1. Social and Emergent Models of Learning.
2. The Genetic Algorithm.
3. The Genetic Algorithm applied to CNF SAT and Traveling Salesman.
4. Evaluating the GA.
5. Classifier Systems and Genetic Programming. Use of LISP.
6. Artificial Life and Social-Based Learning.
7. The "game of life".
8. Evolutionary Programming.
9. A Case Study in Emergence from Crutchfield and Mitchell.

1. Emergent approach.
   
   
2. Fundamentals of evolutionary methods. What is Evolutionary computation. Philosophy of modeling evolution. Genetic Algorithms. Encoding the problem. Fitness function. Evolution strategies. Comparizon and examples.
   
   
3. Genetic Programming, Evolutionary Strategies, Evolutionary Design. Fogel, state machines. Evolution versus intelligence. Intellectual adaptation. Learning state machines. Prediction experiments. Pattern recognition and classification. Humans versus machines. Control System Design.
   
   
4. Genetic Programming. What is Genetic Programming. Data structures. History. How are genetic principles applied. Computer programs as trees. Fitness. Mutation. Examples of applications.

1. Extrinsic Evolvable Hardware. Extrinsic approaches to Evolvable Hardware, including the approach from PSU based on learning and optimizing finite state machines. "Evolvable Hardware" or "Learning Hardware"? Machine Learning is designing a network. Induction of State Machines from Temporal Logic Constraints. Hardware speed-up of learning algorithms. Technologies for learning hardware. Phases of learning. Comparison of approaches. Multi-valued logic language to represent the learning data in hardware. Logic Patterns in tables. Regular automata. Types of cube calculus used for learning. Logic Synthesis appraoch to learning. Strong Criterium for Learning. Decomposition and Constructive Induction. Extrinsic versus Intrinsic approaches to Evolvable Hardware. Learning using FPGAs. Temporal Logic Constraints as an example of a language to describe behavior. Man, Wolf, Goat and Cabbage problem. Example of software.
   
   
2. Intrinsic Evolvable Hardware. Intrinsic approach to Evolvable Hardware, including Brain Bulder from ATR and Learning Hardware from PSU. Recent ambitious approaches to build robots. Approach of ATR in Japan and its critics. Hardware and concepts of CAM-brain machine. Evolving Neural Nets on top of Cellular Automata. Universal Logic Machine and approach of PSU. What worked, what not. Good guys versus bad guys, our approach to Constructive Induction - decomposition. Schematics diagram and components of our approach. Previous related undergraduate and graduate projects.

1. Evolutionary Methods in Design. Art. Lego bridges of Pollack. Evolutionary art and how it relates to robotics. Mutator program. Computer Evolution of Buildable Objects of Pablo Funes and Jordan Pollack.
   
   
2. Levi. HereBoy Algorithm. Multi-Mutation. Relation to Simulated Annealing. Claims. Comparisons. Evolutionary time. HereBoy versus previous approaches.
   
   
3. Vasiliev and Miller. Array Genetic Algorithm for Multiplier Design. Towards the Automatic Design of More Efficient Digital Circuits. Evolving Digital Circuits. Multipliers. Fitness landscapes. Cartesian genetic programming. Genotype-phenotype mapping. Scalability problem of digital circuits.
   
   
4. Approach to evolvable learning and self-repair based on embriology ideas. Switzerland. The Embryonics Approach. Chromosome of Caenorhabditis Elegans. Multicellular organization. Why Embryonics. Cellular Division. Cellular Differentiation. Kinds of programmable arrays. Arrays based on switching. MUXtree molecule - organization. Embryonics Landscape. Stop Watch example: cellular differentiation and Self-replication. Self-repair. Artificial Genome.
   
   
5. Evolving game strategies Characteristics of Intelligent Agents. Evolving Game Playing Strategies. Emergent properties. Prisoner's Dilemma. Prisoner's Dilemma as a model of nature. Games and strategies based on Prisoner's Dilemma. Axelrod's Tournaments. Deterministic versus stochastic strategies. Strategies in nature. Examples. Evolution of Behavior in Nature. Kin selection. Cooperative Breeding. Intersexual selection - charm.

1. Towards Robot Theatre and fundamental models in robotics.
2. Mandatory Reading.
   Read all slides for this week.
   Read Chapter 21 of the textbook. "Genetic Programming".
   Read Chapter 22 of the textbook. "Behavior-Based Systems".
   Read Chapter 23 of the textbook. "Evolution of Walking Gaits".
   Homework 1. Assigned in class on Tuesday. Due Tuesday of next week.
   This homework is related to grammars for motion, probabilistic movement generation and simple Braitenberg Vehicles (reactive architecture).

1. 7ROBODEM.AVI Our robot.
2. ApplicationsClip.mpg Teenage student Yale Fan talks about his nationally acclaimed research at PSU.
3. Talking-lions-2.avi Our lions who argue permanently - from Hahoe Theatre. Built by Martin Lukac and Kyle.
4. biped.MOV The Lego biped built by a teenager Michal Woyke in Poland.
5. braiten-clp1.wmv Braitenberg Vehicle of Arushi, a teenage girl who works on quantum robots. This robot is afraid of light.
6. braiten-clp2.wmv Braitenberg Vehicle of Arushi. This robot is afraid of objects, obstacles.
7. braiten-clp3.wmv Braitenberg Vehicle of Arushi. Agressive robot.
8. braiten-clp4.wmv Braitenberg Vehicle of Arushi. Light following with switched behavior.
9. braiten-clp5.wmv Braitenberg Vehicle of Arushi. Turns left from objects.
10. braiten-clp6.wmv Braitenberg Vehicle of Arushi. Likes music.
11. braiten-clp7.wmv Braitenberg Vehicle of Arushi. Attacks music source but is afraid of a hand.
12. robohead-clp1.wmv Braitenberg Vehicle of Arushi. Einsteind-Podolsky-Rosen robot called Mister Quantum Potato Head. Illustrates EPR entanglement with movement of mouth and eyes. (11 measured). Pay attention to delay when it does nothing (00 measured)
13. robohead-clp2.wmv Braitenberg Vehicle of Arushi. Einsteind-Podolsky-Rosen robot called Mister Quantum Potato Head.
14. robohead-clp3.wmv Braitenberg Vehicle of Arushi. Einsteind-Podolsky-Rosen robot called Mister Quantum Potato Head.

1. Link to student Alex Perez with information about Sonbi Robot and Robot Theatre.

1. Robot Examples. My requirement is: the robot should have 2 hands, head and torso and should be a mobile robot.
2. X000. Tetrix_Short_Promo.wmv
3. X001. Intro - Lego NXT, Tetrix, ROBOTC and Motors.ppt
4. X002. Simple programming in ROBOTC.ppt
5. X003. Tetrix Mechanical Design.pptx
6. X003A. Tetrix Robot Mechanical Examples.pptx
7. X004. Robot C complete example - robot_head_1.pdf
8. X004A. Smarter BV.ppt

1. X005. Lego NXT complete example with NXC.pptx
2. X006. Projects with simple mobile robots.ppt
3. X007. Simple Sensors in ROBOTC.ppt
4. X008. IR sensors and Encoders. LCDs.ppt
5. X009. Accelerometers and Gyros.ppt
6. X011. Behavior based robots and architectures.ppt

1. X012. Affective-Computing.ppt
2. X013. More advanced Projects.ppt
3. X016. NQC Programming.ppt
4. X017. NXC Programming.ppt
5. Example of software in RobotC.
6. X018. OUR project..ppt Some ideas about our project this summer. I will add more when we will know more what we want to build.
7. Flagbot project The practical project that we have done last year.
8. How to Design a Robot.ppt
9. Class with robot project. A robot connected to quantum computer.

1. Simple and Quantum Braitenberg Vehicles.
2. Quantum Braitenberg and advanced behavioral architectures. Homework is here.
3. Mandatory Reading.
   Read all slides for this week.
   Review chapters 20 - 23 of the textbook.
   Read Chapter 16 of the textbook. "Maze Exploration".
   Homework 2. Assigned in class. Due Tuesday of week 5.
   This homework is related to behavior robot in a maze or other environment. See slides.

1. Genetic Algorithm for Lego.
2. Genetic Algorithm for logic synthesis.
3. Generative Art. How to use evolutionary ideas to produce art.
4. THE HOMEWORK. The homework. Applying the class ideas to the robot of your choice, you must use a GA or GP.

SLIDES BELOW THIS LINE ARE NOT VALID FOR YEAR 2009

1. Subsumption Theory.
2. Examples-of-Subsumption Architectures in Mobile Robots.
3. Reactive software.
4. Behavioral Architectures.
5. Introduction to Machine Learning.
6. Decision Trees.

WEEK 9.


Maze Running. Soccer. Behavioral Robots in environments. Social behaviors.


1. Mandatory Reading.
   Read all slides for this week.
   Read chapter 17 from the textbook. "Map Generation".
   Read chapter 18 from the textbook. "Robot Soccer".
   Review all covered material before the midterm exam next week.
   Homework 3.
   Assigned in class. Neural Network for robot.

WEEK 10.


Tuesday. Review: Maze Running. Soccer. Behavioral Robots in environments. Social behaviors.


1. Thursday.
   Midterm exam. 2 hours. Open book.
   
   
   
   
   1. Introduction to the class. What is a robot. In PPT format.
      
      
   2. History. In PPT format.
      
      
   3. What robots are good for. In PPT format. Learn about various types of robots that exist and can be build by us.
      
      
   4. Inexpensive robots. In PPT format. Learn about various types of robots that we already built in the lab. You can reuse, extend them or modify. You can build your own similar robots.
      
      
   5. Lego Robots. In PPT format. Inexpensive but complete robots that we used for fast prototyping of robots in the past. Used in high schools and in various competitions.
      
      
   6. Pneumatic Robots. In PPT format. We built several such robots in the past but next we switched to standard electric servos.
      
      
   7. Robots from Robix kit, the fastest way to control movements. In PPT format. We have two Robix kits available for students to use.
      
      
   8. Robot soccer competitions. Learn about 2002 year class project. In PPT format.
      
      
   9. Look to the webpages of Martin Lukac and Jacob Biamonte for project descriptions and source codes for robots. Contact them if something is not clear.
      
      
   1. Explanation of soccer project.
   2. Legged Robotic Soccer Project at PSU. In the past we built a robot team of walking robots. The image processing can be partially re-used in this year projects.
   1. Introduction to Talking Head and theatre projects for year 2004. Find more on these topics in Martin's www page, below, and on the link to my Korean lectures from the bottom of my main www page.
      
   2. Introduction to our theatre in year 2005.
      
      
   1. Introduction to robot theatre.
   2. Our inexpensive robots. Ideas for practical realization. This will teach you inexpensive methods and items in our lab that can be used for talking head robots.
   3. Emotional Robots and Immitation. This will teach you about robots' emotions and how they immitate humans. These methods can be all used in our talking heads.
   4. Arms of Industrial robots. Part I This material is useful in projects related to building various types of robot arms.
   5. Arms of Industrial robots. Part II This material is useful in projects related to building various types of robot arms.
   6. Design of walking robots. This is an introductory material useful if your project is a walking robot.
   7. This is a simple explanation of servos and their use. Futaba and Hitec servos are used in our projects.
   8. Hexapods for robot soccer.
   9. Pattern Recognition Project for PSU Big Hexapod.
   10. Uland Wong and Marek Perkowski, ``A New Approach to Robot's Imitation of Behaviors by Decomposition of Multiple-Valued Relations.'' Proc. Boolean Problems, 2002. Freiberg, Germany.
   
   
   

   SECTION 1.3. BASIC LISP. LISTS AND RECURSION. BASIC DATA STRUCTURES AND FUNCTIONS.

   
   
   SECTION 1.3.1. MANDATORY READING FROM TEXTBOOK.
   
   Please read from Chapter 15 from Luger's book.
   1. Symbolic Expressions. The syntactic Basis of LISP.
   2. Control of LISP Evaluation: quote and eval.
   3. Programming in LISP; creating new functions.
   4. Program Control in LISP: Conditionals and Predicates.
   5. Functions, Lists and Symbolic Computing.
   6. Lists as recursive structures.
   7. Nested lists, Structure and CAD/CDR recursion.
   8. Binding variables using SET.
   9. Defining local variables using LET.
   10. Data Types in Common Lisp.
   11. Search in LISP. A functional approach to the Farmer, Wolf, Goat and Cabbage Problem.
   12. Higher order functions and abstractions.
   13. Maps and filters.
   14. Functional arguments and lambda expressions.
   15. Search strategies in LISP: breadth-first. Depth-First. Best-First search.
   16. Pattern matching in LISP.
   17. Recursive Unification.
   
   
   SECTION 1.3.2. ADDITIONAL READING FROM TEXTBOOK.
   
   Please read from Chapter 15 from Luger's book.
   1. Interpreters and embedded languages (not required at this time, but very useful for projects).
   2. Loigc Programming in LISP (whole section is not required this quarter).
   3. A simple Logic Programming Language in LISP.
   4. Streams and Stream Processing.
   5. A Stream-based Logic Programming Interpreter.
   6. Streams and delayed Evaluation. (whole section is not required this quarter).
   7. An Expert System Shell in LISP.
   8. Semantic Networks and Inheritance in LISP.
   9. Object oriented programming using CLOS. (whole section is not required this quarter).
   10. Learning in LISP. The ID3 Algorithm. (very useful in projects).
   
   
   
   SECTION 1.3.3. MANDATORY LECTURES.
   
   
   1. Basic functions in Lisp. In PPT format.
   2. Lists in Lisp and functions on them. In PPT format.
   3. Destructive list operations. In PPT format.
   4. Problem solving in Common Lisp. In PPT format.
   
   
   HOMEWORK 2.
   Recursion in Lisp. Simple puzzles and behaviors.
   
   

   SECTION 1.4. SEARCH IN LISP.

   
   
   SECTION 1.4.1. MANDATORY READING FROM TEXTBOOKS.
   
   Please read Chapter 3 in Luger's book.
   1. Introduction.
   2. Graph Theory.
   3. Structures for state space search.
   4. State Space Representation of Problems.
   5. Strategies for State Space Search.
   6. Data-Driven and Goal-driven Search.
   7. Implementing graph search.
   8. Depth-First and Breadth-First Search.
   9. Iterative Deepening.
   
   
   SECTION 1.4.2. ADDITIONAL READING FROM TEXTBOOKS.
   
   
   
   1. Using the State Space to Represent Reasoning with Predicate Calculus.
   2. State Space Description of a Logical System.
   3. AND/OR graphs.
   4. Further examples and applications.
   
   
   SECTION 1.4.3. MANDATORY LECTURES.
   
   
   
   1. Introduction to search in Artificial Intelligence. In PPT format.
      
      
   2. Search. In PPT format.
   
   SECTION 1.4.4. AUXILIARY PROGRAMS.
   1. Many examples of very useful Lisp programs for robotics.
   2. Farmer_wolf_etc_rules_only.lisp
   3. farmer_wolf_goat_cabbage.lisp
   4. Best First Search. Program in Lisp.
   5. Breadth First Search. Search_1.lisp
   6. Breadth First Search. Search_2.lisp
   7. Depth First Search. Program in Lisp
   
   
   

   SECTION 1.5. INTRODUCTION TO MACHINE LEARNING.

   
   SECTION 1.5.1. MANDATORY READING.
   Please read from chapter 9 in Luger's book.
   1. Introduction.
   2. A framework for Symbol-Based Learning. (SBL).
   3. Version Space Search.
   4. Generalization Operators and the Concept Space.
   5. The candidate elimination algorithm.
   6. LEX: inducing Search Heuristics.
   7. Evaluating Candidate Elimination.
   8. The ID3 Decision Tree Induction Algorithm.
   9. Top-Down Decision Tree Induction.
   10. Information theoretic test selection.
   11. Evaluating ID3.
   
   
   SECTION 1.5.2. ADDITIONAL READING.
   
   Please read from chapter 9 in Luger's book.
   1. Decision Tree Data Issues: Bagging, Boosting.
   2. Inductive Bias and Learnability.
   3. The theory of learnability.
   4. Knowledge and Learning.
   5. Meta-Dendral.
   6. Explanation-Based Learning. (EBL).
   7. EBL and Knowledge-Level Learning.
   8. Analogical Reasoning.
   9. Unsupervised Learning.
   10. Discovery and Unsupervised Learning.
   11. Conceptual Clustering.
   12. COBWEB and Structure of Taxonomic Knowledge.
   13. Reinforcement learning.
   14. The components of Reinforcement learning.
   15. An Example: Tic-Tac-Toe Revisited.
   16. Inference Algorithms and Applications of Reinforcement Learning.
   
   
   SECTION 1.5.3. MANDATORY LECTURES.
   
   
   1. Introduction to Machine Learning. In PPT format. What is Machine Learning. Why Machine Learning. A General Model of Learning Agents. Aspects of Learning systems. Recent research in Machine Learning. Major paradigms of Machine Learning. The Inductive Learning Problem. Supervised Concept Learning. Inductive Learning Framework. Nearest Neighbor Classification. Learning Decision Trees. Preference Bias. Ockham's Razor. Inductive Learning and Bias. Using probabilities to create trees: Huffman code. Rule and Decision Tree learning. Neural Network Learning. Specifying a Learning Problem. Learning to play checkers. Data Mining. Planning and control applications. Research Disciplines related to learning.
   2. Decision-trees. In PPT format. Decision tree learning. Training examples. ID3 algorithm. Greedy Search. Information Theory background. Entropy. Use of information measures in ID3. Gain. Bias. Examples. Extensions of Decision Tree algorithms. Real-Valued Data. Noise and Overfitting. Prunning Decision Trees. Incremental Learning. Evaluation Methodology. From Decision Trees to Rules. C4.5. Homeworks and projects.
   3. Trees and Diagrams for Learning. In PPT format. This is an advanced material that is not covered every years. It is covered only depending on projects. Information Theoretic Approach to Minimization in Logic Expressions, Trees, Decision Diagrams and Circuits. Shannon entropy and conditional entropy. Information theoretic measures. History of approaches. Optimal Decision Trees. Arithmetic Spectrum. Extensions to Davio and Shannon expansions. New types of decision trees and diagrams. Pseudo-Kronecker. Word-Level Decision Trees for arithmetic functions. Free Word-Level Decision Trees. Galois logic. Minimization of arithmetic expressions. Linear arithmetic.
   4. Introduction to Machine Learning methods. For possible use in our projects. Some of the methods presented here are below, presented in more detail and with examples. Why learning is working. Error. Computational Learning Theory. Hypothesis evaluation. Bayesian approach to learning. Bayesian Networks. Reinforcement Learning. Passive Reinforcement Learning. Active RL. Q-learning.
   
   SECTION 1.5.4. AUXILIIARY LECTURE SLIDES AND MATERIALS.
   
   1. Machine Learning, Decision Trees and Entropy.
      
      
   2. Decision Diagrams and Expressions. Galois and Arithmetic Logic.
      
      
   3. Machine Learning, Data Mining, Robotics.
      
      
   4. Introduction to Machine Learning.
      
      
   5. Decision Trees and Inductive Learning.
      
      
   6. Learning and Adaptation.
      
      
   7. Animal behavior and taking ideas from biology
      
      
   8. Intro to Machine Learning and Data Mining in Robotics.
      
      
   9. Famous ID3 program that learns Decision Trees. Widely used in industry. This is the basic industrial method for learning. We have several codes. Easy to program. We used in the past on the talking head projects. Read below the theory. Read theory in the Luger's book.
      
      
   10. plot.lsp
   
   
   

   SECTION 1.6. ADVANCED LISP AND PROGRAMMING EXAMPLES.

   
   
   Read the remaining of chapter 15 from Luger about LISP, if you are not done yet.
   
   SECTION 1.6.1. EXAMPLES OF PROGRAMS IN LISP.
   
   
   1. Credit Program.
      
      
   2. Expert System Shell.
      
      
   3. logic_shell
      
      
   4. newtons_method.lisp
      
      
   5. Semantic networks. Logical reasoning systems.
      
      
   6. semantic_network
      
      
   7. thermostat_simulation.clos
      
      
   8. OOPS in LISP.
      
      
   9. thermostat_simulation.oops
   
   
   

   CHAPTER 2. HARDWARE REALIZATION OF SPECIAL PROCESSORS USEFUL IN ROBOTICS.

   
   
   Please read again the Chapter 15 from Luger's book if you have not done so yet.
   
   (formerly called, "Computers for logic and logic programming" in the catalog).
   
   The entire chapter 2 is not covered this academic year. It is covered by individual projects only.
   
   
   SECTION 2.1. MANDATORY LECTURE SLIDES.
   1. Description of DEC PERLE board that uses Xilinx 3090 chips and was the best accelerator at the time of its creation. This is the board that we have. Slides about the DEC-PERLE board and its applications. Xilinx 6200 chips, ideas and realization. Software, emulator boards.
      
      
   2. DEC PERLE board as an example.
      
      
   3. More on DEC PERLE applications.
      
      
   4. XC6200A chip.
      
      
   5. Cube Calculus Machine.
      
      
   6. Basic Combinational Problems. Coloring, covering, cliques, decomposition. These are basic algorithms that have been speed up in hardware accelerators.
      
      
   7. Basic Combinational Problems. Coloring, covering, cliques, decomposition. II
      
      
   8. Basic Combinational Problems. Coloring, covering, cliques, decomposition. III
      
      
   
   SECTION 2.2. AUXILIARY MATERIALS.
   
   Material from this section is only auxiliary. It is of interest only to students like Kavitha who work on Evolvable Hardware for robot control projects.
   
   
   1. Machines of Svoboda and Sasao. August Stern. Tautology and satisfiability. Simulation Machines.
   2. Logic Design Machines. ICCAD 85.
      
      
   3. Information about DEC-PERLE-1 board and projects using it.
      
      
   4. Cube Calculus Machine.
      1. Text of Qihong Chen and Figures. Latex and eps formats.
      2. Perkowski et al, Learning Hardware using Multiple-Valued Logic. IEEE Micro 2002. In PDF Format. Part 1.
      3. Perkowski et al, Learning Hardware using Multiple-Valued Logic. IEEE Micro 2002. In PDF Format. Part 2.
      4. Cube Calculus Machine. This is a special purpose computer for solving combinatorial decision and optimization problems. Several versions exist.
      
      
      
   5. In-chip minimization of logic functions from DAC 2003.
      
   6. Computer to operate on BDDs.
   7. Evolvable Hardware - simple robot controller using FPGAs. Fall project using VHDL.
      
      
   8. Cellular Automata - simple robot controller using FPGAs. Fall project using VHDL.
      
      
   9. Decomposition Machine project. This is a computer to learn in hardware using the Ashenhurst/Curtis Decomposition.
      
      
   10. Satisfiability Machine. This machine solves the satisfiability formula in hardware. Various variants of this machine has been designed in previous classes. Code not here.
       
       
   11. Rough Set Theory Machine This machine is a cellular processor to process rough sets from the Pawlak's Theory used in Data Mining.
       
       
   12. Image Processing and Pattern Recognition computer. Processors for various image processing and pattern recognition tasks. They have been never put together to one complete system.
       
       
   13. Vision for robots.
       
       
   14. Robot. Very simple robot controller. Many useful remarks about using VHDL tools from Mentor.
       
       
   15. MACHINE FOR THE COVERING PROBLEM.
       This is a more advanced variant of the machine from the class. Similar to project above, Petrick machine and Satisfiability Machine.
       
       

SECTION 4.1. MANDATORY LECTURES
1. Please read chapter 13 of Luger's book. (it will be covered again later on and in projects).
   1. Role of Knowledge in Language Understanding.
   2. Deconstructing Language: A Symbolic Analysis.
   3. Syntax.
   4. Syntax and Knowledge with ATN Parsers.
   5. Stochastic Tools for Language Analysis.
   6. Natural Language Applications.
   
   
   
2. Agents. What is an agent? Definitions. Examples. Rationality and autonomy. Types of agents. Properties and characteristics of Environments. Simple Reflex Agent. Reflex Agent with Internal State. Brooks Subsumption Architecture from the point of view of agents. Goal-based agents. Utility-based agent. The Prisoner Dilemma again and respective agents.
   
   
3. Learning fundamentals. Types of learning. Role of trees. Examples. Neural Nets - learning without explanation. Example: Neuro-Control of a robotic arm. Applications of Neural Networks. Swarm intelligence. Possible uses of learning in natural language projects.
   
   
4. Rule-Based Systems, Natural Language, Expert Systems. Natural Language Interfaces to robots, examples. Block-world. Knowledge representation in rule based expert systems. What is knowledge. Components of an expert system. Human expert characteristics. Practical examples of expert systems. Natural Language use in Expert systems. Natural Language Recognition. Chinese room discussion. Recent trends in AI.
   
   
5. Homework. Eliza-like programs for our robots. Possible approaches to AI: think vs act, like humans vs well. Eliza. Parry. How to improve these models by adding learning, reasoning and expert knowledge. The Loebner Test. Explanation of homework. Invitation to creativity. The software developed in this homework will be next used on each of your robots, especially in your talking head robots.
   
   

SECTION 4.2. AUXILIARY MATERIALS
1. Multi-agents. Environments. Coordination. Communication and Messages. Protocols. KQML. Ontologies. Blackboard Systems. From Luger. In PDF format.
2. Luger. Distributed Agents. Distributed Problem Solving. Heterogeneous task sharing. Result sharing. Cooperation. Distributed Constrained Heuristic Search. Organizational Structuring. Planning. Frame Problem. STRIPS. Distributed Planning. Centralized versus Distributed Planning. Plan Merging. Distributed Hierarchical Planning. Hierarchical Behavior-Space Search. Distributed Planning and Negotiation. Distributed Plan Representation. Post Planning Coordination. Interleaving plan, coordination and execution.
   
   
3. Planning. I. Planning Problem. Planning versus Problem Solving. Major approaches to planning. General Problem Solver (GPS). Situation Calculus Planning. Basic representations for planning. Operator/action representation. Blocks world. STRIPS planning. Goals interaction. Sussman Anomaly. State-space planning. Plan-space planning. Partial-order planning. Least committement. Non-linear plans.
   
   
4. Planning. II. Rule-based Deduction Systems. Forward production systems. AND-OR graph search. Backward Production Systems. Information Retrieval System. Planning and the Frame Problem. Control Strategies in Planning. Triangle Table. Homework and exam problems.
   
   
5. Paper by Lenat about agents. Very good. This paper explains philosophy of the famous CYC system based on agents interaction.
   
   
6. Natural Language Processing. Objectives of Natural Language Processing (NLP). History. Results. Major issues. Ambiguity. Context. Meaning. NLP Systems. SHRDLU. LUNAR. TAUM-METEO. SYSTRAN.




CHAPTER 5. TREE-SEARCH METHODS AND ARCHITECTURES.


SECTION 5.1. MANDATORY READING ASSIGNMENTS

Please read chapter 4 of Luger's book. (Do not read section 4.3).
1. Introduction.
2. Implementing the "Best-First" Search.
3. Implementing Heuristic Evaluation Functions.
4. Heuristic Search and Expert Systemes.
5. Admissibility, Monotonicity and Informedness.
6. Complexity of search algorithms.


Please read chapter 5 of Luger's book. (Section 5.3 is not mandatory at this time).
1. Introduction.
2. Recursion-Based Search.
3. Pattern-Directed Search.
4. Recursive search example: A variant of Knight's Tour Problem.
5. Refining the pattern search algorithm.
6. Production Systems.
7. Definition and History.
8. Examples of Production Systems.
9. Control of search in production systems.
10. Advantages of Production Systems for AI.
11. The Blackboard Architecture for Problem-Solving.



SECTION 5.2. MANDATORY LECTURES


1. Uninformed Search Methods. Building goal-based agents. Problem-solving agents. States. Actions. Closed world assumption. Knowledge representation issues. Single-state and multiple-state problems. The 8 puzzle. Search formulation. Examples of Problems. Basic strategies. Depth First. Breadth first. Illustration of spaces as trees. Search Complexity. Avoidance of duplication. State space versus solution (search) space. Formalization. General search algorithms. Bookkeeping. Evaluation search strategies. Uniform cost strategy. Depth-first iterative deepening. Bidirectional search. Methodological aspects. Types of problems to be represented as search. How to apply to real-world problems. Traveling Salesman, route and VLSI design problems. The importance of search outside robotics.
   
   
2. Informed search methods. Informed methods and domain-specific information. Heuristics. Best-first search. Greedy search. Beam search. Algorithm A. Algorithm A*. Iterative Deepening A*. Automatic generation of h functions. Complexity of A* search. Iterative Improvement Search. Hill Climbing. Local and Global Maxima (Minima). Simulated Annealing. Genetic Algorithm versus search: general discussion.
   
   

SECTION 5.3. AUXILIARY MATERIALS


1. Expert systems. Knowledge-based problem solving. Expert Systems. Their characteristics. System architectures. Example of inference in expert system. Rule based reasoning. Case study: MYCIN expert system. Certainty factors in MYCIN. Rule representation. Knowledge engineering. Example Expert System Shell. Auto-Repair knowledge base. Examples in Prolog. Adding uncertainty to expert system.
   
   
2. Introduction to search. Search in Symbolic versus non-symbolic representations. Symbolic representation. Problem formulations. Reasoning, pattern matching and inferencing. Artificial Intelligence versus Natural Language. Search in NL. Differences of AI and standard computing. Types of AI computing. General characteristics of AI. Robotics as part of AI. Commercial applications of AI. Computer-Aided Instruction and Intelligent Support systems. Self-Evolving systems. Future of AI. A difficult homework related to planar geometry. (only for ambitious students).
   
   
3. Search strategies from Russell and Norvig. More examples of basic informed and uninformed strategies. Detailed analysis of Tower of Hanoi problem. Bi-directional search. Complexity of strategies.
   
   
4. Artificial Intelligence versus Robotics. Architectures. Is robotics engineering or science? Essential components of a robot. Fundamentals of robot control architectures. Reactive versus deliberative systems. Hybrid Systems. Behavior-Based Systems. Feedback Control and Cybernetics. Turtle robots of Walter Grey. Breitenberg Vehicles. History of AI and robotics. Change of paradigms. Generations of robots.
   
   
5. Advanced backtracking. In PPT format. Constraint Satisfaction Problem. Example: Map Coloring. Graph Coloring. review. Satisfiability. Applications of SAT in real world. Formal definition of a Constraint Network. Binary CSP. Crosswords. Solving constraints problems. Systematic search. Backtracking. Problems with backtracking. Consistency. Constraint propagation. K-consistency. Improving backtracking.
   
   
6. Advanced Backtracking 2. A famous example of constraints: labeling line drawings. Ordered constraint graphs. Tree-structured constraint graphs. Backtrack-free CSP. Interleaving constraint-propagation and search. Variable ordering. Value ordering. Iterative repair. Min-conflicts heuristics. Intelligent backtracking. Challenges and open problems.
   
   
7. Hardware for search. (not covered this year).
   
   




CHAPTER 6. PREDICATE CALCULUS AND RESOLUTION METHOD.


SECTION 6.1. MANDATORY READINGS.

Please read from Chapter 2 of Luger book (2002 edition) on Predicate Calculus.
1. The propositional calculus.
2. Symbols and sentences.
3. The semantics of Propositional calculus.
4. Predicate Calculus.
5. The syntax of predicates and sentence.
6. A semantics for the Predicate Calculus.
7. A "Blocks World" Example of Semantic Meaning.
8. Using Inference Rules to Produce Predicate Calculus Expressions.
9. Unification.
10. Examples of Unification.


SECTION 6.2. ADDITIONAL READINGS.

Please read Chapter 2 of Luger book (2002 edition) on Predicate Calculus.
1. Application: A Logic-Based Financial Advisor.


SECTION 6.3. MANDATORY LECTURES.


1. Wumpus World and Predicate Calculus. You can read more about Wumpus world in Russell/Norvig Book. Architecture of a knowledge-based agent. Levels of a knowledge-based agent. The Wumpus world environment. Agents in Wumpus world. Representation, reasoning and logic. The connection between sentences and facts. Logic as a Knowledge-Representation Language. Ontology and Epistemology. Propositional Logic. Semantics. Interpretation. Model. Tautology. Contradiction. Entailment. Truth Tables. Venn Diagrams. Inference rules. Soundness of Resolution Principle. Normal forms of PL sentences. Horn sentences. Entailment and derivation. Weakness of PL. Hunt the Wumpus agent.
   
   
2. First Order Logic. First Order (Predicate) Logic (FOL). Syntax. Quantifiers. Scope of quantifiers. Terms and atoms. Sentences. Translating English to FOL. Semantics of FOL. Axioms. Definitions. Theorems. Higher Order Logic. Representing Change. Situation Calculus. Frame Problem.
   
   
3. Inference in First Order Logic. Inference rules for FOL. Generalized Modus Ponens. Resolution for FOL. Converting FOL sentences to clause form. Unification examples. Unification Algorithm. Resolution refutation. Refutation Resolution Procedure. Control Strategies. Examples of Automatic Theorem Proving. Horn Clauses. Logic Programming. Prolog and issues in implementing logic programming.
   
   
4. Introduction to Knowledge Representation. In PPT Format. Knowledge representation. Types of knowledge. Semantic Data Models. Semantic Nets. Classes. Objects. Attributes. Inheritance. Encapsulation. Polymorphism. Object Orientation. Rule Representations. Types of rules. Generalizations. Hierarchical representations. Deep Knowledge. Predicates versus frames. Facets.
   
   
5. Semantic Networks. Logical Reasoning Systems. In PPT format. Forward-chaining Production Systems. Components of Production Systems. Basic Inference Procedure. Conflict Resolution Strategies. Default Reasoning. Comparing PS and FOL. Semantic Networks. Reification. Inference by Association. ISA hierarchy. Individuals and classes. Inference by inheritance. Multiple Inheritance. Nixon Diamond. Exceptions in ISA hierarchy. From Semantic Nets to Frames. Facets. Procedural attachment. Description logic. Subsumption. KR paradigms.
   
   
6. Rule Based Deduction Systems Planning. In PPT format. Rule-Based Deduction Systems. Forward Production Systems. Backward Production Systems. Homeworks. Rule-Based approach to planning. Frame problem in planning. Examples.
   
   
7. Planning. In PPT format. The Planning Problem. Planning versus Problem Solving. Typical planning assumptions. Major approaches to planning. General Problem Solver. Situation Calculus Planning. Basic representations for planning. Operator/Action representation. Blocks world. Operators. STRIPS planning. Goal Interaction. Sussman anomaly. State-space planning. Plan-space planning. Partial-order planning. Least commitment. Non-linear planning.
   
   
8. Abduction. Uncertainty. Probabilistic Reasoning. In PPT format. What is abduction. Comparing abduction, deduction and induction. Characteristics of abductive reasoning. Sources of uncertainty. Probabilistic inference. Probability of events. Bayes' Theorem. Naive Bayesian Approach. Assessment of Assumptions. Limitations of Naive Bayesian System. Bayes Belief Networks (BBN). Inference with BBN. Algorithmic Approach. Joint Tree. Stochastic simulation. MAP Problems. Noisy-OR BBN. Minimum Description Length approach. Neural Network approach. Dempster-Shafer Theory. Evidence combination. Link to Fuzzy sets and fuzzy logic. Uncertainty in rule-based systems.


SECTION 6.4. AUXILIARY LECTURES.
1. Introduction to logic, logic reasoning. Statements. Transitive syllogisms. Simple and compound statements. Translating English to propositional statements. Truth tables. Order of operations. Equivalence of metalanguage. Conditionals and their translations. Valid forms of reasoning. Quantifiers and negations.
   
   
2. Propositional Logic. A review and few new ideas. Review. Tautologies and contradictions. Propositional versus predicate calculus. Examples of using quantifiers. Logic and ambiquity. Semidecidability.
   
   
3. Example of using a propositional calculus based agent in robotics Validity. Rules of inference. Complexity of propositional inference. Wumpus world.
   
   
4. Representation and logic. Satisfiability. Soundness. Model. Interpretation. Completeness. Resolution.
   
   
5. First Order Logic Translation from English to FOPC. Interpretation. Representing change. Block world. Problems with formalization.
   
   
6. Resolution Based Automatic Theorem Proving. Resolution examples. Resolution for question-answering. Factoring. Equality. Resolution strategies. Logic Programming.
   
   




CHAPTER 7. LOGIC PROGRAMMING AND ARTIFICIAL INTELLIGENCE IN ROBOT SYSTEMS.


SECTION 7.1. MANDATORY READING.

Please read from chapter 12 in Luger's book.
1. Introduction to weak methods in Theorem Proving.
2. The General Problem Solver and Difference Tables.
3. Resolution Theorem Proving.
4. Producing the Clause Form for Resolution Refutations.
5. PROLOG and Automated Reasoning.
6. Logic Programming and Prolog.


SECTION 7.2. ADDITIONAL READING.

Please read from chapter 12 in Luger's book.
1. The binary Resolution Proof Procedure.
2. Strategies and Simplification Techniques for Resolution.
3. Answer Extraction from Resolution Refutations.
4. Further Issues in Automated Reasoning.
5. Uniform Representation for Weak Method Solutions.
6. Alternative Inference Rules.
7. Search Strategies and their use.



SECTION 7.3. MANDATORY LECTURE SLIDES.


1. Introduction to logic programming language Prolog
   
   
2. More Prolog
   
   

SECTION 7.4. AUXILIARY MATERIALS.



1. Representation of problems using Prolog.
   
   
2. Prolog Environment.
3. METHODS OF FORMULATING AND SOLVING PROBLEMS IN LOGIC PROGRAMMING.
4. UNIFICATION IN HARDWARE.
5. MACHINES FOR LOGIC PROGRAMMING AND ARTIFICIAL INTELLIGENCE.
6. Warren Abstract Machine.
7. Japanese Fifth Generation Computer Program.
8. Data Bases.
9. Search.
10. Natural Language Processing.
11. unification.lisp
12. Ulug and Bowen.
13. Matching instead unification.
14. Lisp Machines.
15. Prolog Machines.
16. Parallel Prolog Machines.
17. CAM from MIT.
18. CBM in ATR Japan.




CHAPTER 8. FUZZY LOGIC AND FUZZY LOGIC MACHINES.



SECTION 8.1. MANDATORY READING.
1. Read section 8.2.2. Reasoning with Fuzzy Sets from Luger.




SECTION 8.2. MANDATORY LECTURE SLIDES.
1. Decomposition, Formal Synthesis and Minimization of simplified fuzzy logic circuits. Efficiency of decomposition. Minimization of fuzzy functions. Graphical method for minimization. Algebraic identities. Transformations. Differences between Boolean Logic and Fuzzy logic. Approaches to fuzzy logic decomposition. Method by conversion to multiple-valued logic. Fuzzy functions ternary maps. Operators in maps, conversions. Canonical and non-canonical forms. Entire method illustrated on examples. Generalization of the Ashenhurst-Curtis decomposition (review). Column compatibility, coloring, functions versus relations. Examples from real life. Learning from medical data bases. Variable ordering, vacuous variables. Subsumption rule for fuzzy logic minimization. Fuzzy decision trees and fuzzy multiplexers.
2. Applications of Fuzzy Logic Expert Systems.
   
   SECTION 8.3. AUXILIARY MATERIALS.
   
   
   1. Lecture on Fuzzy Sets and Logic.
      
      
   2. Lecture on Fuzzy Logic Fundamentals.
      
      
   3. Lecture on Fuzzy Logic System Examples.
      
      
   4. Lecture on Fuzzy Logic Minimization and Decomposition.
      
      
   5. Lecture on Fuzzy Logic Decomposition.
      
      
   6. See Marek Perkowski's publications for year 2003. Article from International Conference on Fuzzy Information Processing from Beijing, China.
      
      
   7. Texas Instruments, ``Enhanced Control of Alternating Current Motor Using Fuzzy Logic and a TMS320 Digital Signal Processor''. Application Report. 1996. In PDF format.
      
      
   8. S. Guo, L. Peters, and H. Surmann, ``Design and Application of an Analog Fuzzy Logic Controller,'' Paper in PDF format.
      
      
   9. S. Ghosh, Q. Razouqi, H.J. Schumacher and A. Celmins, ``A Survey of Recent Advances in Fuzzy Logic in Telecommunications Networks and New Challenges,'' IEEE Trans. on Fuzzy Systems.
      
      
   10. E. Tunsel and Mo Jamshidi, ``On Embedded Fuzzy Controllers,'' 1st World Automation Congress. 1994.
       
       
   11. Fuzzy Controllers.
   12. Papers of Marek Patyra.
   13. Analog realization of Fuzzy Logic.
   14. Mamdani.
   15. Sendai trains.
   
   
   

   CHAPTER 9. REASONING BY ANALOGY AND INDUCTION.

   
   In some years, this unit is taught instead of unit 1.5, called "Introduction to Machine Learning." In years that both units are taught, you should skip reading the two first lectures below about decision trees and decision diagrams.
   
   SECTION 9.1. MANDATORY LECTURES.
   
   Review again chapter 9 from Luger and especially trees induction and section 9.5.4. on Analogical Reasoning.
   
   1. Decision Trees
      
      
   2. Information theory in learning. Trees, expressions. Advanced Material. Taught only sometimes in detail, when useful in a project.
      
      
   3. Reasoning types. In PPT format.
      
      
   4. Types of reasoning. In PPT format.
      
      
   5. Reasoning. In PPT format.
      
      
   6. Reasoning by analogy. In PPT format.
      
      
   7. Reasoning by analogy in Prolog. In PPT format.
      
      
   
   SECTION 9.2. AUXILIARY MATERIALS:
   1. Functional Decomposition Machine.
      
      
   2. Physical Design for EHW.
      
      
   3. auxiliary1.ppt
      
      
   4. auxiliary2.ppt
      
      
   5. auxiliary3-Analogical-learning.ppt
      
      
   6. Marek Perkowski et al, ``Use of Machine Learning based on Constructive Induction in Dialogs with Robotics Heads''. Proc. ICORR 2003, the International Conference on Rehabilitation Robotics, Daejeon, Korea.
      
      
   7. Marek Perkowski and Stanislaw Grygiel, ``Decomposition of Relations: A New Approach to Constructive Induction in Machine Learning and Data Mining - An Overview.'' Proc. Telecommunication Conference, Poland. May 2001.
      
      
   8. Slides to the above.
      
      
   9. B. Steinbach, A. Mishchenko, M. Perkowski, ``An Algorithm for Bi-Decomposition of Logic Functions,'' Proceedings of DAC 2001.
      
      




CHAPTER 10. ARTIFICIAL NEURAL NETWORKS AND ASSOCIATIVE PROCESSORS.



This year we are going to skip entirely neural networks and associative processors during the Winter quarter. We will cover them in Spring quarter.

SECTION 10.1. MANDATORY LECTURES.


1. Connectionist Approach.
2. Lecture on associative processors.

SECTION 10.2. AUXILIARY MATERIALS.
1. Realization of modern associative processors.
2. Abduction. Uncertainty. Probabilistic Reasoning.