymorphisms with major functional effects (e.g., premature stop codons, or deletions of coding sequence) were validated by dideoxy sequencing. Using this data set, we provide the first systematic description of the types of genes that harbor major effect polymorphisms in natural populations at moderate allele frequencies. The data also provide an unprecedented resource for the study of genetic variation in an experimentally tractable, multicellular model organism.

[BibTex]

[BibTex]


no image
Cross-Validation Optimization for Structured Hessian Kernel Methods

Seeger, M., Chapelle, O.

Max-Planck Institute for Biological Cybernetics, Tübingen, Germany, February 2006 (techreport)

Abstract
We address the problem of learning hyperparameters in kernel methods for which the Hessian of the objective is structured. We propose an approximation to the cross-validation log likelihood whose gradient can be computed analytically, solving the hyperparameter learning problem efficiently through nonlinear optimization. Crucially, our learning method is based entirely on matrix-vector multiplication primitives with the kernel matrices and their derivatives, allowing straightforward specialization to new kernels or to large datasets. When applied to the problem of multi-way classification, our method scales linearly in the number of classes and gives rise to state-of-the-art results on a remote imaging task.

PDF Web [BibTex]

PDF Web [BibTex]

2003


no image
Support Vector Channel Selection in BCI

Lal, T., Schröder, M., Hinterberger, T., Weston, J., Bogdan, M., Birbaumer, N., Schölkopf, B.

(120), Max Planck Institute for Biological Cybernetics, Tuebingen, Germany, December 2003 (techreport)

Abstract
Designing a Brain Computer Interface (BCI) system one can choose from a variety of features that may be useful for classifying brain activity during a mental task. For the special case of classifying EEG signals we propose the usage of the state of the art feature selection algorithms Recursive Feature Elimination [3] and Zero-Norm Optimization [13] which are based on the training of Support Vector Machines (SVM) [11]. These algorithms can provide more accurate solutions than standard filter methods for feature selection [14]. We adapt the methods for the purpose of selecting EEG channels. For a motor imagery paradigm we show that the number of used channels can be reduced significantly without increasing the classification error. The resulting best channels agree well with the expected underlying cortical activity patterns during the mental tasks. Furthermore we show how time dependent task specific information can be visualized.

PDF Web [BibTex]

2003

PDF Web [BibTex]


no image
Image Reconstruction by Linear Programming

Tsuda, K., Rätsch, G.

(118), Max Planck Institute for Biological Cybernetics, Tübingen, Germany, October 2003 (techreport)

PDF [BibTex]

PDF [BibTex]


no image
Statistical Learning Theory

Bousquet, O.

Machine Learning Summer School, August 2003 (talk)

PDF [BibTex]

PDF [BibTex]


no image
Remarks on Statistical Learning Theory

Bousquet, O.

Machine Learning Summer School, August 2003 (talk)

PDF [BibTex]

PDF [BibTex]


no image
Ranking on Data Manifolds

Zhou, D., Weston, J., Gretton, A., Bousquet, O., Schölkopf, B.

(113), Max Planck Institute for Biological Cybernetics, 72076 Tuebingen, Germany, June 2003 (techreport)

Abstract
The Google search engine has had a huge success with its PageRank web page ranking algorithm, which exploits global, rather than local, hyperlink structure of the World Wide Web using random walk. This algorithm can only be used for graph data, however. Here we propose a simple universal ranking algorithm for vectorial data, based on the exploration of the intrinsic global geometric structure revealed by a huge amount of data. Experimental results from image and text to bioinformatics illustrates the validity of our algorithm.

PDF [BibTex]

PDF [BibTex]


no image
Kernel Hebbian Algorithm for Iterative Kernel Principal Component Analysis

Kim, K., Franz, M., Schölkopf, B.

(109), MPI f. biologische Kybernetik, Tuebingen, June 2003 (techreport)

Abstract
A new method for performing a kernel principal component analysis is proposed. By kernelizing the generalized Hebbian algorithm, one can iteratively estimate the principal components in a reproducing kernel Hilbert space with only linear order memory complexity. The derivation of the method, a convergence proof, and preliminary applications in image hyperresolution are presented. In addition, we discuss the extension of the method to the online learning of kernel principal components.

PDF [BibTex]

PDF [BibTex]


no image
Learning with Local and Global Consistency

Zhou, D., Bousquet, O., Lal, T., Weston, J., Schölkopf, B.

(112), Max Planck Institute for Biological Cybernetics, Tuebingen, Germany, June 2003 (techreport)

Abstract
We consider the learning problem in the transductive setting. Given a set of points of which only some are labeled, the goal is to predict the label of the unlabeled points. A principled clue to solve such a learning problem is the consistency assumption that a classifying function should be sufficiently smooth with respect to the structure revealed by these known labeled and unlabeled points. We present a simple algorithm to obtain such a smooth solution. Our method yields encouraging experimental results on a number of classification problems and demonstrates effective use of unlabeled data.

[BibTex]

[BibTex]


no image
Implicit Wiener Series

Franz, M., Schölkopf, B.

(114), Max Planck Institute for Biological Cybernetics, June 2003 (techreport)

Abstract
The Wiener series is one of the standard methods to systematically characterize the nonlinearity of a neural system. The classical estimation method of the expansion coefficients via cross-correlation suffers from severe problems that prevent its application to high-dimensional and strongly nonlinear systems. We propose a new estimation method based on regression in a reproducing kernel Hilbert space that overcomes these problems. Numerical experiments show performance advantages in terms of convergence, interpretability and system size that can be handled.

PDF [BibTex]

PDF [BibTex]


no image
Machine Learning approaches to protein ranking: discriminative, semi-supervised, scalable algorithms

Weston, J., Leslie, C., Elisseeff, A., Noble, W.

(111), Max Planck Institute for Biological Cybernetics, Tübingen, Germany, June 2003 (techreport)

Abstract
A key tool in protein function discovery is the ability to rank databases of proteins given a query amino acid sequence. The most successful method so far is a web-based tool called PSI-BLAST which uses heuristic alignment of a profile built using the large unlabeled database. It has been shown that such use of global information via an unlabeled data improves over a local measure derived from a basic pairwise alignment such as performed by PSI-BLAST's predecessor, BLAST. In this article we look at ways of leveraging techniques from the field of machine learning for the problem of ranking. We show how clustering and semi-supervised learning techniques, which aim to capture global structure in data, can significantly improve over PSI-BLAST.

PDF [BibTex]

PDF [BibTex]


no image
The Geometry Of Kernel Canonical Correlation Analysis

Kuss, M., Graepel, T.

(108), Max Planck Institute for Biological Cybernetics, Tübingen, Germany, May 2003 (techreport)

Abstract
Canonical correlation analysis (CCA) is a classical multivariate method concerned with describing linear dependencies between sets of variables. After a short exposition of the linear sample CCA problem and its analytical solution, the article proceeds with a detailed characterization of its geometry. Projection operators are used to illustrate the relations between canonical vectors and variates. The article then addresses the problem of CCA between spaces spanned by objects mapped into kernel feature spaces. An exact solution for this kernel canonical correlation (KCCA) problem is derived from a geometric point of view. It shows that the expansion coefficients of the canonical vectors in their respective feature space can be found by linear CCA in the basis induced by kernel principal component analysis. The effect of mappings into higher dimensional feature spaces is considered critically since it simplifies the CCA problem in general. Then two regularized variants of KCCA are discussed. Relations to other methods are illustrated, e.g., multicategory kernel Fisher discriminant analysis, kernel principal component regression and possible applications thereof in blind source separation.

PDF [BibTex]

PDF [BibTex]


no image
The Kernel Mutual Information

Gretton, A., Herbrich, R., Smola, A.

Max Planck Institute for Biological Cybernetics, April 2003 (techreport)

Abstract
We introduce two new functions, the kernel covariance (KC) and the kernel mutual information (KMI), to measure the degree of independence of several continuous random variables. The former is guaranteed to be zero if and only if the random variables are pairwise independent; the latter shares this property, and is in addition an approximate upper bound on the mutual information, as measured near independence, and is based on a kernel density estimate. We show that Bach and Jordan‘s kernel generalised variance (KGV) is also an upper bound on the same kernel density estimate, but is looser. Finally, we suggest that the addition of a regularising term in the KGV causes it to approach the KMI, which motivates the introduction of this regularisation. The performance of the KC and KMI is verified in the context of instantaneous independent component analysis (ICA), by recovering both artificial and real (musical) signals following linear mixing.

PostScript [BibTex]

PostScript [BibTex]


no image
Rademacher and Gaussian averages in Learning Theory

Bousquet, O.

Universite de Marne-la-Vallee, March 2003 (talk)

PDF [BibTex]

PDF [BibTex]


no image
Introduction: Robots with Cognition?

Franz, MO.

6, pages: 38, (Editors: H.H. Bülthoff, K.R. Gegenfurtner, H.A. Mallot, R. Ulrich, F.A. Wichmann), 6. T{\"u}binger Wahrnehmungskonferenz (TWK), February 2003 (talk)

Abstract
Using robots as models of cognitive behaviour has a long tradition in robotics. Parallel to the historical development in cognitive science, one observes two major, subsequent waves in cognitive robotics. The first is based on ideas of classical, cognitivist Artificial Intelligence (AI). According to the AI view of cognition as rule-based symbol manipulation, these robots typically try to extract symbolic descriptions of the environment from their sensors that are used to update a common, global world representation from which, in turn, the next action of the robot is derived. The AI approach has been successful in strongly restricted and controlled environments requiring well-defined tasks, e.g. in industrial assembly lines. AI-based robots mostly failed, however, in the unpredictable and unstructured environments that have to be faced by mobile robots. This has provoked the second wave in cognitive robotics which tries to achieve cognitive behaviour as an emergent property from the interaction of simple, low-level modules. Robots of the second wave are called animats as their architecture is designed to closely model aspects of real animals. Using only simple reactive mechanisms and Hebbian-type or evolutionary learning, the resulting animats often outperformed the highly complex AI-based robots in tasks such as obstacle avoidance, corridor following etc. While successful in generating robust, insect-like behaviour, typical animats are limited to stereotyped, fixed stimulus-response associations. If one adopts the view that cognition requires a flexible, goal-dependent choice of behaviours and planning capabilities (H.A. Mallot, Kognitionswissenschaft, 1999, 40-48) then it appears that cognitive behaviour cannot emerge from a collection of purely reactive modules. It rather requires environmentally decoupled structures that work without directly engaging the actions that it is concerned with. This poses the current challenge to cognitive robotics: How can we build cognitive robots that show the robustness and the learning capabilities of animats without falling back into the representational paradigm of AI? The speakers of the symposium present their approaches to this question in the context of robot navigation and sensorimotor learning. In the first talk, Prof. Helge Ritter introduces a robot system for imitation learning capable of exploring various alternatives in simulation before actually performing a task. The second speaker, Angelo Arleo, develops a model of spatial memory in rat navigation based on his electrophysiological experiments. He validates the model on a mobile robot which, in some navigation tasks, shows a performance comparable to that of the real rat. A similar model of spatial memory is used to investigate the mechanisms of territory formation in a series of robot experiments presented by Prof. Hanspeter Mallot. In the last talk, we return to the domain of sensorimotor learning where Ralf M{\"o}ller introduces his approach to generate anticipatory behaviour by learning forward models of sensorimotor relationships.

Web [BibTex]

Web [BibTex]


no image
A Note on Parameter Tuning for On-Line Shifting Algorithms

Bousquet, O.

Max Planck Institute for Biological Cybernetics, Tübingen, Germany, 2003 (techreport)

Abstract
In this short note, building on ideas of M. Herbster [2] we propose a method for automatically tuning the parameter of the FIXED-SHARE algorithm proposed by Herbster and Warmuth [3] in the context of on-line learning with shifting experts. We show that this can be done with a memory requirement of $O(nT)$ and that the additional loss incurred by the tuning is the same as the loss incurred for estimating the parameter of a Bernoulli random variable.

PDF PostScript [BibTex]

PDF PostScript [BibTex]


no image
Interactive Images

Toyama, K., Schölkopf, B.

(MSR-TR-2003-64), Microsoft Research, Cambridge, UK, 2003 (techreport)

Abstract
Interactive Images are a natural extension of three recent developments: digital photography, interactive web pages, and browsable video. An interactive image is a multi-dimensional image, displayed two dimensions at a time (like a standard digital image), but with which a user can interact to browse through the other dimensions. One might consider a standard video sequence viewed with a video player as a simple interactive image with time as the third dimension. Interactive images are a generalization of this idea, in which the third (and greater) dimensions may be focus, exposure, white balance, saturation, and other parameters. Interaction is handled via a variety of modes including those we call ordinal, pixel-indexed, cumulative, and comprehensive. Through exploration of three novel forms of interactive images based on color, exposure, and focus, we will demonstrate the compelling nature of interactive images.

Web [BibTex]

Web [BibTex]

2002


no image
Kernel Dependency Estimation

Weston, J., Chapelle, O., Elisseeff, A., Schölkopf, B., Vapnik, V.

(98), Max Planck Institute for Biological Cybernetics, August 2002 (techreport)

Abstract
We consider the learning problem of finding a dependency between a general class of objects and another, possibly different, general class of objects. The objects can be for example: vectors, images, strings, trees or graphs. Such a task is made possible by employing similarity measures in both input and output spaces using kernel functions, thus embedding the objects into vector spaces. Output kernels also make it possible to encode prior information and/or invariances in the loss function in an elegant way. We experimentally validate our approach on several tasks: mapping strings to strings, pattern recognition, and reconstruction from partial images.

PDF [BibTex]

2002

PDF [BibTex]


no image
A compression approach to support vector model selection

von Luxburg, U., Bousquet, O., Schölkopf, B.

(101), Max Planck Institute for Biological Cybernetics, 2002, see more detailed JMLR version (techreport)

Abstract
In this paper we investigate connections between statistical learning theory and data compression on the basis of support vector machine (SVM) model selection. Inspired by several generalization bounds we construct ``compression coefficients'' for SVMs, which measure the amount by which the training labels can be compressed by some classification hypothesis. The main idea is to relate the coding precision of this hypothesis to the width of the margin of the SVM. The compression coefficients connect well known quantities such as the radius-margin ratio R^2/rho^2, the eigenvalues of the kernel matrix and the number of support vectors. To test whether they are useful in practice we ran model selection experiments on several real world datasets. As a result we found that compression coefficients can fairly accurately predict the parameters for which the test error is minimized.

[BibTex]

[BibTex]