Bandit-Based Algorithms for Budgeted Learning

We explore the problem of budgeted machine learning, in which the learning algorithm has free access to the training examples’ labels but has to pay for each attribute that is specified. This learning model is appropriate in many areas, including medical applications. We present new algorithms for choosing which attributes to purchase of which examples in the budgeted learning model based on algorithms for the multi-armed bandit problem. All of our approaches outperformed the current state of the art. Furthermore, we present a new means for selecting an example to purchase after the attribute is selected, instead of selecting an example uniformly at random, which is typically done. Our new example selection method improved performance of all the algorithms we tested, both ours and those in the literature.

Critical behavior in a cross-situational lexicon learning scenario

The associationist account for early word learning is based on the co-occurrence between referents and words. Here we introduce a noisy cross-situational learning scenario in which the referent of the uttered word is eliminated from the context with probability gamma, thus modeling the noise produced by out-of-context words. We examine the performance of a simple associative learning algorithm and find a critical value of the noise parameter gamma(c) above which learning is impossible. We use finite-size scaling to show that the sharpness of the transition persists across a region of order tau(-1/2) about gamma(c), where tau is the number of learning trials, as well as to obtain the learning error (scaling function) in the critical region. In addition, we show that the distribution of durations of periods when the learning error is zero is a power law with exponent -3/2 at the critical point. Copyright (C) EPLA, 2012

Control design and genetic algorithm optimization for electrostatic MEMS

This dissertation discusses structural-electrostatic modeling techniques, genetic algorithm based optimization and control design for electrostatic micro devices. First, an alternative modeling technique, the interpolated force model, for electrostatic micro devices is discussed. The method provides improved computational efficiency relative to a benchmark model, as well as improved accuracy for irregular electrode configurations relative to a common approximate model, the parallel plate approximation model. For the configuration most similar to two parallel plates, expected to be the best case scenario for the approximate model, both the parallel plate approximation model and the interpolated force model maintained less than 2.2% error in static deflection compared to the benchmark model. For the configuration expected to be the worst case scenario for the parallel plate approximation model, the interpolated force model maintained less than 2.9% error in static deflection while the parallel plate approximation model is incapable of handling the configuration. Second, genetic algorithm based optimization is shown to improve the design of an electrostatic micro sensor. The design space is enlarged from published design spaces to include the configuration of both sensing and actuation electrodes, material distribution, actuation voltage and other geometric dimensions. For a small population, the design was improved by approximately a factor of 6 over 15 generations to a fitness value of 3.2 fF. For a larger population seeded with the best configurations of the previous optimization, the design was improved by another 7% in 5 generations to a fitness value of 3.0 fF. Third, a learning control algorithm is presented that reduces the closing time of a radiofrequency microelectromechanical systems switch by minimizing bounce while maintaining robustness to fabrication variability. Electrostatic actuation of the plate causes pull-in with high impact velocities, which are difficult to control due to parameter variations from part to part. A single degree-of-freedom model was utilized to design a learning control algorithm that shapes the actuation voltage based on the open/closed state of the switch. Experiments on 3 test switches show that after 5-10 iterations, the learning algorithm lands the switch with an impact velocity not exceeding 0.2 m/s, eliminating bounce.

Hybrid Approach Combining Machine Learning and a Rule-Based Expert System for Text Categorization

This paper discusses a novel hybrid approach for text categorization that combines a machine learning algorithm, which provides a base model trained with a labeled corpus, with a rule-based expert system, which is used to improve the results provided by the previous classifier, by filtering false positives and dealing with false negatives. The main advantage is that the system can be easily fine-tuned by adding specific rules for those noisy or conflicting categories that have not been successfully trained. We also describe an implementation based on k-Nearest Neighbor and a simple rule language to express lists of positive, negative and relevant (multiword) terms appearing in the input text. The system is evaluated in several scenarios, including the popular Reuters-21578 news corpus for comparison to other approaches, and categorization using IPTC metadata, EUROVOC thesaurus and others. Results show that this approach achieves a precision that is comparable to top ranked methods, with the added value that it does not require a demanding human expert workload to train

Contributions to Bayesian network learning with applications to neuroscience

Neuronal morphology is a key feature in the study of brain circuits, as it is highly related to information processing and functional identification. Neuronal morphology affects the process of integration of inputs from other neurons and determines the neurons which receive the output of the neurons. Different parts of the neurons can operate semi-independently according to the spatial location of the synaptic connections. As a result, there is considerable interest in the analysis of the microanatomy of nervous cells since it constitutes an excellent tool for better understanding cortical function. However, the morphologies, molecular features and electrophysiological properties of neuronal cells are extremely variable. Except for some special cases, this variability makes it hard to find a set of features that unambiguously define a neuronal type. In addition, there are distinct types of neurons in particular regions of the brain. This morphological variability makes the analysis and modeling of neuronal morphology a challenge. Uncertainty is a key feature in many complex real-world problems. Probability theory provides a framework for modeling and reasoning with uncertainty. Probabilistic graphical models combine statistical theory and graph theory to provide a tool for managing domains with uncertainty. In particular, we focus on Bayesian networks, the most commonly used probabilistic graphical model. In this dissertation, we design new methods for learning Bayesian networks and apply them to the problem of modeling and analyzing morphological data from neurons. The morphology of a neuron can be quantified using a number of measurements, e.g., the length of the dendrites and the axon, the number of bifurcations, the direction of the dendrites and the axon, etc. These measurements can be modeled as discrete or continuous data. The continuous data can be linear (e.g., the length or the width of a dendrite) or directional (e.g., the direction of the axon). These data may follow complex probability distributions and may not fit any known parametric distribution. Modeling this kind of problems using hybrid Bayesian networks with discrete, linear and directional variables poses a number of challenges regarding learning from data, inference, etc. In this dissertation, we propose a method for modeling and simulating basal dendritic trees from pyramidal neurons using Bayesian networks to capture the interactions between the variables in the problem domain. A complete set of variables is measured from the dendrites, and a learning algorithm is applied to find the structure and estimate the parameters of the probability distributions included in the Bayesian networks. Then, a simulation algorithm is used to build the virtual dendrites by sampling values from the Bayesian networks, and a thorough evaluation is performed to show the model’s ability to generate realistic dendrites. In this first approach, the variables are discretized so that discrete Bayesian networks can be learned and simulated. Then, we address the problem of learning hybrid Bayesian networks with different kinds of variables. Mixtures of polynomials have been proposed as a way of representing probability densities in hybrid Bayesian networks. We present a method for learning mixtures of polynomials approximations of one-dimensional, multidimensional and conditional probability densities from data. The method is based on basis spline interpolation, where a density is approximated as a linear combination of basis splines. The proposed algorithms are evaluated using artificial datasets. We also use the proposed methods as a non-parametric density estimation technique in Bayesian network classifiers. Next, we address the problem of including directional data in Bayesian networks. These data have some special properties that rule out the use of classical statistics. Therefore, different distributions and statistics, such as the univariate von Mises and the multivariate von Mises–Fisher distributions, should be used to deal with this kind of information. In particular, we extend the naive Bayes classifier to the case where the conditional probability distributions of the predictive variables given the class follow either of these distributions. We consider the simple scenario, where only directional predictive variables are used, and the hybrid case, where discrete, Gaussian and directional distributions are mixed. The classifier decision functions and their decision surfaces are studied at length. Artificial examples are used to illustrate the behavior of the classifiers. The proposed classifiers are empirically evaluated over real datasets. We also study the problem of interneuron classification. An extensive group of experts is asked to classify a set of neurons according to their most prominent anatomical features. A web application is developed to retrieve the experts’ classifications. We compute agreement measures to analyze the consensus between the experts when classifying the neurons. Using Bayesian networks and clustering algorithms on the resulting data, we investigate the suitability of the anatomical terms and neuron types commonly used in the literature. Additionally, we apply supervised learning approaches to automatically classify interneurons using the values of their morphological measurements. Then, a methodology for building a model which captures the opinions of all the experts is presented. First, one Bayesian network is learned for each expert, and we propose an algorithm for clustering Bayesian networks corresponding to experts with similar behaviors. Then, a Bayesian network which represents the opinions of each group of experts is induced. Finally, a consensus Bayesian multinet which models the opinions of the whole group of experts is built. A thorough analysis of the consensus model identifies different behaviors between the experts when classifying the interneurons in the experiment. A set of characterizing morphological traits for the neuronal types can be defined by performing inference in the Bayesian multinet. These findings are used to validate the model and to gain some insights into neuron morphology. Finally, we study a classification problem where the true class label of the training instances is not known. Instead, a set of class labels is available for each instance. This is inspired by the neuron classification problem, where a group of experts is asked to individually provide a class label for each instance. We propose a novel approach for learning Bayesian networks using count vectors which represent the number of experts who selected each class label for each instance. These Bayesian networks are evaluated using artificial datasets from supervised learning problems. Resumen La morfología neuronal es una característica clave en el estudio de los circuitos cerebrales, ya que está altamente relacionada con el procesado de información y con los roles funcionales. La morfología neuronal afecta al proceso de integración de las señales de entrada y determina las neuronas que reciben las salidas de otras neuronas. Las diferentes partes de la neurona pueden operar de forma semi-independiente de acuerdo a la localización espacial de las conexiones sinápticas. Por tanto, existe un interés considerable en el análisis de la microanatomía de las células nerviosas, ya que constituye una excelente herramienta para comprender mejor el funcionamiento de la corteza cerebral. Sin embargo, las propiedades morfológicas, moleculares y electrofisiológicas de las células neuronales son extremadamente variables. Excepto en algunos casos especiales, esta variabilidad morfológica dificulta la definición de un conjunto de características que distingan claramente un tipo neuronal. Además, existen diferentes tipos de neuronas en regiones particulares del cerebro. La variabilidad neuronal hace que el análisis y el modelado de la morfología neuronal sean un importante reto científico. La incertidumbre es una propiedad clave en muchos problemas reales. La teoría de la probabilidad proporciona un marco para modelar y razonar bajo incertidumbre. Los modelos gráficos probabilísticos combinan la teoría estadística y la teoría de grafos con el objetivo de proporcionar una herramienta con la que trabajar bajo incertidumbre. En particular, nos centraremos en las redes bayesianas, el modelo más utilizado dentro de los modelos gráficos probabilísticos. En esta tesis hemos diseñado nuevos métodos para aprender redes bayesianas, inspirados por y aplicados al problema del modelado y análisis de datos morfológicos de neuronas. La morfología de una neurona puede ser cuantificada usando una serie de medidas, por ejemplo, la longitud de las dendritas y el axón, el número de bifurcaciones, la dirección de las dendritas y el axón, etc. Estas medidas pueden ser modeladas como datos continuos o discretos. A su vez, los datos continuos pueden ser lineales (por ejemplo, la longitud o la anchura de una dendrita) o direccionales (por ejemplo, la dirección del axón). Estos datos pueden llegar a seguir distribuciones de probabilidad muy complejas y pueden no ajustarse a ninguna distribución paramétrica conocida. El modelado de este tipo de problemas con redes bayesianas híbridas incluyendo variables discretas, lineales y direccionales presenta una serie de retos en relación al aprendizaje a partir de datos, la inferencia, etc. En esta tesis se propone un método para modelar y simular árboles dendríticos basales de neuronas piramidales usando redes bayesianas para capturar las interacciones entre las variables del problema. Para ello, se mide un amplio conjunto de variables de las dendritas y se aplica un algoritmo de aprendizaje con el que se aprende la estructura y se estiman los parámetros de las distribuciones de probabilidad que constituyen las redes bayesianas. Después, se usa un algoritmo de simulación para construir dendritas virtuales mediante el muestreo de valores de las redes bayesianas. Finalmente, se lleva a cabo una profunda evaluaci ón para verificar la capacidad del modelo a la hora de generar dendritas realistas. En esta primera aproximación, las variables fueron discretizadas para poder aprender y muestrear las redes bayesianas. A continuación, se aborda el problema del aprendizaje de redes bayesianas con diferentes tipos de variables. Las mixturas de polinomios constituyen un método para representar densidades de probabilidad en redes bayesianas híbridas. Presentamos un método para aprender aproximaciones de densidades unidimensionales, multidimensionales y condicionales a partir de datos utilizando mixturas de polinomios. El método se basa en interpolación con splines, que aproxima una densidad como una combinación lineal de splines. Los algoritmos propuestos se evalúan utilizando bases de datos artificiales. Además, las mixturas de polinomios son utilizadas como un método no paramétrico de estimación de densidades para clasificadores basados en redes bayesianas. Después, se estudia el problema de incluir información direccional en redes bayesianas. Este tipo de datos presenta una serie de características especiales que impiden el uso de las técnicas estadísticas clásicas. Por ello, para manejar este tipo de información se deben usar estadísticos y distribuciones de probabilidad específicos, como la distribución univariante von Mises y la distribución multivariante von Mises–Fisher. En concreto, en esta tesis extendemos el clasificador naive Bayes al caso en el que las distribuciones de probabilidad condicionada de las variables predictoras dada la clase siguen alguna de estas distribuciones. Se estudia el caso base, en el que sólo se utilizan variables direccionales, y el caso híbrido, en el que variables discretas, lineales y direccionales aparecen mezcladas. También se estudian los clasificadores desde un punto de vista teórico, derivando sus funciones de decisión y las superficies de decisión asociadas. El comportamiento de los clasificadores se ilustra utilizando bases de datos artificiales. Además, los clasificadores son evaluados empíricamente utilizando bases de datos reales. También se estudia el problema de la clasificación de interneuronas. Desarrollamos una aplicación web que permite a un grupo de expertos clasificar un conjunto de neuronas de acuerdo a sus características morfológicas más destacadas. Se utilizan medidas de concordancia para analizar el consenso entre los expertos a la hora de clasificar las neuronas. Se investiga la idoneidad de los términos anatómicos y de los tipos neuronales utilizados frecuentemente en la literatura a través del análisis de redes bayesianas y la aplicación de algoritmos de clustering. Además, se aplican técnicas de aprendizaje supervisado con el objetivo de clasificar de forma automática las interneuronas a partir de sus valores morfológicos. A continuación, se presenta una metodología para construir un modelo que captura las opiniones de todos los expertos. Primero, se genera una red bayesiana para cada experto y se propone un algoritmo para agrupar las redes bayesianas que se corresponden con expertos con comportamientos similares. Después, se induce una red bayesiana que modela la opinión de cada grupo de expertos. Por último, se construye una multired bayesiana que modela las opiniones del conjunto completo de expertos. El análisis del modelo consensuado permite identificar diferentes comportamientos entre los expertos a la hora de clasificar las neuronas. Además, permite extraer un conjunto de características morfológicas relevantes para cada uno de los tipos neuronales mediante inferencia con la multired bayesiana. Estos descubrimientos se utilizan para validar el modelo y constituyen información relevante acerca de la morfología neuronal. Por último, se estudia un problema de clasificación en el que la etiqueta de clase de los datos de entrenamiento es incierta. En cambio, disponemos de un conjunto de etiquetas para cada instancia. Este problema está inspirado en el problema de la clasificación de neuronas, en el que un grupo de expertos proporciona una etiqueta de clase para cada instancia de manera individual. Se propone un método para aprender redes bayesianas utilizando vectores de cuentas, que representan el número de expertos que seleccionan cada etiqueta de clase para cada instancia. Estas redes bayesianas se evalúan utilizando bases de datos artificiales de problemas de aprendizaje supervisado.

Arquitectura de Control para la Conducción Autónoma de Vehículos en Entornos Urbanos y Autovías

Hoy en día, el desarrollo tecnológico en el campo de los sistemas inteligentes de transporte (ITS por sus siglas en inglés) ha permitido dotar a los vehículos con diversos sistemas de ayuda a la conducción (ADAS, del inglés advanced driver assistance system), mejorando la experiencia y seguridad de los pasajeros, en especial del conductor. La mayor parte de estos sistemas están pensados para advertir al conductor sobre ciertas situaciones de riesgo, como la salida involuntaria del carril o la proximidad de obstáculos en el camino. No obstante, también podemos encontrar sistemas que van un paso más allá y son capaces de cooperar con el conductor en el control del vehículo o incluso relegarlos de algunas tareas tediosas. Es en este último grupo donde se encuentran los sistemas de control electrónico de estabilidad (ESP - Electronic Stability Program), el antibloqueo de frenos (ABS - Anti-lock Braking System), el control de crucero (CC - Cruise Control) y los más recientes sistemas de aparcamiento asistido. Continuando con esta línea de desarrollo, el paso siguiente consiste en la supresión del conductor humano, desarrollando sistemas que sean capaces de conducir un vehículo de forma autónoma y con un rendimiento superior al del conductor. En este trabajo se presenta, en primer lugar, una arquitectura de control para la automatización de vehículos. Esta se compone de distintos componentes de hardware y software, agrupados de acuerdo a su función principal. El diseño de la arquitectura parte del trabajo previo desarrollado por el Programa AUTOPIA, aunque introduce notables aportaciones en cuanto a la eficiencia, robustez y escalabilidad del sistema. Ahondando un poco más en detalle, debemos resaltar el desarrollo de un algoritmo de localización basado en enjambres de partículas. Este está planteado como un método de filtrado y fusión de la información obtenida a partir de los distintos sensores embarcados en el vehículo, entre los que encontramos un receptor GPS (Global Positioning System), unidades de medición inercial (IMU – Inertial Measurement Unit) e información tomada directamente de los sensores embarcados por el fabricante, como la velocidad de las ruedas y posición del volante. Gracias a este método se ha conseguido resolver el problema de la localización, indispensable para el desarrollo de sistemas de conducción autónoma. Continuando con el trabajo de investigación, se ha estudiado la viabilidad de la aplicación de técnicas de aprendizaje y adaptación al diseño de controladores para el vehículo. Como punto de partida se emplea el método de Q-learning para la generación de un controlador borroso lateral sin ningún tipo de conocimiento previo. Posteriormente se presenta un método de ajuste on-line para la adaptación del control longitudinal ante perturbaciones impredecibles del entorno, como lo son los cambios en la inclinación del camino, fricción de las ruedas o peso de los ocupantes. Para finalizar, se presentan los resultados obtenidos durante un experimento de conducción autónoma en carreteras reales, el cual se llevó a cabo en el mes de Junio de 2012 desde la población de San Lorenzo de El Escorial hasta las instalaciones del Centro de Automática y Robótica (CAR) en Arganda del Rey. El principal objetivo tras esta demostración fue validar el funcionamiento, robustez y capacidad de la arquitectura propuesta para afrontar el problema de la conducción autónoma, bajo condiciones mucho más reales a las que se pueden alcanzar en las instalaciones de prueba. ABSTRACT Nowadays, the technological advances in the Intelligent Transportation Systems (ITS) field have led the development of several driving assistance systems (ADAS). These solutions are designed to improve the experience and security of all the passengers, especially the driver. For most of these systems, the main goal is to warn drivers about unexpected circumstances leading to risk situations such as involuntary lane departure or proximity to other vehicles. However, other ADAS go a step further, being able to cooperate with the driver in the control of the vehicle, or even overriding it on some tasks. Examples of this kind of systems are the anti-lock braking system (ABS), cruise control (CC) and the recently commercialised assisted parking systems. Within this research line, the next step is the development of systems able to replace the human drivers, improving the control and therefore, the safety and reliability of the vehicles. First of all, this dissertation presents a control architecture design for autonomous driving. It is made up of several hardware and software components, grouped according to their main function. The design of this architecture is based on the previous works carried out by the AUTOPIA Program, although notable improvements have been made regarding the efficiency, robustness and scalability of the system. It is also remarkable the work made on the development of a location algorithm for vehicles. The proposal is based on the emulation of the behaviour of biological swarms and its performance is similar to the well-known particle filters. The developed method combines information obtained from different sensors, including GPS, inertial measurement unit (IMU), and data from the original vehicle’s sensors on-board. Through this filtering algorithm the localization problem is properly managed, which is critical for the development of autonomous driving systems. The work deals also with the fuzzy control tuning system, a very time consuming task when done manually. An analysis of learning and adaptation techniques for the development of different controllers has been made. First, the Q-learning –a reinforcement learning method– has been applied to the generation of a lateral fuzzy controller from scratch. Subsequently, the development of an adaptation method for longitudinal control is presented. With this proposal, a final cruise control controller is able to deal with unpredictable environment disturbances, such as road slope, wheel’s friction or even occupants’ weight. As a testbed for the system, an autonomous driving experiment on real roads is presented. This experiment was carried out on June 2012, driving from San Lorenzo de El Escorial up to the Center for Automation and Robotics (CAR) facilities in Arganda del Rey. The main goal of the demonstration was validating the performance, robustness and viability of the proposed architecture to deal with the problem of autonomous driving under more demanding conditions than those achieved on closed test tracks.

Técnicas de machine learning para la detección de la negación en textos clínicos en español

En los últimos años han surgido nuevos campos de las tecnologías de la información que exploran el tratamiento de la gran cantidad de datos digitales existentes y cómo transformarlos en conocimiento explícito. Las técnicas de Procesamiento del Lenguaje Natural (NLP) son capaces de extraer información de los textos digitales presentados en forma narrativa. Además, las técnicas de machine learning clasifican instancias o ejemplos en función de sus atributos, en distintas categorías, aprendiendo de otros previamente clasificados. Los textos clínicos son una gran fuente de información no estructurada; en consecuencia, información no explotada en su totalidad. Algunos términos usados en textos clínicos se encuentran en una situación de afirmación, negación, hipótesis o histórica. La detección de esta situación es necesaria para la estructuración de información, pero a su vez tiene una gran complejidad. Extrayendo características lingüísticas de los elementos, o tokens, de los textos mediante NLP; transformando estos tokens en instancias y las características en atributos, podemos mediante técnicas de machine learning clasificarlos con el objetivo de detectar si se encuentran afirmados, negados, hipotéticos o históricos. La selección de los atributos que cada token debe tener para su clasificación, así como la selección del algoritmo de machine learning utilizado son elementos cruciales para la clasificación. Son, de hecho, los elementos que componen el modelo de clasificación. Consecuentemente, este trabajo aborda el proceso de extracción de características, selección de atributos y selección del algoritmo de machine learning para la detección de la negación en textos clínicos en español. Se expone un modelo para la clasificación que, mediante el algoritmo J48 y 35 atributos obtenidos de características lingüísticas (morfológicas y sintácticas) y disparadores de negación, detecta si un token está negado en 465 frases provenientes de textos clínicos con un F-Score del 73%, una exhaustividad del 66% y una precisión del 81% con una validación cruzada de 10 iteraciones. ---ABSTRACT--- New information technologies have emerged in the recent years which explore the processing of the huge amount of existing digital data and its transformation into knowledge. Natural Language Processing (NLP) techniques are able to extract certain features from digital texts. Additionally, through machine learning techniques it is feasible to classify instances according to different categories, learning from others previously classified. Clinical texts contain great amount of unstructured data, therefore information not fully exploited. Some terms (tokens) in clinical texts appear in different situations such as affirmed, negated, hypothetic or historic. Detecting this situation is necessary for the structuring of this data, however not simple. It is possible to detect whether if a token is negated, affirmed, hypothetic or historic by extracting its linguistic features by NLP; transforming these tokens into instances, the features into attributes, and classifying these instances through machine learning techniques. Selecting the attributes each instance must have, and choosing the machine learning algorithm are crucial issues for the classification. In fact, these elements set the classification model. Consequently, this work approaches the features retrieval as well as the attributes and algorithm selection process used by machine learning techniques for the detection of negation in clinical texts in Spanish. We present a classification model which, through J48 algorithm and 35 attributes from linguistic features (morphologic and syntactic) and negation triggers, detects whether if a token is negated in 465 sentences from historical records, with a result of 73% FScore, 66% recall and 81% precision using a 10-fold cross-validation.

New Techniques in Deep Representation Learning

Thesis (Ph.D.)--University of Washington, 2016-06

Applying machine learning to catalogue matching in astrophysics

We present the results of applying automated machine learning techniques to the problem of matching different object catalogues in astrophysics. In this study, we take two partially matched catalogues where one of the two catalogues has a large positional uncertainty. The two catalogues we used here were taken from the H I Parkes All Sky Survey (HIPASS) and SuperCOSMOS optical survey. Previous work had matched 44 per cent (1887 objects) of HIPASS to the SuperCOSMOS catalogue. A supervised learning algorithm was then applied to construct a model of the matched portion of our catalogue. Validation of the model shows that we achieved a good classification performance (99.12 per cent correct). Applying this model to the unmatched portion of the catalogue found 1209 new matches. This increases the catalogue size from 1887 matched objects to 3096. The combination of these procedures yields a catalogue that is 72 per cent matched.

An Incremental EM-based Learning Approach for On-Line Prediction of Hospital Resource Utilization

Objective: Inpatient length of stay (LOS) is an important measure of hospital activity, health care resource consumption, and patient acuity. This research work aims at developing an incremental expectation maximization (EM) based learning approach on mixture of experts (ME) system for on-line prediction of LOS. The use of a batchmode learning process in most existing artificial neural networks to predict LOS is unrealistic, as the data become available over time and their pattern change dynamically. In contrast, an on-line process is capable of providing an output whenever a new datum becomes available. This on-the-spot information is therefore more useful and practical for making decisions, especially when one deals with a tremendous amount of data. Methods and material: The proposed approach is illustrated using a real example of gastroenteritis LOS data. The data set was extracted from a retrospective cohort study on all infants born in 1995-1997 and their subsequent admissions for gastroenteritis. The total number of admissions in this data set was n = 692. Linked hospitalization records of the cohort were retrieved retrospectively to derive the outcome measure, patient demographics, and associated co-morbidities information. A comparative study of the incremental learning and the batch-mode learning algorithms is considered. The performances of the learning algorithms are compared based on the mean absolute difference (MAD) between the predictions and the actual LOS, and the proportion of predictions with MAD < 1 day (Prop(MAD < 1)). The significance of the comparison is assessed through a regression analysis. Results: The incremental learning algorithm provides better on-line prediction of LOS when the system has gained sufficient training from more examples (MAD = 1.77 days and Prop(MAD < 1) = 54.3%), compared to that using the batch-mode learning. The regression analysis indicates a significant decrease of MAD (p-value = 0.063) and a significant (p-value = 0.044) increase of Prop(MAD

The exact sample complexity of PAC-learning problems with unit VC dimension

The Vapnik-Chervonenkis (VC) dimension is a combinatorial measure of a certain class of machine learning problems, which may be used to obtain upper and lower bounds on the number of training examples needed to learn to prescribed levels of accuracy. Most of the known bounds apply to the Probably Approximately Correct (PAC) framework, which is the framework within which we work in this paper. For a learning problem with some known VC dimension, much is known about the order of growth of the sample-size requirement of the problem, as a function of the PAC parameters. The exact value of sample-size requirement is however less well-known, and depends heavily on the particular learning algorithm being used. This is a major obstacle to the practical application of the VC dimension. Hence it is important to know exactly how the sample-size requirement depends on VC dimension, and with that in mind, we describe a general algorithm for learning problems having VC dimension 1. Its sample-size requirement is minimal (as a function of the PAC parameters), and turns out to be the same for all non-trivial learning problems having VC dimension 1. While the method used cannot be naively generalised to higher VC dimension, it suggests that optimal algorithm-dependent bounds may improve substantially on current upper bounds.

Incorporating curvature information into on-line learning

We analyse the dynamics of a number of second order on-line learning algorithms training multi-layer neural networks, using the methods of statistical mechanics. We first consider on-line Newton's method, which is known to provide optimal asymptotic performance. We determine the asymptotic generalization error decay for a soft committee machine, which is shown to compare favourably with the result for standard gradient descent. Matrix momentum provides a practical approximation to this method by allowing an efficient inversion of the Hessian. We consider an idealized matrix momentum algorithm which requires access to the Hessian and find close correspondence with the dynamics of on-line Newton's method. In practice, the Hessian will not be known on-line and we therefore consider matrix momentum using a single example approximation to the Hessian. In this case good asymptotic performance may still be achieved, but the algorithm is now sensitive to parameter choice because of noise in the Hessian estimate. On-line Newton's method is not appropriate during the transient learning phase, since a suboptimal unstable fixed point of the gradient descent dynamics becomes stable for this algorithm. A principled alternative is to use Amari's natural gradient learning algorithm and we show how this method provides a significant reduction in learning time when compared to gradient descent, while retaining the asymptotic performance of on-line Newton's method.

Protein-protein interactions classification from text via local learning with class priors

Text classification is essential for narrowing down the number of documents relevant to a particular topic for further pursual, especially when searching through large biomedical databases. Protein-protein interactions are an example of such a topic with databases being devoted specifically to them. This paper proposed a semi-supervised learning algorithm via local learning with class priors (LL-CP) for biomedical text classification where unlabeled data points are classified in a vector space based on their proximity to labeled nodes. The algorithm has been evaluated on a corpus of biomedical documents to identify abstracts containing information about protein-protein interactions with promising results. Experimental results show that LL-CP outperforms the traditional semisupervised learning algorithms such as SVMand it also performs better than local learning without incorporating class priors.

The evolution of learning systems:to Bayes or not to be

Bayesian algorithms pose a limit to the performance learning algorithms can achieve. Natural selection should guide the evolution of information processing systems towards those limits. What can we learn from this evolution and what properties do the intermediate stages have? While this question is too general to permit any answer, progress can be made by restricting the class of information processing systems under study. We present analytical and numerical results for the evolution of on-line algorithms for learning from examples for neural network classifiers, which might include or not a hidden layer. The analytical results are obtained by solving a variational problem to determine the learning algorithm that leads to maximum generalization ability. Simulations using evolutionary programming, for programs that implement learning algorithms, confirm and expand the results. The principal result is not just that the evolution is towards a Bayesian limit. Indeed it is essentially reached. In addition we find that evolution is driven by the discovery of useful structures or combinations of variables and operators. In different runs the temporal order of the discovery of such combinations is unique. The main result is that combinations that signal the surprise brought by an example arise always before combinations that serve to gauge the performance of the learning algorithm. This latter structures can be used to implement annealing schedules. The temporal ordering can be understood analytically as well by doing the functional optimization in restricted functional spaces. We also show that there is data suggesting that the appearance of these traits also follows the same temporal ordering in biological systems. © 2006 American Institute of Physics.

Self-Learning Fuzzy Spiking Neural Network as a Nonlinear Pulse-Position Threshold Detection Dynamic System Based on Second-Order Critically Damped Response Units

Architecture and learning algorithm of self-learning spiking neural network in fuzzy clustering task are outlined. Fuzzy receptive neurons for pulse-position transformation of input data are considered. It is proposed to treat a spiking neural network in terms of classical automatic control theory apparatus based on the Laplace transform. It is shown that synapse functioning can be easily modeled by a second order damped response unit. Spiking neuron soma is presented as a threshold detection unit. Thus, the proposed fuzzy spiking neural network is an analog-digital nonlinear pulse-position dynamic system. It is demonstrated how fuzzy probabilistic and possibilistic clustering approaches can be implemented on the base of the presented spiking neural network.