Recurrent neural networks of analog units are computers for realvalued functions. We study the time complexity of real computation in general recurrent neural networks. These have sigmoidal, linear, and product units of unlimited order as nodes and no restrictions on the weights. For networks operating in discrete time, we exhibit a family of functions with arbitrarily high complexity, and we derive almost tight bounds on the time required to compute these functions. Thus, evidence is given of the computational limitations that time-bounded analog recurrent neural networks are subject to. 1 Introduction Analog recurrent neural networks are known to have computational capabilities that exceed those of classical Turing machines (see, e.g., Siegelmann and Sontag, 1995; Kilian and Siegelmann, 1996; Siegelmann, 1999).

Recently, Jaakkola and Haussler proposed a method for constructing kernelfunctions from probabilistic models. Their so called "Fisher kernel" has been combined with discriminative classifiers such as SVM and applied successfully in e.g.

In this paper we show that online algorithms for classification and regression canbe naturally used to obtain hypotheses with good datadependent tailbounds on their risk. Our results are proven without requiring complicated concentration-of-measure arguments and they hold for arbitrary online learning algorithms. Furthermore, when applied to concrete online algorithms, our results yield tail bounds that in many cases are comparable or better than the best known bounds.

Competition between actions is based on the motivating characteristics of their consequent states in this sense. Substantial, careful, experiments reviewed in Dickinson & Balleine,12,13 into the neurobiology and psychology ofmotivation shows that this view is incomplete. In many cases, animals are faced with the choice not between many different actionsat a given state, but rather whether a single response isworth executing at all. Evidence suggests that the motivational process underlying this choice has different psychological andneural properties from that underlying action choice. We describe and model these motivational systems, and consider the way they interact.

Hinton Gatsby Unit, UCL London, UK WCIN 3AR hinton@gatsby.ucl.ac.uk Abstract Logistic units in the first hidden layer of a feedforward neural network computethe relative probability of a data point under two Gaussians. This leads us to consider substituting other density models. We present an architecture for performing discriminative learning of Hidden Markov Models using a network of many small HMM's. Experiments on speech data show it to be superior to the standard method of discriminatively training HMM's. 1 Introduction A standard way of performing classification using a generative model is to divide the training cases into their respective classes and then train a set of class conditional models. This unsupervised approach to classification is appealing for two reasons.

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.

This paper presents reinforcement learning with a Long Short Term Memory recurrent neural network: RL-LSTM. Model-free RL-LSTM using Advantage(,x) learning and directed exploration can solve non-Markovian tasks with long-term dependencies between relevantevents. This is demonstrated in a T-maze task, as well as in a difficult variation of the pole balancing task. 1 Introduction Reinforcement learning (RL) is a way of learning how to behave based on delayed reward signals [12]. Among the more important challenges for RL are tasks where part of the state of the environment is hidden from the agent. Such tasks are called non-Markovian tasks or Partially Observable Markov Decision Processes. Many real world tasks have this problem of hidden state. For instance, in a navigation task different positions in the environment may look the same, but one and the same action may lead to different next states or rewards. Thus, hidden state makes RL more realistic.

This paper deals with a neural network architecture which establishes a portfolio management system similar to the Black / Litterman approach. This allocation scheme distributes funds across various securities or financial marketswhile simultaneously complying with specific allocation constraints which meet the requirements of an investor. The portfolio optimization algorithm is modeled by a feedforward neural network. The underlying expected return forecasts are based on error correction neural networks (ECNN), which utilize the last model error as an auxiliary input to evaluate their own misspecification. The portfolio optimization is implemented such that (i.) the allocations comply with investor's constraints and that (ii.) the risk of the portfolio canbe controlled.

We introduce a new type of Self-Organizing Map (SOM) to navigate in the Semantic Space of large text collections. We propose a "hyperbolic SOM"(HSOM) based on a regular tesselation of the hyperbolic plane, which is a non-euclidean space characterized by constant negative gaussian curvature. The exponentially increasing size of a neighborhood around a point in hyperbolic space provides more freedom to map the complex information space arising from language into spatial relations. We describe experiments, showing that the HSOM can successfully be applied to text categorization tasks and yields results comparable to other state-of-the-art methods.

An important issue in applying SVMs to speech recognition is the ability to classify variable length sequences. This paper presents extensions to a standard scheme for handling this variable length data, the Fisher score. A more useful mapping is introduced based on the likelihood-ratio. The score-space defined by this mapping avoids some limitations of the Fisher score. Class-conditional generative modelsare directly incorporated into the definition of the score-space.