Artificial Intelligence

CSCI 440, fall semester, 2008

 

To the Bottom of the Page

 

Instructor:        Dr. Shieu-Hong Lin

Class:      T R 10:30-11:45 am at Lib 141

Office Hours: Tuesday, Thursday 2:00-4:00 pm, Math & CS department

 

Course Syllabus

 

About weekly progress report: (Download this template)

By Thursday each week, you should spend around 5~10 minutes to

  • Add the latest progress made (from last Thursday to this Thursday) into the report, and
  • email the report to Dr. Lin.

 

*************************************************************************************************

Week 1. Could the computer ¡§act¡¨ or ¡§think¡¨ like human beings?  Progress report due Thursday, Sept. 4.

 

HAL and the Construction of Artificial Intelligence.

 

Reading #1: due Thursday, Sept. 4.

  1. Watch a YouTube video on computer vision and automatic recognition of facial emotion expression and read the introduction section of a Wikipedia article on automatic facial recognition.
  2. Watch the impressive YouTube videos on the impressive Honda Asimo humanoid robot like 1 or 2 and read a Wikipedia article on Asimo and its cognition capability.
  3. Watch two YouTube videos on autonomous intelligent computer-controlled vehicles: (i) in a desert setting (DARPA 2005 challenge) and (ii) in an urban setting (DARPA urban challenge 2007) and then read Sections 1~ 4 of this Wikipedia article on DARPA urban challenge 2007.
  4. Read the introduction section in this Wikipedia article on automatic translation and the introduction section in this Wikipedia article on speech recognition. Play with this online automatic Google translation service and their FAQs to get a sense of the state of the art in automatic translation.
  5. Read at least the first three sections (up to the section on computers versus human) of this Wikipedia article on computer chess programs about computer chess programs, watch this YouTube document about Deep Blue, and read this article about GNU Chess (Optional: Download and play with GNU Chess).
  6. Read the introduction and history sections in this Wikipedia article on the Turing test. Take an online Turing test here. Download and play with the chatter-bot Claude.

 

Homework #1: Reflection on AI due Thursday, Sept. 4.

To what extent, do you think the imagination in Stanley Kubrick¡¦s ¡§2001: a space odyssey¡¨ has been or has NOT been realized? Do you think the relationship between the software (computer programs) and the hardware of a computer is very much like the relationship between the mind and the body of a human being? Why or why not? Do you agree or disagree with them some of the sample reflections quoted here. Put down at least 500 words of your opinions (For example, could the computer software such as the chatter-bot in this lab and the intelligent chess programs implement rationality, emotion, unpredictability, or will to some extent? Are the human genomic sequences to some extent like the ¡§executable code¡¨ in control of human behavior? ) Include your opinions into the accumulative progress report when you email me your accumulative progress report.

 

*************************************************************************************************

 

Week 2. Speech recognition and Hidden Markov models (HMMs). Progress report due Thursday, Sept. 11.

 

Reading #2: due Thursday, Sept. 11.

  • Thoroughly read and do things as described in this spelling recognition project to gain an understanding of speech recognition from that of spelling recognition.
  • Read Sections 1.1~2.1 in Introduction to Data Mining.

 

 

Homework #2: An automatic game-playing agent due Thursday, Sept. 11.

 

 

*************************************************************************************************

 

Week 3. More on HMMs and introduction to machine learning. Progress report due Thursday, Sept. 18.

 

Reading #3: due Thursday, Sept. 18.

  • Examine the following 8 text files: A, B, C, D, E, F, G, and H (or download them as a zip file together) resulting from four persons Johnny, Winnie, Manny, and Cathy with following typing characteristic parameters (Johnny, Winnie, Manny, and Cathy) typing the Biola vision text two times each. Try to determine for each text who types the text. Come up with a plan for automatic identity recognition using the demo source code provided in spelling recognition project.
  • Read Sections 2.2~2.5 in Introduction to Data Mining.

 

 

Homework #3: First, read and understand the application of HMMs in the foreign traveler scenario depicted in the homework and the use of the Viterbi-like algorithm to calculate key probabilistic information as demonstrated in the homework. Then use the algorithms to answer three questions asked in the homework. Due Thursday, Sept. 18.

 

 

 

*************************************************************************************************

 

Week 4. Machine learning & decision tree induction + more on search. Progress report due Thursday, Sept. 25.

 

Reading #4: due Thursday, Sept. 25.

  • Read the Wikipedia article on alpha-beta pruning and browse Russell¡¦s AIMA lecture slides on search and the related subject.
  • Read Sections 3.1~4.3 in Introduction to Data Mining.
  • Download and install WEKA, browse this introductory PPT tutorial on WEKA, download and unzip this zip file to get the zoo data set and the Iris data set, apply the classifier J48 under the WEKA explorer (in the tree section of the classifier menu) to the two datasets, and report in your progress report the decision trees you found above for the zoo data set and the Iris data set.
  • You can also download uci-20070111.tar.gz from the WEKA website to see many of the public UCI and UCI KDD datasets.

 

 

Homework #4: (i) For each document d in {A, B, C, D, E, F, G} and each person p in {Johnny, Winnie, Manny, and Cathy}, calculate the logarithm of the probability Pr(d/p) that document d is the resulting text when person p types the Biola vision text. (ii) Based on the probabilities calculated in (i) above, for each document d determine the top 2 most likely persons that may have generated the document d. Due Thursday, Sept. 25.

 

 

*************************************************************************************************

 

Week 5. Machine learning: Overfitting and performance evaluation. Progress report due Thursday, Oct. 2.

 

Reading #5: due Thursday, Oct. 2.

  • Read the Wikipedia article on alpha-beta pruning and browse Russell¡¦s AIMA lecture slides on search and the related subject.
  • Read Sections 4.4~4.6 in Introduction to Data Mining.
  • Download and install WEKA, browse this introductory PPT tutorial on WEKA, download and unzip this zip file to get the zoo data set and the Iris data set, apply the classifier J48 under the WEKA explorer (in the tree section of the classifier menu) to the two datasets, and report in your progress report the decision trees you found above for the zoo data set and the Iris data set.

 

 

Homework #5: From the exercises in Chapter 2, Eddie: exercises 4X, Harrison exercises 4X+1, Hee-Chun: 4X+2, Matthew: 4X+3.  Do at least 4 out of the 7 exercises assigned to you. Discussion: Tuesday, Sept. 30. Due Thursday, Oct. 2.

 

 

*************************************************************************************************

 

Week 6. More on classification and machine learning. Progress report due Thursday, Oct. 9.

 

Reading #6: due Thursday, Oct. 9.

  • Automatic reasoning with rule-based knowledge representation: (i).Browse through this paper on automatic reasoning about events with time-interval information to get a sense of how a computer like HAL may conduct planning and reasoning about causes and consequences. (iii) Read the skeleton source code and play with a fully function executable and a sample instance in testInstance.txt zipped in this zip file.
  • More on classification and machine learning: Read Sections 5.1~5.3 in Introduction to Data Mining on rule-based classifiers, nearest-neighbor classifiers, and Bayesian classifiers.

 

 

Homework #6: From the exercises in Chapter 4, Eddie: exercises 4X, Harrison exercises 4X+1, Hee-Chun: 4X+2, Matthew: 4X+3.  Do at least 2 out of the 3 exercises assigned to you and pick at least one more from other exercises assigned to others. Discussion: Tuesday, Oct. 7. Due Thursday, Oct. 9.

 

 

 

*************************************************************************************************

Week 7. Machine learning and its applications. Progress report due Thursday, Oct. 16.

 

Reading #7: due Thursday, Oct. 16.

Homework #7: Turn in a summary of the assigned papers. Discussion: Tuesday, Oct. 14 and Thursday, Oct. 16.

 

 

*************************************************************************************************

 

Week 8. Torrey Conference & the Midterm

 

  • Data mining project assignment 1: Do and report your progress in the following tasks. (i) Convert the datasets into the arff format and visualize the datasets using WEKA. (ii) Derive discreteized versions of the datasets by discretizing the numerical attributes under WEKA.  Due: Tuesday, Oct. 21.
  • Take-home Midterm (due: Thursday, Oct. 30).

 

 

*************************************************************************************************

 

Weeks 9~10. Basics of intelligent agents and knowledge representation.

 

Reading #9: due Thursday, Nov. 13.

 

Data mining project assignment 2: Do and report your progress in the following tasks. (i) Use the WEKA explorer to conduct classification experiments using clarifiers such as various Bayes classifiers, nearest-neighbor classifiers, rule-based classifiers, and decision tree classifiers. Try to balance the number of not-retained cases in the training dataset by duplicating them to bump the number of such cases up. (ii) What are the accuracies of the classification evaluated based on cross validation, or based on the hold-out scheme, or based on the test dataset? Are there problems of overfitting? If so, what are the evidences?  Due: Thursday, Nov. 13.

 

 

 

*************************************************************************************************

 

Weeks 11~13. More on Logic, knowledge representation, and intelligent agents; Thanksgiving recess.

 

Reading #11: Due Tuesday, Nov. 25.

 

Data mining project assignment 3: Based on what you have done so far, develop a mini manual describing the whole procedure of (i) cleaning up the raw data set as an Excel worksheet, (ii) selection of attributes (i..e. removal of some columns in the Excel worksheet) , (iii) transforming it into an appropriate file in the arff format file, (iv) building a classification model using at least one of the classifiers provided by WEKA, (v) the steps for applying the model to classify new ceases, and (vi) a summary of the empirical results you have seen. Due Tuesday, Nov. 25.

 

Homework #8: Linear Regression. Due Thursday, Dec. 4.

 

*************************************************************************************************

 

Weeks 14~15. Association analysis and Clustering

 

Reading #14: Due Thursday, Dec. 11.

  • Read Chapter 8 of Introduction to Data Mining on clustering.

 

Take-home exam on knowledge representation and automatic reasoning

 

Discussion of chapter 6 and chapter 8: Week 15.

 

Data mining project report: Wrap up your data mining project assignments 1~3 as a report.

 

Answers to Homework#4: here. (Congratulations to Eddie for the 100% accuracy of prediction.)

 

*************************************************************************************************

 

 

 

To the Top of the Page