The curse of dimensionality gives rise to prohibitive computational requirements that render infeasible the exact solution of large-scale stochastic control problems. We study an efficient method based on linear programming for approximating solutions to such problems. Theapproach "fits" a linear combination of pre-selected basis functions to the dynamic programming cost-to- go function. We develop bounds on the approximation error and present experimental resultsin the domain of queueing network control, providing empirical support for the methodology.

We describe a neural network which enhances and completes salient closed contours. Our work is different from all previous work in three important ways. First, like the input provided to V1 by LGN, the input toour computation is isotropic. That is, the input is composed of spots not edges. Second, our network computes a well defined function of the input based on a distribution of closed contours characterized by a random process. Third, even though our computation is implemented in a discrete network, its output is invariant to continuous rotations and translations of the input pattern.

Sham Kakade Gatsby Computational Neuroscience Unit 17 Queen Square, London, UK WC1N 3AR http://www.gatsby.ucl.ac.uk sham@gatsby.ucl.ac.uk Abstract We provide a natural gradient method that represents the steepest descent direction based on the underlying structure of the parameter space.Although gradient methods cannot make large changes in the values of the parameters, we show that the natural gradient ismoving toward choosing a greedy optimal action rather than just a better action. These greedy optimal actions are those that would be chosen under one improvement step of policy iteration with approximate, compatible value functions, as defined by Sutton etal. We then show drastic performance improvements in simple MDPs and in the more challenging MDP of Tetris. 1 Introduction There has been a growing interest in direct policy-gradient methods for approximate planning in large Markov decision problems (MDPs). Unfortunately, the standard gradient descent rule is noncovariant. In this paper, we present a covariant gradient by defining a metric based on the underlying structure of the policy.

Neural activity appears to be a crucial component for shaping the receptive fieldsof cortical simple cells into adjacent, oriented subregions alternately receivingON-and OFFcenter excitatory geniculate inputs. It is known that the orientation selective responses of V1 neurons are refined by visual experience. After eye opening, the spatiotemporal structure of neural activity in the early stages of the visual pathway depends both on the visual environment and on how the environment is scanned. We have used computational modeling to investigate how eye movements might affect the refinement of the orientation tuning of simple cells in the presence ofa Hebbian scheme of synaptic plasticity. Levels of correlation between theactivity of simulated cells were examined while natural scenes were scanned so as to model sequences of saccades and fixational eye movements, such as microsaccades, tremor and ocular drift. The specific patterns of activity required for a quantitatively accurate development of simple cell receptive fields with segregated ON and OFF subregions were observed during fixational eye movements, but not in the presence of saccades or with static presentation of natural visual input.

In contrast to standard statistical learning theory which studies uniform bounds on the expected error we present a framework that exploits the specific learning algorithm used. Motivated by the luckiness framework [8] we are also able to exploit the serendipity of the training sample. The main difference to previous approaches lies in the complexity measure; rather than covering all hypotheses ina given hypothesis space it is only necessary to cover the functions which could have been learned using the fixed learning algorithm. We show how the resulting framework relates to the VC, luckiness and compression frameworks. Finally, we present an application of this framework to the maximum margin algorithm for linear classifiers which results in a bound that exploits both the margin and the distribution of the data in feature space. 1 Introduction Statistical learning theory is mainly concerned with the study of uniform bounds on the expected error of hypotheses from a given hypothesis space [9, 1].

The question of whether the nervous system produces movement through the combination of a few discrete elements has long been central to the study of motor control. Muscle synergies, i.e. coordinated patterns of muscle activity, have been proposed as possible building blocks. Here we propose a model based on combinations of muscle synergies with a specific amplitudeand temporal structure. Time-varying synergies provide a realistic basis for the decomposition of the complex patterns observed in natural behaviors. To extract time-varying synergies from simultaneous recordingof EMG activity we developed an algorithm which extends existing nonnegative matrix factorization techniques.

A mixed-signal paradigm is presented for high-resolution parallel innerproduct computationin very high dimensions, suitable for efficient implementation ofkernels in image processing. At the core of the externally digital architecture is a high-density, low-power analog array performing binary-binary partial matrix-vector multiplication. Full digital resolution is maintained even with low-resolution analog-to-digital conversion, owing torandom statistics in the analog summation of binary products. A random modulation scheme produces near-Bernoulli statistics even for highly correlated inputs. The approach is validated with real image data, and with experimental results from a CID/DRAM analog array prototype in 0.5

We present an algorithm that induces a class of models with thin junction trees--models that are characterized by an upper bound on the size of the maximal cliques of their triangulated graph. By ensuring that the junction tree is thin, inference in our models remains tractable throughout the learning process. This allows both an efficient implementation of an iterative scaling parameter estimation algorithm and also ensures that inference can be performed efficiently with the final model. We illustrate the approach with applications in handwritten digit recognition and DNA splice site detection.

When constructing a classifier, the probability of correct classification offuture data points should be maximized. In the current paper this desideratum is translated in a very direct way into an optimization problem, which is solved using methods from convex optimization.We also show how to exploit Mercer kernels in this setting to obtain nonlinear decision boundaries. A worst-case bound on the probability of misclassification of future data is obtained explicitly. 1 Introduction Consider the problem of choosing a linear discriminant by minimizing the probabilities thatdata vectors fall on the wrong side of the boundary. One way to attempt to achieve this is via a generative approach in which one makes distributional assumptions aboutthe class-conditional densities and thereby estimates and controls the relevant probabilities. The need to make distributional assumptions, however, casts doubt on the generality and validity of such an approach, and in discriminative solutionsto classification problems it is common to attempt to dispense with class-conditional densities entirely.

Statistical learning and probabilistic inference techniques are used to infer thehand position of a subject from multi-electrode recordings of neural activityin motor cortex. First, an array of electrodes provides training dataof neural firing conditioned on hand kinematics. We learn a nonparametric representationof this firing activity using a Bayesian model and rigorously compare it with previous models using cross-validation. Second, we infer a posterior probability distribution over hand motion conditioned on a sequence of neural test data using Bayesian inference. The learned firing models of multiple cells are used to define a non-Gaussian likelihood term which is combined with a prior probability for the kinematics. A particle filtering method is used to represent, update, and propagate the posterior distribution over time. The approach is compared withtraditional linear filtering methods; the results suggest that it may be appropriate for neural prosthetic applications.