Bayesian online algorithms for learning in discrete Hidden Markov Models

We propose and analyze two different Bayesian online algorithms for learning in discrete Hidden Markov Models and compare their performance with the already known Baldi-Chauvin Algorithm. Using the Kullback-Leibler divergence as a measure of generalization we draw learning curves in simplified situations for these algorithms and compare their performances.

HYBRID AND INCREMENTAL FUZZY LEARNING FOR HUMAN SKIN DETECTION

In this paper, a framework for detection of human skin in digital images is proposed. This framework is composed of a training phase and a detection phase. A skin class model is learned during the training phase by processing several training images in a hybrid and incremental fuzzy learning scheme. This scheme combines unsupervised-and supervised-learning: unsupervised, by fuzzy clustering, to obtain clusters of color groups from training images; and supervised to select groups that represent skin color. At the end of the training phase, aggregation operators are used to provide combinations of selected groups into a skin model. In the detection phase, the learned skin model is used to detect human skin in an efficient way. Experimental results show robust and accurate human skin detection performed by the proposed framework.

Stochastic stability analysis for the constant-modulus algorithm

We derive an easy-to-compute approximate bound for the range of step-sizes for which the constant-modulus algorithm (CMA) will remain stable if initialized close to a minimum of the CM cost function. Our model highlights the influence, of the signal constellation used in the transmission system: for smaller variation in the modulus of the transmitted symbols, the algorithm will be more robust, and the steady-state misadjustment will be smaller. The theoretical results are validated through several simulations, for long and short filters and channels.

Learning dependent sources using mixtures of Dirichlet: applications on hyperspectral unmixing

This paper is an elaboration of the DECA algorithm [1] to blindly unmix hyperspectral data. The underlying mixing model is linear, meaning that each pixel is a linear mixture of the endmembers signatures weighted by the correspondent abundance fractions. The proposed method, as DECA, is tailored to highly mixed mixtures in which the geometric based approaches fail to identify the simplex of minimum volume enclosing the observed spectral vectors. We resort then to a statitistical framework, where the abundance fractions are modeled as mixtures of Dirichlet densities, thus enforcing the constraints on abundance fractions imposed by the acquisition process, namely non-negativity and constant sum. With respect to DECA, we introduce two improvements: 1) the number of Dirichlet modes are inferred based on the minimum description length (MDL) principle; 2) The generalized expectation maximization (GEM) algorithm we adopt to infer the model parameters is improved by using alternating minimization and augmented Lagrangian methods to compute the mixing matrix. The effectiveness of the proposed algorithm is illustrated with simulated and read data.

Contextual intelligent load management with ANN adaptive learning module

With the current increase of energy resources prices and environmental concerns intelligent load management systems are gaining more and more importance. This paper concerns a SCADA House Intelligent Management (SHIM) system that includes an optimization module using deterministic and genetic algorithm approaches. SHIM undertakes contextual load management based on the characterization of each situation. SHIM considers available generation resources, load demand, supplier/market electricity price, and consumers’ constraints and preferences. The paper focus on the recently developed learning module which is based on artificial neural networks (ANN). The learning module allows the adjustment of users’ profiles along SHIM lifetime. A case study considering a system with fourteen discrete and four variable loads managed by a SHIM system during five consecutive similar weekends is presented.

Machine Learning Algorithms at Myocardial Perfusion Imaging - a Preliminary Study

Introduction: A major focus of data mining process - especially machine learning researches - is to automatically learn to recognize complex patterns and help to take the adequate decisions strictly based on the acquired data. Since imaging techniques like MPI – Myocardial Perfusion Imaging on Nuclear Cardiology, can implicate a huge part of the daily workflow and generate gigabytes of data, there could be advantages on Computerized Analysis of data over Human Analysis: shorter time, homogeneity and consistency, automatic recording of analysis results, relatively inexpensive, etc.Objectives: The aim of this study relates with the evaluation of the efficacy of this methodology on the evaluation of MPI Stress studies and the process of decision taking concerning the continuation – or not – of the evaluation of each patient. It has been pursued has an objective to automatically classify a patient test in one of three groups: “Positive”, “Negative” and “Indeterminate”. “Positive” would directly follow to the Rest test part of the exam, the “Negative” would be directly exempted from continuation and only the “Indeterminate” group would deserve the clinician analysis, so allowing economy of clinician’s effort, increasing workflow fluidity at the technologist’s level and probably sparing time to patients. Methods: WEKA v3.6.2 open source software was used to make a comparative analysis of three WEKA algorithms (“OneR”, “J48” and “Naïve Bayes”) - on a retrospective study using the comparison with correspondent clinical results as reference, signed by nuclear cardiologist experts - on “SPECT Heart Dataset”, available on University of California – Irvine, at the Machine Learning Repository. For evaluation purposes, criteria as “Precision”, “Incorrectly Classified Instances” and “Receiver Operating Characteristics (ROC) Areas” were considered. Results: The interpretation of the data suggests that the Naïve Bayes algorithm has the best performance among the three previously selected algorithms. Conclusions: It is believed - and apparently supported by the findings - that machine learning algorithms could significantly assist, at an intermediary level, on the analysis of scintigraphic data obtained on MPI, namely after Stress acquisition, so eventually increasing efficiency of the entire system and potentially easing both roles of Technologists and Nuclear Cardiologists. In the actual continuation of this study, it is planned to use more patient information and significantly increase the population under study, in order to allow improving system accuracy.

Algorithm for the Parkinson’s disease behavioural models characterization using a biosensor

Dissertação para obtenção do Grau de Mestre em Engenharia Biomédica

Learning domain-specific sentiment lexicons with applications to recommender systems

Search is now going beyond looking for factual information, and people wish to search for the opinions of others to help them in their own decision-making. Sentiment expressions or opinion expressions are used by users to express their opinion and embody important pieces of information, particularly in online commerce. The main problem that the present dissertation addresses is how to model text to find meaningful words that express a sentiment. In this context, I investigate the viability of automatically generating a sentiment lexicon for opinion retrieval and sentiment classification applications. For this research objective we propose to capture sentiment words that are derived from online users’ reviews. In this approach, we tackle a major challenge in sentiment analysis which is the detection of words that express subjective preference and domain-specific sentiment words such as jargon. To this aim we present a fully generative method that automatically learns a domain-specific lexicon and is fully independent of external sources. Sentiment lexicons can be applied in a broad set of applications, however popular recommendation algorithms have somehow been disconnected from sentiment analysis. Therefore, we present a study that explores the viability of applying sentiment analysis techniques to infer ratings in a recommendation algorithm. Furthermore, entities’ reputation is intrinsically associated with sentiment words that have a positive or negative relation with those entities. Hence, is provided a study that observes the viability of using a domain-specific lexicon to compute entities reputation. Finally, a recommendation system algorithm is improved with the use of sentiment-based ratings and entities reputation.

Dynamic group formation in mobile computer supported collaborative learning environment

Forming suitable learning groups is one of the factors that determine the efficiency of collaborative learning activities. However, only a few studies were carried out to address this problem in the mobile learning environments. In this paper, we propose a new approach for an automatic, customized, and dynamic group formation in Mobile Computer Supported Collaborative Learning (MCSCL) contexts. The proposed solution is based on the combination of three types of grouping criteria: learner’s personal characteristics, learner’s behaviours, and context information. The instructors can freely select the type, the number, and the weight of grouping criteria, together with other settings such as the number, the size, and the type of learning groups (homogeneous or heterogeneous). Apart from a grouping mechanism, the proposed approach represents a flexible tool to control each learner, and to manage the learning processes from the beginning to the end of collaborative learning activities. In order to evaluate the quality of the implemented group formation algorithm, we compare its Average Intra-cluster Distance (AID) with the one of a random group formation method. The results show a higher effectiveness of the proposed algorithm in forming homogenous and heterogeneous groups compared to the random method.

A Survey of Active Learning Algorithms for Supervised Remote Sensing Image Classification

Defining an efficient training set is one of the most delicate phases for the success of remote sensing image classification routines. The complexity of the problem, the limited temporal and financial resources, as well as the high intraclass variance can make an algorithm fail if it is trained with a suboptimal dataset. Active learning aims at building efficient training sets by iteratively improving the model performance through sampling. A user-defined heuristic ranks the unlabeled pixels according to a function of the uncertainty of their class membership and then the user is asked to provide labels for the most uncertain pixels. This paper reviews and tests the main families of active learning algorithms: committee, large margin, and posterior probability-based. For each of them, the most recent advances in the remote sensing community are discussed and some heuristics are detailed and tested. Several challenging remote sensing scenarios are considered, including very high spatial resolution and hyperspectral image classification. Finally, guidelines for choosing the good architecture are provided for new and/or unexperienced user.

Active Learning of Very-High Resolution Optical Imagery with SVM: Entropy vs Margin Sampling

An active learning method is proposed for the semi-automatic selection of training sets in remote sensing image classification. The method adds iteratively to the current training set the unlabeled pixels for which the prediction of an ensemble of classifiers based on bagged training sets show maximum entropy. This way, the algorithm selects the pixels that are the most uncertain and that will improve the model if added in the training set. The user is asked to label such pixels at each iteration. Experiments using support vector machines (SVM) on an 8 classes QuickBird image show the excellent performances of the methods, that equals accuracies of both a model trained with ten times more pixels and a model whose training set has been built using a state-of-the-art SVM specific active learning method

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