Results 1  10
of
22
Learning to select and generalize striking movements in robot table tennis
 In Proceedings of the AAAI 2012 Fall Symposium on robots that Learn Interactively from Human Teachers
, 2012
"... Learning new motor tasks from physical interactions is an important goal for both robotics and machine learning. However, when moving beyond basic skills, most monolithic machine learning approaches fail to scale. For more complex skills, methods that are tailored for the domain of skill learning ar ..."
Abstract

Cited by 27 (12 self)
 Add to MetaCart
Learning new motor tasks from physical interactions is an important goal for both robotics and machine learning. However, when moving beyond basic skills, most monolithic machine learning approaches fail to scale. For more complex skills, methods that are tailored for the domain of skill learning are needed. In this paper, we take the task of learning table tennis as an example and present a new framework that allows a robot to learn cooperative table tennis from physical interaction with a human. The robot first learns a set of elementary table tennis hitting movements from a human table tennis teacher by kinesthetic teachin, which is compiled into a set of motor primitives represented by dynamical systems. The robot subsequently generalizes these movements to a wider range of situations using our mixture of motor primitives approach. The resulting policy enables the robot to select appropriate motor primitives as well as to generalize between them. Finally, the robot plays with a human table tennis partner and learns online to improve its behavior. We show that the resulting setup is capable of playing table tennis using an anthropomorphic robot arm. 1
Learning sparse overcomplete codes for images
 J. VLSI Signal Process. Syst
"... Abstract. Images can be coded accurately using a sparse set of vectors from a learned overcomplete dictionary, with potential applications in image compression and feature selection for pattern recognition. We present a survey of algorithms that perform dictionary learning and sparse coding and make ..."
Abstract

Cited by 24 (3 self)
 Add to MetaCart
(Show Context)
Abstract. Images can be coded accurately using a sparse set of vectors from a learned overcomplete dictionary, with potential applications in image compression and feature selection for pattern recognition. We present a survey of algorithms that perform dictionary learning and sparse coding and make three contributions. First, we compare our overcomplete dictionary learning algorithm (FOCUSSCNDL) with overcomplete independent component analysis (ICA). Second, noting that once a dictionary has been learned in a given domain the problem becomes one of choosing the vectors to form an accurate, sparse representation, we compare a recently developed algorithm (sparse Bayesian learning with adjustable variance Gaussians, SBLAVG) to well known methods of subset selection: matching pursuit and FOCUSS. Third, noting that in some cases it may be necessary to find a nonnegative sparse coding, we present a modified version of the FOCUSS algorithm that can find such nonnegative codings. Efficient parallel implementations in VLSI could make these algorithms more practical for many applications.
Nonparametric Bayesian kernel models
 Discussion Paper 200509, Duke University ISDS
, 2007
"... Kernel models for classification and regression have emerged as widely applied tools in statistics and machine learning. We discuss a Bayesian framework and theory for kernel methods, providing a new rationalisation of kernel regression based on nonparametric Bayesian models. Functional analytic re ..."
Abstract

Cited by 14 (5 self)
 Add to MetaCart
(Show Context)
Kernel models for classification and regression have emerged as widely applied tools in statistics and machine learning. We discuss a Bayesian framework and theory for kernel methods, providing a new rationalisation of kernel regression based on nonparametric Bayesian models. Functional analytic results ensure that such a nonparametric prior specification induces a class of functions that span the reproducing kernel Hilbert space corresponding to the selected kernel. Bayesian analysis of the model allows for direct and formal inference on the uncertain regression or classification functions. Augmenting the model with Bayesian variable selection priors over kernel bandwidth parameters extends the framework to automatically address the key practical questions of kernel feature selection. Novel, customised MCMC methods are detailed and used in example analyses. The practical benefits and modelling flexibility of the Bayesian kernel framework are illustrated in both simulated and real data examples that address prediction and classification inference with highdimensional data.
Bayesian models for Largescale Hierarchical Classification
"... A challenging problem in hierarchical classification is to leverage the hierarchical relations among classes for improving classification performance. An even greater challenge is to do so in a manner that is computationally feasible for the large scale problems usually encountered in practice. This ..."
Abstract

Cited by 7 (0 self)
 Add to MetaCart
(Show Context)
A challenging problem in hierarchical classification is to leverage the hierarchical relations among classes for improving classification performance. An even greater challenge is to do so in a manner that is computationally feasible for the large scale problems usually encountered in practice. This paper proposes a set of Bayesian methods to model hierarchical dependencies among class labels using multivariate logistic regression. Specifically, the parentchild relationships are modeled by placing a hierarchical prior over the children nodes centered around the parameters of their parents; thereby encouraging classes nearby in the hierarchy to share similar model parameters. We present variational algorithms for tractable posterior inference in these models, and provide a parallel implementation that can comfortably handle largescale problems with hundreds of thousands of dimensions and tens of thousands of classes. We run a comparative evaluation on multiple largescale benchmark datasets that highlights the scalability of our approach, and shows a significant performance advantage over the other stateoftheart hierarchical methods. 1
A Sparse Nonlinear Bayesian Online Kernel Regression
 in Proceedings of the Second IEEE International Conference on Advanced Engineering Computing and Applications in Sciences (AdvComp 2008
, 2008
"... In a large number of applications, engineers have to estimate values of an unknown function given some observed samples. This task is referred to as function approximation or as generalization. One way to solve the problem is to regress a family of parameterized functions so as to make it fit at be ..."
Abstract

Cited by 6 (5 self)
 Add to MetaCart
(Show Context)
In a large number of applications, engineers have to estimate values of an unknown function given some observed samples. This task is referred to as function approximation or as generalization. One way to solve the problem is to regress a family of parameterized functions so as to make it fit at best the observed samples. Yet, usually batch methods are used and parameterization is habitually linear. Moreover, very few approaches try to quantify uncertainty reduction occurring when acquiring more samples (thus more information), which can be quite useful depending on the application. In this paper we propose a sparse nonlinear bayesian online kernel regression. Sparsity is achieved in a preprocessing step by using a dictionary method. The nonlinear bayesian kernel regression can therefore be considered as achieved online by a Sigma Point Kalman Filter. First experiments on a cardinal sine regression show that our approach is promising. 1
No freelunch and bayesian optimality
 IJCNN Workshop on MetaLearning
, 2007
"... We take a Bayesian approach to the issues of bias, meta bias, transfer, overfit, and NoFreeLunch in the context of supervised learning. If we accept certain relationships between the function class, on training set data, and off training set data, then a graphical model can be created that represe ..."
Abstract

Cited by 6 (2 self)
 Add to MetaCart
(Show Context)
We take a Bayesian approach to the issues of bias, meta bias, transfer, overfit, and NoFreeLunch in the context of supervised learning. If we accept certain relationships between the function class, on training set data, and off training set data, then a graphical model can be created that represents the supervised learning problem. This graphical model dictates a specific algorithm which will be the “optimal” approach to learning the parameters of any given function representation given the variable relationships. Thus, there is an optimal technique for supervised learning. We reconcile this idea of an optimal technique with the ideas of NoFreeLunch and show how these ideas relate to the concepts of meta and transfer learning through hierarchical versions of the graphical model. 1
A Bayesian decision theoretical approach to supervised learning, selective sampling, and empirical function optimization
, 2010
"... This Dissertation is brought to you for free and open access by BYU ScholarsArchive. It has been accepted for inclusion in All Theses and Dissertations by an authorized administrator of BYU ScholarsArchive. For more information, please contact scholarsarchive@byu.edu. ..."
Abstract

Cited by 3 (1 self)
 Add to MetaCart
(Show Context)
This Dissertation is brought to you for free and open access by BYU ScholarsArchive. It has been accepted for inclusion in All Theses and Dissertations by an authorized administrator of BYU ScholarsArchive. For more information, please contact scholarsarchive@byu.edu.
A utile function optimizer
 in The Proceedings of the IEEE Congress on Evolutionary Computation (CEC) (accepted
, 2007
"... Abstract — We recast the problem of unconstrained continuous evolutionary optimization as inference in a fixed graphical model. This approach allows us to address several pervasive issues in optimization, including the traditionally difficult problem of selecting an algorithm that is most appropriat ..."
Abstract

Cited by 3 (3 self)
 Add to MetaCart
(Show Context)
Abstract — We recast the problem of unconstrained continuous evolutionary optimization as inference in a fixed graphical model. This approach allows us to address several pervasive issues in optimization, including the traditionally difficult problem of selecting an algorithm that is most appropriate for a given task. This is accomplished by placing a prior distribution over the expected class of functions, then employing inference and intuitively defined utilities and costs to transform the evolutionary optimization problem into one of active sampling. This allows us to pose an approach to optimization that is optimal for each expressly stated function class. The resulting solution methodology can optimally navigate explorationexploitation tradeoffs using wellmotivated decision theory, while providing the process with a natural stopping criterion. Finally, the model naturally accommodates the expression of dynamic and noisy functions, setting it apart from most existing algorithms that address these issues as an afterthought. We demonstrate the characteristics and advantages of this algorithm formally and with examples. I.
BAYESIAN MODELS AND MACHINE LEARNING WITH GENE EXPRESSION ANALYSIS APPLICATIONS
, 2005
"... The present thesis is divided into two major parts. The first part focuses on developing modelbased estimates for gene expression indices in the Bayesian framework. In the application of oligonucleotide expression array technology, reliable estimation of expression indices is critical for “highlev ..."
Abstract

Cited by 2 (1 self)
 Add to MetaCart
The present thesis is divided into two major parts. The first part focuses on developing modelbased estimates for gene expression indices in the Bayesian framework. In the application of oligonucleotide expression array technology, reliable estimation of expression indices is critical for “highlevel analysis ” such as classification, clustering and regulatory network exploration. A statistical model (Li and Wong, 2001a) has been proposed to develop modelbased estimates for gene expression indices and outlier detection. Chapter 1 illustrates an extension of the model in the Bayesian framework. Proper constraints on model parameters, heavytail distributions for noise, and mixture priors are introduced with the help of Gibbs sampling. Our model is applied to both artificial probe data and real microarray probe data, with a demonstration that it is more robust and reliable than the original model. The second part of the thesis concerns a novel Bayesian models for the problem of nonlinear regression for prediction. Recently, kernel methods have been introduced
Introducing Dynamic Prior Knowledge to PartiallyBlurred Image Restoration
"... Abstract. The paper presents an unsupervised method for partiallyblurred image restoration without influencing unblurred regions or objects. Maximum a posteriori estimation of parameters in Bayesian regularization is equal to minimizing energy of a dataset for a given number of classes. To estimate ..."
Abstract

Cited by 1 (1 self)
 Add to MetaCart
(Show Context)
Abstract. The paper presents an unsupervised method for partiallyblurred image restoration without influencing unblurred regions or objects. Maximum a posteriori estimation of parameters in Bayesian regularization is equal to minimizing energy of a dataset for a given number of classes. To estimate the point spread function (PSF), a parametric model space is introduced to reduce the searching uncertainty for PSF model selection. Simultaneously, PSF selfinitializing does not rely on supervision or thresholds. In the image domain, a gradient map as apriori knowledge is derived not only for dynamically choosing nonlinear diffusion operators but also for segregating blurred and unblurred regions via an extended graphtheoretic method. The cost functions with respect to the image and the PSF are alternately minimized in a convex manner. The algorithm is robust in that it can handle images that are formed in variational environments with different blur and stronger noise. 1