Learning Linear, Sparse, Factorial Codes

In previous work (Olshausen & Field 1996), an algorithm was described for learning linear sparse codes which, when trained on natural images, produces a set of basis functions that are spatially localized, oriented, and bandpass (i.e., wavelet-like). This note shows how the algorithm may be interpreted within a maximum-likelihood framework. Several useful insights emerge from this connection: it makes explicit the relation to statistical independence (i.e., factorial coding), it shows a formal relationship to the algorithm of Bell and Sejnowski (1995), and it suggests how to adapt parameters that were previously fixed.

Fast Learning by Bounding Likelihoods in Sigmoid Type Belief Networks

Sigmoid type belief networks, a class of probabilistic neural networks, provide a natural framework for compactly representing probabilistic information in a variety of unsupervised and supervised learning problems. Often the parameters used in these networks need to be learned from examples. Unfortunately, estimating the parameters via exact probabilistic calculations (i.e, the EM-algorithm) is intractable even for networks with fairly small numbers of hidden units. We propose to avoid the infeasibility of the E step by bounding likelihoods instead of computing them exactly. We introduce extended and complementary representations for these networks and show that the estimation of the network parameters can be made fast (reduced to quadratic optimization) by performing the estimation in either of the alternative domains. The complementary networks can be used for continuous density estimation as well.

On Convergence Properties of the EM Algorithm for Gaussian Mixtures

"Expectation-Maximization'' (EM) algorithm and gradient-based approaches for maximum likelihood learning of finite Gaussian mixtures. We show that the EM step in parameter space is obtained from the gradient via a projection matrix $P$, and we provide an explicit expression for the matrix. We then analyze the convergence of EM in terms of special properties of $P$ and provide new results analyzing the effect that $P$ has on the likelihood surface. Based on these mathematical results, we present a comparative discussion of the advantages and disadvantages of EM and other algorithms for the learning of Gaussian mixture models.

Learning from Incomplete Data

Real-world learning tasks often involve high-dimensional data sets with complex patterns of missing features. In this paper we review the problem of learning from incomplete data from two statistical perspectives---the likelihood-based and the Bayesian. The goal is two-fold: to place current neural network approaches to missing data within a statistical framework, and to describe a set of algorithms, derived from the likelihood-based framework, that handle clustering, classification, and function approximation from incomplete data in a principled and efficient manner. These algorithms are based on mixture modeling and make two distinct appeals to the Expectation-Maximization (EM) principle (Dempster, Laird, and Rubin 1977)---both for the estimation of mixture components and for coping with the missing data.

Learning and adaptation in physical heterogeneous teams of robots

In this paper we present a novel approach to assigning roles to robots in a team of physical heterogeneous robots. Its members compete for these roles and get rewards for them. The rewards are used to determine each agent’s preferences and which agents are better adapted to the environment. These aspects are included in the decision making process. Agent interactions are modelled using the concept of an ecosystem in which each robot is a species, resulting in emergent behaviour of the whole set of agents. One of the most important features of this approach is its high adaptability. Unlike some other learning techniques, this approach does not need to start a whole exploitation process when the environment changes. All this is exemplified by means of experiments run on a simulator. In addition, the algorithm developed was applied as applied to several teams of robots in order to analyse the impact of heterogeneity in these systems

Efficient learning of reactive robot behaviors with a Neural-Q_learning approach

The purpose of this paper is to propose a Neural-Q_learning approach designed for online learning of simple and reactive robot behaviors. In this approach, the Q_function is generalized by a multi-layer neural network allowing the use of continuous states and actions. The algorithm uses a database of the most recent learning samples to accelerate and guarantee the convergence. Each Neural-Q_learning function represents an independent, reactive and adaptive behavior which maps sensorial states to robot control actions. A group of these behaviors constitutes a reactive control scheme designed to fulfill simple missions. The paper centers on the description of the Neural-Q_learning based behaviors showing their performance with an underwater robot in a target following task. Real experiments demonstrate the convergence and stability of the learning system, pointing out its suitability for online robot learning. Advantages and limitations are discussed

Semi-online neural-Q_leaming for real-time robot learning

Reinforcement learning (RL) is a very suitable technique for robot learning, as it can learn in unknown environments and in real-time computation. The main difficulties in adapting classic RL algorithms to robotic systems are the generalization problem and the correct observation of the Markovian state. This paper attempts to solve the generalization problem by proposing the semi-online neural-Q_learning algorithm (SONQL). The algorithm uses the classic Q_learning technique with two modifications. First, a neural network (NN) approximates the Q_function allowing the use of continuous states and actions. Second, a database of the most representative learning samples accelerates and stabilizes the convergence. The term semi-online is referred to the fact that the algorithm uses the current but also past learning samples. However, the algorithm is able to learn in real-time while the robot is interacting with the environment. The paper shows simulated results with the "mountain-car" benchmark and, also, real results with an underwater robot in a target following behavior

A behavior-based scheme using reinforcement learning for autonomous underwater vehicles

This paper presents a hybrid behavior-based scheme using reinforcement learning for high-level control of autonomous underwater vehicles (AUVs). Two main features of the presented approach are hybrid behavior coordination and semi on-line neural-Q_learning (SONQL). Hybrid behavior coordination takes advantages of robustness and modularity in the competitive approach as well as efficient trajectories in the cooperative approach. SONQL, a new continuous approach of the Q_learning algorithm with a multilayer neural network is used to learn behavior state/action mapping online. Experimental results show the feasibility of the presented approach for AUVs

Towards Direct Policy Search Reinforcement Learning for Robot Control

This paper proposes a high-level reinforcement learning (RL) control system for solving the action selection problem of an autonomous robot. Although the dominant approach, when using RL, has been to apply value function based algorithms, the system here detailed is characterized by the use of direct policy search methods. Rather than approximating a value function, these methodologies approximate a policy using an independent function approximator with its own parameters, trying to maximize the future expected reward. The policy based algorithm presented in this paper is used for learning the internal state/action mapping of a behavior. In this preliminary work, we demonstrate its feasibility with simulated experiments using the underwater robot GARBI in a target reaching task

Predicting sense of community and participation by applying machine learning to open government data

Community capacity is used to monitor socio-economic development. It is composed of a number of dimensions, which can be measured to understand the possible issues in the implementation of a policy or the outcome of a project targeting a community. Measuring community capacity dimensions is usually expensive and time consuming, requiring locally organised surveys. Therefore, we investigate a technique to estimate them by applying the Random Forests algorithm on secondary open government data. This research focuses on the prediction of measures for two dimensions: sense of community and participation. The most important variables for this prediction were determined. The variables included in the datasets used to train the predictive models complied with two criteria: nationwide availability; sufficiently fine-grained geographic breakdown, i.e. neighbourhood level. The models explained 77% of the sense of community measures and 63% of participation. Due to the low geographic detail of the outcome measures available, further research is required to apply the predictive models to a neighbourhood level. The variables that were found to be more determinant for prediction were only partially in agreement with the factors that, according to the social science literature consulted, are the most influential for sense of community and participation. This finding should be further investigated from a social science perspective, in order to be understood in depth.

An artificial economy based on reinforcement learning and agent based modeling

In this paper, we employ techniques from artificial intelligence such as reinforcement learning and agent based modeling as building blocks of a computational model for an economy based on conventions. First we model the interaction among firms in the private sector. These firms behave in an information environment based on conventions, meaning that a firm is likely to behave as its neighbors if it observes that their actions lead to a good pay off. On the other hand, we propose the use of reinforcement learning as a computational model for the role of the government in the economy, as the agent that determines the fiscal policy, and whose objective is to maximize the growth of the economy. We present the implementation of a simulator of the proposed model based on SWARM, that employs the SARSA(λ) algorithm combined with a multilayer perceptron as the function approximation for the action value function.

Unsupervised genetic algorithm deployed for intrusion detection

This paper represents the first step in an on-going work for designing an unsupervised method based on genetic algorithm for intrusion detection. Its main role in a broader system is to notify of an unusual traffic and in that way provide the possibility of detecting unknown attacks. Most of the machine-learning techniques deployed for intrusion detection are supervised as these techniques are generally more accurate, but this implies the need of labeling the data for training and testing which is time-consuming and error-prone. Hence, our goal is to devise an anomaly detector which would be unsupervised, but at the same time robust and accurate. Genetic algorithms are robust and able to avoid getting stuck in local optima, unlike the rest of clustering techniques. The model is verified on KDD99 benchmark dataset, generating a solution competitive with the solutions of the state-of-the-art which demonstrates high possibilities of the proposed method.

An on-line algorithm of uncertain time delay estimation in a continuous system

In this paper, we present an on-line estimation algorithm for an uncertain time delay in a continuous system based on the observational input-output data, subject to observational noise. The first order Pade approximation is used to approximate the time delay. At each time step, the algorithm combines the well known Kalman filter algorithm and the recursive instrumental variable least squares (RIVLS) algorithm in cascade form. The instrumental variable least squares algorithm is used in order to achieve the consistency of the delay parameter estimate, since an error-in-the-variable model is involved. An illustrative example is utilized to demonstrate the efficacy of the proposed approach.

Example uses of a learning object

A self study course for learning to program using the C programming language has been developed. A Learning Object approach was used in the design of the course. One of the benefits of the Learning Object approach is that the learning material can be reused for different purposes. 'Me course developed is designed so that learners can choose the pedagogical approach most suited to their personal learning requirements. For all learning approaches a set of common Assessment Learning Objects (ALOs or tests) have been created. The design of formative assessments with ALOs can be carried out by the Instructional Designer grouping ALOs to correspond to a specific assessment intention. The course is non-credit earning, so there is no summative assessment, all assessment is formative. In this paper examples of ALOs and their uses is presented together with their uses as decided by the Instructional Designer and learner. Personalisation of the formative assessment of skills can be decided by the Instructional Designer or the learner using a repository of pre-designed ALOs. The process of combining ALOs can be carried out manually or in a semi-automated way using metadata that describes the ALO and the skill it is designed to assess.