In this month, I continued my studies in Relational Reinforcement Learning by reviewing the article Towards Informed Reinforcement Learning from the proceedings of the 2004 Machine Learning workshop of Relational Reinforcement Learning.   Basically the articles summarizes that an agent with limited information can find an optimal policy and can achieve a goal or goal states with limited information about its environment.  The experiments reported seems to suggest this type of exploration is possible.  According to Google Scholar search, there are 11 subsequent articles that reference this one.  In the RRL arena, my goal is to repeat the block’s world experiment as reported in Relational Reinforcement Learning article by Dzeroski, De Raedt, and Blockeel.

Also, I have been focused on Robert Winkler’s book An Introduction to Bayesian Analysis and Inference.  From the Winkler book, my goal is to gain a better understanding of Bayesian Inference to better understand the Decision Theoretic models of machine learning.  In the machine learning world, Bayesian approaches to various problems are resulting interesting solutions to various problems in Multiagent Reinforcement Learning.

There are some interesting articles in the Journal of Bayesian Analysis and in JAIR.

In another topic of interest, I am continuing my reintroduction of LISP by reading the Patrick Winston and Berthold Horn classic.

From a statistical point of view, this month achieve a new record number of hits with 806 hits, boosting my total for the current year to over 3200 hits.  The hotest pages are my About page, followed by Wumpus World and Wumpus World Revisited.

Thank you for your interest and support.

This month marks my first year in utilizing wordpress to host my blog.  It has been a great journey so far and hopefully will be better this coming year.

The month of September has been busy.  Unfortunately I did not write any blog entries due to my busy schedule.  However, from statistical point of view, my blog had the second highest total of number of views (362 in total).  Also, I covered various topics this month with Probability and Statistics, Bayesian Inference, Reinforcement Learning, and finally a LISP review.

I started to review the Reinforcement Learning book (Sutton and Barto 1997) only reading Part 1 and part of Part 2.  I did download the lisp code associated with the book.  I ran the tic-tac-toe program in my test environment.

In addition, I read Part 1 of the PAIP (Norving 1992).  What this book forced me to review LISP (Winston and Horn 1989), only covering the first three chapters in just a few days.

Also I completed the third chapter of Inductive Logic Programming (Bergadano and Gunetti 1996), which covered bottom up ILP methods.  Material is diverse since the authors integrated many papers together to discuss the important bottom up methods.

Next I completed the first four chapters of  Introduction to Bayesian Inference and Decision (Winkler 2003).  The books covers probability theory as well as the the introduction to Bayes Theorem.  Chapter 3 covers Bayesian Inference for Discrete Probability Models.  Very excellent material for those of you interested in Bayesian Inference and Decision as to gain insights into Decision Theoretic models.  The exercises are very good as well.  Looking forward to completing book followed by reading AIMA Part V.  For interested in Bayesian Methods, download the Journal of Bayesian Analysis.

All this material has been great.  I started thinking about some ideas such as using an agent approach for the farmer, wolf, goat, and cabbage problem.  Although it may seem a bit overkill, I was interested in seeing various methods to this problem with reinforcement learning methods.

The hottest page was my About page, followed by Wumpus World, AI Links page, the Warnsdorf algorithm, and last month’s progress report.

Thank you for your interest and support.

During the month of August, I focused on Relational Reinforcement Learning, a field that combines Relational Learning and Reinforcement Learning fields.  Please read my blog entry for Relational Reinforcement Learning.  Afterwards, I have read a number of different articles regarding RRL learning from different authors, including the RRL workshop at the ICML 2004 conference.  My initial preference is the work performed and researched by Dr. Eduardo F. Morales.

I posted my review and comments regarding part I of the book The Art of Prolog (Sterling and Shapiro 1994).  There are very few books regarding Logic Programming in recent years.  Since the mid 1990s, much of the effort in Logic Programming has been with Inductive Logic Programming and Relational Learning.  The emerging field of Statistical Relational Learning has become a new field of research.  In addition, the prolog language has been very stable since the ISO Prolog standard was established.  Some the prolog has minor differences with certain predefined predicates, but it has been compatible with Edinbrugh Prolog.  The current Prolog books cover some basic AI applications such as games, mini expert systems, and meta interpreters.  What is needed is agent/environment based applications using modern tools such as Q-Learning or TD-methods.

I finally completed Scott Sanner’s video lecture in Reinforcement Learning.  An excellent tutorial.  The lecture was discussing topics such as model based and model free methods.  The subjects MDPs, dynamic programming, Monte Carlo, temporal difference, and function approximations are discussed.

I also performed research on the Warnsdorf algorithm and posted my initial results.  Most of this work was done using procedural languages and with square boards of size five and greater.

In addition, I found the ICML-06 Workshop on Structural Knowledge Transfer for Machine Learning.

I finally completed the Summer Schools in Logic and Learning (SSLL 09) video lecture for Reinforcement Learning presented by Scott Sanner.  It was a good introduction into Reinforcement Learning.  It is a six part lecture.  Plan for time to watch the videos.  He starts with an introduction into the topic.  Next is model based solutions with Markov Decision Process (MDP) and dynamic programming (DP) with value iteration and policy iteration.  The Bellman equations are introduced.  Lastly, he reviews model free approaches from Chapters 5, 6, and 7 of the Reinforcement Learning: An Introduction (Sutton and Barto 1998).  He covers Monte Carlo and Temporal Difference (TD) algorithms, in particular TD lamdba and SARSA.  However due to large state spaces of problems, he discussion function approximators (both linear and nonlinear) along with gradient descend approaches.  A worth while video lecture.

As part of my follow research to my paper, I did some preliminary research for the Warnsdorf’s algorithm for finding knight’s tours.  I found three relevant papers on the subject (Pohl 1967, Pohl-Stockmeyer 2004, and Ganzfried 2004).  Basically, the Warnsdorf algorithm is to traverse the path with least degree.  In the case of a tie, the a path is randomly selected.  In reality, the Warnsdorf algorithm is not consistently successful as pointed out by (Ganzfried 2004).  The (Ganzfried 2004) paper modifies the Warnsdorf algorithm with key squares in which the move order changes, which improves finding knight tours.  The next step is generate a prolog program using the Pohl-Warnsdorf and Ganzfried algorithms.

The book The Art of Prolog (Sterling and Shapiro 1994) is an excellent treatise in the subject of Logic Programming and Prolog.  The book is divided into four parts – Logic Programming, The Prolog Language, Advance Prolog Programming Techniques, and Applications.

Beginning with Logic Programming, chapter one provides the best introduction into Logic Programming as well as the first definition of the abstract interpreter for logic programming.  The definitons for term, functor, compound term, clause or rules, logic programs, and the meaning for logic programs are presented.  The summary is worth reading.

Chapter Two goes into database programming and it connection to logic programming.

Next is Recursion in Chapter Three, which illustrates many techniques and links the original logic programming symbols.  For example the definition of a natural number is recursive; e.g. s(0), which is the successor function.  Many example programs are illustrated in this chapter.

The fourth chapter covers unification and expands on the simple abstract interpreter for logic programming.

The section ends with a chapter on Theory of Logic Programming by introducing terms such as the Hebrand Base and Universe of a Logic Program.  Each Chapter has a background section explaining the various directions and original papers on the subject matter.  So far, a very good book to read on logic programming.

I start this month reading the articles in Relational Reinforcement Learning.  It began as a curiosity due to a response in the Reinforcement Learning mailing list.  The authors point a speech by Rich Sutton and Leslie Pack Kaelbling at the IJCAI 1997 conference in Japan, in which they recommended the combination of Induction Logic Programming (aka Relational Learning) and Reinforcement Learning.  The authors of the paper demonstrate this concept by using the blocks world domain to illustrate the combination of both fields.  Perhaps, this is exactly what I needed since I have done much reading, studying, and video lectures with logic programming, ILP, and reinforcement learning.  Of course, this paper spawn much researching according to CiteSeerX and Google Scholarly.

I found the web site to the 2004 Workshop in Relational Reinforcement Learning and downloaded the proceedings.

I shall continue this path of research interest.

In this progress report, I attended the ICAI 09 conference in Las Vegas (the IJCAI 2009 is located in Pasadena).  I presented my work on A Brute Force Approach to Solving the Knights Tour Using Prolog.  One thing that I learned is the participants are proud to share their work with others.  Another lesson is to make your presentation within the time allotted.  A few presenters did not finish their presentations in the 20 minutes time slot.  I rehearsed my presentation a few times to have the timing within 20 minutes.  I also learned that I need to get more training with MS Power Point.

I am also continue to view the video lectures from the Summer Schools in Logic and Learning from Video Lectures.  Thus, I have completed the Intelligent Agents video lectures by John Lloyd.  The presentation references the AIMA book and the material is from Chapters 1, 2, 26 and 27.  He illustrates various agent architectures with the main conclusion that reinforcement learning agents can determine the best utility function.  Then he discusses the history of AI and Philosophy of AI.  A good introduction into the topic of agents.  He also makes a point about statistical relational learning, a new field combining the relational learning with statisticial learning methods – this field was discussed at the ICML 2007 conference.

Another lecture series that I also viewed was the Scott Sanner video lecture in Reinforcement Learning.  He begins with an interesting approach, but eventually ties into the RL book.

I have completed reading Part VI (Learning) of the AIMA book.  Back in 1995, the material was relevant to its time.  In the second edition, the authors replaced the chapter on neural networks with Statistical Learning Methods.  This new chapter incorporates Probabilistic Inference, Learning with Incomplete data, Bayesian Networks, Learning with Hidden data, Instance Based Learning, Neural Networks, and Kernel Machines.  Perhaps the most interesting topic is kernel machines, but the topic was superficially covered.  Though statistical learning methods are not as good as decision learning trees for some domains.  Since then, Relational Learning has expanded to include statistical methods, creating a new field of Statistical Relational Learning (SRL).

In additionally, I wrote a post on advanced Prolog programming.  I believed that this item is missing in books.

In other news, the International Joint Conference in Artificial Intelligence (IJCAI) 2009 was held in Pasadena this year.  I had reviewed the accepted papers list, which was very interesting.

In order to generate interest in my blog, I posted announcements in various news groups.  This generated over 700 hits to the site.  The hot blog entries and pages are as follows:

About (38), Wumpus World (33), AI Links(30), and Progress Report – June 2009 (26).

Thank you for your support.

Reading the comp.ai.prolog news group, I noticed the prolog FAQ did not make any references to any advance prolog programming websites.  Though it mentions the Craft of Prolog as the book for the advance reader, it was published before the ISO Prolog standard was finalized.  I have in my wish list in Amazon the book Prolog Programming in Depth by Michael Covington et. al (Fascimile edition), which seems from the table contents a book that could compliment the Bratko’s Prolog: Programming for Artificial Intelligence and The Art of Prolog.  Here is an example:

cycle(X) :-
   (    X < 1 -> true
        ;
        write(X),
        nl,
        X1 is X - 1,
        cycle(X1)
   ).

The above example is used to count down from X to 1.  In all the online tutorials, I have not seen example using the -> control construct.  I will on occasion will publish some prolog code.

Part VI of the AIMA book covers learning and various techniques on agents can learn.  The material covers Learning from Observations, Neural Networks, Reinforcement Learning, and Knowledge in Learning.  In the second edition, the chapter in Neural Networks is replaced with Statistical Learning Methods.

I read Chapter 18 of the AIMA book called Learning from Observations.  The Chapter focuses on decision trees and decision lists as some computational learning theory.  The main classification algorithm is the ID3 algorithm for classifying examples and generating decision trees.  The ID3 algorithm is based on Hunt’s Concept Learning System (CLS).  The ID3 algorithm uses information theory for obtaining the decision tree covering most the examples.  I have studied the ID3 algorithm quite extensively since the original application was to classify the winning position in the king and rook versus the king and knight endgame (Quinlan 1983).

Chapter 19 covers Neural Networks (in the second edition this chapter is replaced with Statistical Learning Theory).  The chapter reviews single and multilayer network networks as well as the perceptron, hopfield networks, and boltzmann machines.  Also, the Bayesian Belief Network is discussed.  Although there has been with some successes such as NET talk, hand recognition, and ALVINN.  Still neural networks have not reached a level of success as other implementations of machine learning.

Chapter 20 covers Reinforcement Learning.  This chapter the authors present reinforcement learning by covering the model based approach.  The utility function is maximized based on the state and action yielding the reward.  Unfortunately, the state space in many problems is very learning, and therefore, it is impractical to assign values to a large state space.  Other methods such as temporal difference (TD) methods are used and Q-learning are used to find the optimal policy.  The authors discuss the published successes in games such as Neurogammon, TDgammon, and Othello.  Current game research using reinforcement learning is investigating Go.  Lastly, the authors consider Evolutionary Computing (EC) as a special subset of reinforcement learning since the goal is to find the optimal fitness function.

Finally Chapter 21 covers Knowledge in Learning.  The authors present the case about learning given background knowledge.  The authors discuss the concept of knowledge base inductive learning (KBIL).  First, attribute learning was presented with background knowledge.  The main difference with Chapter 18 is background knowledge is used to narrow the hypothesis space.  Then Explanation Based Learning (EBL) was presented along with some applications.  In addition, the authors discussed an extension to ID3 with background knowledged (RBID3).  Then, Inductive Logic Programming (ILP) was presented.  ILP can solve problems by using Inverse Resolution (IR) or using top down methods like FOIL.  ILP uses horn clauses for the knowledge base and problem definition.  Upon an ILP process based on the examples, a logic program is generated.  ILP takes advantage of relational learning.

Based on my reading experiences, the ID3 algorithm has been successful algorithm since its successors C4.5 and C5 have commercial success.  However, there are other attribute based classification algorithms such as the AQ-15 by Michalski and CN2 by Clark and Niblett.  Also, related to Reinforcement Learning research is Markov Decision Process (MDP) and Partially Observable MDP (POMDP) models, but these concepts are covered in Chapter 17 of Part V.  With ILP, the Lavrac and Dzeroski Inductive Logic Programming: Techniques and Applications book is available online (see the references page for the URL).  There are many active areas of learning research today.