CSE352
ARTIFICIAL INTELLIGENCE
FALL 2011
Course Information
News:
PROJECT PRESENATIONS START TUESDAY December 6 and continue December 8-
last day of classes.
These are 5-10 minutes presentations of your results.
Please send me e-mail to reserve a spot.
I also need a SHORT few pages of Project results Description. You can e-mail me- or bring to class.
NEW Lecture Notes on RESOLUTION L15, L16, L17 POSTED
THANKS for wonderful PRESENTATIONS!
NO FINAL!!! HAve a GEAT FINALS week.
HOMEWORK 3 Solutions are posted; they are solutions submitted by
some of our students in class; good standard!
Time:
Tuesday, Thursday, 2:20 - 3:40 pm
Place:
Humanities 3019
Professor:
Anita Wasilewska
1428 CS Building; 632-8458
e-mail: anitaatcs.sunysb.edu
Office Hours: Tuesday, Thursday at 1 - 2 pm, and by appointments.
Teaching Assistant:
No TA
Book:
The Essence of Artificial Intelligence
Allison Cawsey
Prentice Hall, 1998
General Course Description:
Artifficial Intelligence is a broad and well established field.
The AI textbooks seem to be getting longer and longer. Our cource textbook
attemts to reverse this trend. It provides a concise and accessible
introductionto the field.
The course will closely follow the book and is designed to give
a broad, yet in-depth overview of different fields of AI. It will
examine the most recognized techniques in a more rigorous detail.
For this part we will provide detailed lecture notes distributed in class and
available on the web.
It also will explore the newest trends and developments of the
field in form of students talks based on newest research and applications
from the
field.
The course outcomes and catalog description are in the official course description page.
Student Information
Students ATTENDENCE is essential for the course as students
presentations are integral and as important part of the course
design as Professor's lecture. I will take attendance.
Project Data
Play around with the data and familiarize yourself with it (DOWNLOAD: bakarydata.xls )
Project Description
Project Description
PROJECT HOMEWORK
PROJECT DESCRIPTION SLIDES
Project Tools
WEKA
RapidMiner
Past Project Examples
PROJECT Presentation Example 1
PROJECT Presentation Example 2
PROJECT Presentation Example 3
PROJECT Presentation Example 4
Students Input
Here are the links for the Loebner Prize.
Please send me more, if you find some interesting pages, or articles.
Loebner Prize Main Page
Minsky Comments
2005 Contest
2005 Winner (Jabberwacky)
PRESENTATIONS
Please email Professor
the subject of your presentation as soon as possible.
Presentation PROPOSAL is due THURSADY, September 22
You can discuss it with Professor earlier before you submit the final version. It must contain a TITLE and a short few paragraphs long description.
DATES for presentations are "FIRST COME, FIRST SERVE",
so fix the date of your presentation as soon and possible. You don't need to have a TITLE to reserve the date of the presenation.
PRESENTATIONS SCHEDULE
PRESENTATIONS DAY 1: Tuesday, October 11
Presentation 1
PRESENTER: Sebastian Bryk
TITLE:
Watson: Next Generation AI.
My topic will be on the super computer Watson developed by IBM it is a Question-Answering Systems for answering queries posed in the natural language English.
I believe this computer and the team which created it is a pioneer in AI technology, for a machine to be able to answer question in a natural language is and compete
successfully on a game show is amazing.
Therefore I will focus on the learning algorithms used, the development resources needed to create such a machine, this will include time invested, hardware used, overall design of
the super computer etc. I will try to find out as much information as possible on Watson's ability to learn from his game show challenges and how he adapts to categories offered on the
game show jeopardy. I shall also include information on how he processes and translates natural language in order to arrive at an answer, then give the correct answer in the proper format.
Presentation slides:
Presentation 2
PRESENTER: Xu Fei Jiang
TITLE: "Artificial Intelligence in Medicine"
This presentation will discuss how AI is use in Medical field, and focus primarily on the system called CBR, also known as Case Based Reasoning, how it utilizes Ai to be to aid the doctor in problem solving and decision making support in fields related to health science.
CBR is an artificial intelligence approach that utilize past experience to solve current problems. CBR has proven to be especially applicable to problem solving and decision support in fields related to health science. The reason why CBR are needed are: medicine is a highly data intensive field, a human doctor will never be able to remember the details about all medicine, where a programming can remember every data in its data table and retrieve that piece of information when needed, biological system such as human body is difficult to describe by general models, it can acts as a general framework for clinicians when the proper headlines are available.
Presentation slides:
PRESENTATIONS DAY 2: Thursday, October 13
Presentation 1
PRESENTER: Alexander Niculescu
TITLE: Searle's Chinese Room Argument.
PRESENTATIONS DAY 3: Tuesday, October 25
Presentation 1
PRESENTER: William Fischer
TITLE: Deep Blue
Presentation slides:
PRESENTATIONS DAY 4: THURSDAY, November 3
Presentation 1
TITLE: rts path finding
PRESENTER: David Hayman
Presentation slides:
Presentation 2
PRESENTER: Joseph Scarabino
TITLE:
Presentation slides:
Presentation 3
PRESENTER: Amanpreet Mukker, Victor Angueira
TITLE: Siri software that is on the new iPhone 4S
Presentation slides:
PRESENTATIONS DAY 5: THURSDAY, November 10
Presentation 1
PRESENTER: Anthony "Tony" Arra
TITLE: Vocaloids/Speech-synthesis
Presentation slides:
Presentation 2
PRESENTER: John Paul Pennisi
TITLE: Genetic Algorithms
I would like to do my project on genetic algorithms. Compared to symbolic AI systems
which are very static genetic algorithms can evolve and solve a variety of problems. For a
symbolic AI system if the problem for which the system could solve somehow changes the
system could have a very hard problem adapting to the changes. Genetic algorithms on the other
hand are better yet suited for changes in the problem.
Genetic algorithms are based on natural biological evolution. The architecture of the systems
that implement genetic algorithms is more adapt to a wide range of problems. It works by first
creating a large set of possible solutions to a given problem. Then it evaluates each of these
solutions and assigns it an appropriate fitness level. This is easily comparable to the survival of
the fittest in biological evolution. The solutions then breed new solutions and the parent breeds
that were deemed more fit have a higher chance of reproducing. Overall the solutions evolve
over time. If programmed correctly genetic algorithms can be incredibly efficient.
Applications of genetic algorithms are immense since any problem that has a large search
domain could be suitably tackled by genetic algorithms. Some organizations that use genetic
algorithms are the military that use them to differentiate between different radar returns and
stock companies use genetic algorithm powered programs to predict the stock market. AI Biles
also use genetic algorithms to filter out ‘good’ and ‘bad’ riffs for jazz improvisation.
Presentation slides:
Presentation 3
PRESENTER: Rishabh Raj
TITLE: Artificial Intelligence in Finance
Lately, AI has been playing a major definitive role in the financial markets throughout the globe. Instead of trading taking place by humans, computers armed with highly sophisticated algorithms and AI techniques have taken over to trade in these markets amounting to trillions of dollars in trading volumes daily.
Thus, in my presentation, I would start by giving a very brief overview about this industry and will then then elaborate on a trading algorithm that has been developed by me over the course of the last one year. My algorithm, which uses binomial decision trees and other AI techniques has been developed as a beta version and coded in Java. Although, it has not been deployed yet; I plan to do it soon.
I would also like to spawn upon other major algorithms being currently used in the industry to explain the what, why and how of this trillion dollar industry where AI plays a very significant part.
PRESENTATIONS DAY 6: THURSDAY, November 17
Presentation 1
PRESENTER: Joshua Belanich
TITLE: Cooperative Multiagent Learning
Cooperative Multiagent Learning is when you have multiple intelligent agents attempting to solve a task. However, a key constraint is that no agent can know everything all the other agents know. Therefore they must communicate with each other to solve their common goal.
For this project I will cover what is Multi-Agent learning, why it is important, and overview multiple learning techniques coming from reinforcement learning, game theory, and evolutionary computing. I will also cover modern applications of Multi-Agent learning.
Presentation slides:
Presentation 2
PRESENTER: Sammy Mohamed
TITLE: Natural Language Processing
I am interested in doing research on how machine learning algorithms and statistics applies to the construction of programs that "understand" language. (Additionally on what it means to understand language and what is difficult about this process) To be specific I am currently working on understanding the techniques used to implement word sense disambiguation, semantic translation, textual summarization and machine translation.
Presentation slides:
PRESENTATIONS DAY 7: THURSDAY, December 1
Presentation 1
PRESENTER: Nicholas Lucchesi
TITLE: AI in First Person Shooter Games
First person shooter games are getting more complex as they are created for newer consoles. Enemies need to be “smarter” as players are gaining skill in this type of game. They need to be proactive in actions rather than reactive.
In most games, the AI is a reactionary AI system. In this, the computer player or bots would react based upon what the player does not make their own initial decisions. Players commonly notice the real difficult style of play is playing against each other. To resemble playing against another human, bots must act upon information they currently have and not wait for an action of a player.
This can be simulated in many ways. The bot could begin less intelligent and as the game goes on, it can asses the style the player uses whether it be hiding behind things or charging right at them and change its strategy based upon that info. Another way this is done is to give the bot global knowledge about everything in the map and they can use a difficulty modifier to affect decisions. The bot should also have a database of actions that can be accessed that ranks different actions based upon their effectiveness.
Presentation slides:
Presentation 2
PRESENTER: Ed Linero, Nicholas Trombetta
TITLE: AI in 2D Side-Scrolling Video Games
Presentation slides:
Presentation 3
PRESENTER: Waseem Mir
TITLE:
Artificial Intelligence in Bioinformatics
The amount of experimental data generated in the field of molecular biology is increasing very rapidly due to automated and massively parallel techniques of DNA sequencing, such as micro arrays. This vast amount of information is spread throughout many databases, making it ideal for processing by the data mining and knowledge discovery techniques of artificial intelligence. In particular, AI techniques have been used for sequence reconstruction of entire genomes, for identifying families (clusters) of related genes, and also for identifying genes which may be involved in disease pathways.
Presentation 3
PRESENTER: Isaac Kaplan
TITLE: Facial Recognition
Presentation slides:
PRESENTATIONS DAY 8: TUERSDAY, December 6
Presentation 1
PRESENTER: Hiep Truong
TITLE: Japanese AI technology development
Japanese has come a long way since their first ambitious investment into AI technology field in 1980s. Although experts believe that Japanese does not quite success in completing their initial goals but their achievements are still unmatched and they are undoubtedly the world’s pioneer of the field. In my presentation, I will focus in showing the progress of Japanese’s AI development from its start in 1980 to the present day. The presentation will also cover in details Japanese Actroid (a type of humanoid robot) and its AI. My main goal of the presentation will be informing and drawing interested to the ground breaking development of the A.I field.
PROJECT PRESENTAIONS: 1
PROJECT PRESENTATIONS 2: December 8
Josh Belanich
John Paul Pennisi
These are short, 5-10 minutes presentations of results of your project.
AI TALKS in the DEPARTMENT
tba
Homeworks and Tests Schedule
Homework 1 due Tuesday, September 20
Homework 2, Part 1, 2 due Tuesday, October 13
Homework 3, Part 1,2 due Thursday, November 3
Homework 3, Part 3 due Thursday, November 10
Homework 4 (Resolution) due December 1
PROJECT Homework (YOUR PROJECT DATA) due Tuesday, November 22 (10 extra points)
PROJECT PRESENTATIONS (5-10 minutes): December 6,8
FINAL is a take home test and due on the official day of FINALS or on any day
before.
BUT AS THE CLASS WORKED VERY WELL - NO FINAL!
DOWNLOADS
2011 SYLLABUS
TAKE HOME FINAL- NO FINAL!!!
HOMEWORKS
Homework 1
Homework 2, Part 1
Homework 2, Part 2
Homework 3 Classifier
Homework 4 - Resolution
Homeworks Solutions
Homework 1 Solutions
Present STUDENT Homework 1 Solutions
Past STUDENT 1 Homework 1 Solutions
Past STUDENT 2 Homework 1 Solutions
STUDENT 1 Homework 2 Solutions
STUDENT 2 Homework 2 Solutions
STUDENT 3 Homework 2 Solutions
Homework 3 Solutions 1
Homework 3 Solutions 2
Homework 4, Solutions to be posted
Past Students Presentations
Natural Language Processing
Deep Blue
Fuzzy Logic and its Applications
Autonomous Vehicles
AI in Computer Vision; Past, Present and Future
AI in Chess Playing
Genetic Algorithms
Computer Vision and Facial Recognision
Lecture Notes:
Lecture 1, Chapter 1; Introduction to AI
Lecture 2, Chapter 2; Knowledge Representation
Lecture 3, Chapters 2; Predicate Logic (1)
Lecture 4, Chapters 2, 3; Rule Based Systems (1)
Lecture 5, Chapters 2, 3; Rule Based Systems (2)
Lecture 6, Chapter 7; Introduction to Learning
Lecture 7, Classification-Supervised Learning, Part 1
Lecture 7, Classification-Supervised Learning, Part 2
Lecture 8, Decision Tree, Classification Part 3
Lecture 9, General Majority Voting Examples, Classification Part 4
Lecture 10, BASIC Decision Tree Algorithm, Classification Part 5, Handaut 2
Lecture 11, Testing, Classification Part 6
Lecture 12 Data Preprocessing
Lecture 13, Classification By Neural Networks, Handout 3
Lecture 14, Genetic Algorithms
Bayesian Classification (optional)
Lecture 15, Resolution, Part 1, Handout 4
Lecture 16, Resolution Strategies, Part 2
Lecture 17, Resolution Strategies, Part 3
Lecture 18, Predicate Logic, Part 2
DATASETS
Datasets for learning, data mining and knowledge discovery
University California Irvine KDD Archive
World Bank datasets
ACADEMIC INTEGRITY STATEAMENT
Each student must pursue his or her academic goals honestly and be
personally accountable for all submitted work. Representing another
person's work as your own is always wrong. Any suspected instance of
academic dishonesty will be reported to the Academic Judiciary. For
more comprehensive information on academic integrity, including
categories of academic dishonesty, please refer to the academic
judiciary website at
Academic Judiciary Website
Stony Brook University Syllabus Statement
If you have a physical, psychological, medical, or learning
disability that may impact your course work, please contact
Disability Support Services at (631) 632-6748 or
Disability Support ServicesWebsite
They will determine with
you what accommodations are necessary and appropriate. All
information and documentation is confidential.
Students who require assistance during emergency evacuation are
encouraged to discuss their needs with their professors and
Disability Support Services. For procedures and information go to
the following website:
Disability Support Services Website