On the Convergence of Stochastic Iterative Dynamic Programming Algorithms

Recent developments in the area of reinforcement learning have yielded a number of new algorithms for the prediction and control of Markovian environments. These algorithms, including the TD(lambda) algorithm of Sutton (1988) and the Q-learning algorithm of Watkins (1989), can be motivated heuristically as approximations to dynamic programming (DP). In this paper we provide a rigorous proof of convergence of these DP-based learning algorithms by relating them to the powerful techniques of stochastic approximation theory via a new convergence theorem. The theorem establishes a general class of convergent algorithms to which both TD(lambda) and Q-learning belong.

Improving Multiclass Text Classification with the Support Vector Machine

We compare Naive Bayes and Support Vector Machines on the task of multiclass text classification. Using a variety of approaches to combine the underlying binary classifiers, we find that SVMs substantially outperform Naive Bayes. We present full multiclass results on two well-known text data sets, including the lowest error to date on both data sets. We develop a new indicator of binary performance to show that the SVM's lower multiclass error is a result of its improved binary performance. Furthermore, we demonstrate and explore the surprising result that one-vs-all classification performs favorably compared to other approaches even though it has no error-correcting properties.

FastAero – A Precorrected FFT – Fast Multipole Tree Steady and Unsteady Potential Flow Solver

In this paper a precorrected FFT-Fast Multipole Tree (pFFT-FMT) method for solving the potential flow around arbitrary three dimensional bodies is presented. The method takes advantage of the efficiency of the pFFT and FMT algorithms to facilitate more demanding computations such as automatic wake generation and hands-off steady and unsteady aerodynamic simulations. The velocity potential on the body surfaces and in the domain is determined using a pFFT Boundary Element Method (BEM) approach based on the Green’s Theorem Boundary Integral Equation. The vorticity trailing all lifting surfaces in the domain is represented using a Fast Multipole Tree, time advected, vortex participle method. Some simple steady state flow solutions are performed to demonstrate the basic capabilities of the solver. Although this paper focuses primarily on steady state solutions, it should be noted that this approach is designed to be a robust and efficient unsteady potential flow simulation tool, useful for rapid computational prototyping.

Aplicaciones del análisis funcional del comportamiento a la instrucción superior.

Analizar la situación de la enseñanza universitaria tradicional y diseñar un sistema basado en los principios del análisis del comportamiento como alternativa a la anterior. Alumnos universitarios. Textos confeccionados en base a la tecnología de la enseñanza. Evaluación continua El método experimental que se diseñó arrojó mejores resultados en cuanto a rendimiento que el tradicional. Es posible mejorar el proceso de enseñanza-aprendizaje siempre que se tengan en cuenta los principios teóricos del proceso y se cambie la función actual del profesor, su papel, y se diseñe la enseñanza de acuerdo a los principios anteriores.

Exam 85 pdf

Exam questions and solutions in PDF

Exercise sheet 8 pdf

Exercises and solutions in PDF

Exam 1999 pdf

Exam questions and solutions in PDF

271 Exam 84 pdf

Exam questions and solutions in PDF

Lecture notes

Lecture notes in LaTex

Exercise sheet 4 pdf

Exercises and solutions in PDF

Exercise sheet 1 pdf

Exercises and solutions in PDF

Exercise sheet 9 pdf

Exercises and solutions in PDF

Exam 1999

Exam questions and solutions in LaTex

Exam 99 pdf

Exam questions and solutions in PDF

MA17xexam00pdf

Exam questions and solutions in PDF