CSC 872: Pattern Analysis and Machine Intelligence
Spring 2024 (#8110, 3 units)
|
Instructor: Dr. Kazunori Okada |
|||
|
Lec. Session |
Tue: 4:00 - 6:45 pm |
Lec. Location |
Hensill Hall 543 |
|
Office Phone |
(415) 338-7687 |
Office |
Thornton Hall 911 |
|
Office Fax |
(415) 338-6826 |
Office Hours |
Thr: 11:30 am - 12:30 pm |
|
Email Address |
Web Page |
http://online.sfsu.edu/~kazokada/ |
|
|
Mailing Address |
Computer Science Department, San Francisco State
University |
||
|
Teaching Assistant |
Yoshimasa Iwano |
TA Office Hour TA Office |
Wed: 2:00 pm – 3:00 pm Online (see Canvas) |
|
Lecture Plan (subject to change) |
|||||||
|
Week |
Sub |
Topic: Lecture |
Topic: Exercise |
Notes |
Readers |
Assignments |
Dues |
|
01:01/30 |
INT |
Introduction: PAMI Frameworks |
Project Discussion |
Ch.1 |
|
||
|
02:02/06 |
AI |
Agent-based AI Framework |
MATLAB Exercise 1 |
Ch.2 |
|
|
|
|
03:02/13 |
AI |
Problem Solving: Search Methods |
MATLAB Exercise 2 |
Ch.3-4 |
HW 1 (lec01-03) |
|
|
|
04:02/20 |
AI |
Knowledge Rep.: Propositional Logic |
MATLAB Exercise 3 |
Ch.7-8 |
|
|
|
|
05:02/27 |
AI |
Knowledge Rep: First-Order Logic |
Fast Prototype 1: Modeling: PCA 1 |
Ch.9 |
HW 2 (lec04-05) |
|
|
|
06:03/05 |
PR |
Bayesian Framework |
Fast Prototype 1: Modeling: PCA 2 |
Ch.13-14, 20 |
|
|
|
|
07:03/12 |
PR |
Statistical Modeling: Non-Parametric |
Fast Prototype 1: Modeling: PCA 3 |
Ch.20 |
HW 3 (lec06-07) |
|
|
|
08:03/19 |
PR |
Statistical Modeling: Parametric |
Fast Prototype 2: Segmentation: MS 1 |
Ch.20 |
|
P-Topic (3/19) |
|
|
03/26 |
SPRING BREAK |
----- |
---------- |
------------ |
|
||
|
09:04/02 |
PR |
Statistical Modeling: Mixture Models |
Fast Prototype 2: Segmentation: MS 2 |
|
HW 4 (lec08-09) |
|
|
|
10:04/09 |
ML |
Machine Learning Framework |
Fast Prototype 2: Segmentation: MS 3 |
Ch.14, 18 |
|
|
|
|
11:04/16 |
ML |
Supervised Learning: Classification |
Fast Prototype 3: Classify: LDA 1 |
Ch.3 |
HW 5 (lec10-11) |
|
|
|
12:04/23 |
ML |
Supervised Learning: Regression |
Fast Prototype 3: Classify: LDA 2 |
Ch.3 |
|
|
|
|
13:04/30 |
NN |
Neural Network: Functional Learning |
Fast Prototype 3: Classify: LDA 3 |
Ch.20 |
|
|
|
|
14:05/07 |
NN |
Neural Network: Deep Learning 1 |
Neural Network: Deep Learning 2 |
|
|
Presen (5/12) |
|
|
15:05/14 |
CON |
Project Final Presentation 1 |
Project Final Presentation 2 |
|
|
Final-R (5/14) |
|
Basic Information
Course Summary:
This course offers an introduction to the modern artificial intelligence: the pattern analysis and machine intelligence (PAMI) studies. This research field ranges over a wide variety of well-established subjects, including artificial intelligence (AI), pattern recognition (PR), machine learning (ML), neural network (NN), all of which are intricately related to each other because of the shared underlying concepts and theories. Collectively they contribute to various practical applications, such as graphics/animation, games, factory automation, robotics, video analysis/security, medical imaging, bioinformatics, data mining, to name a few. The main goal of this course is to develop an intuitive understanding of the various fields, as well as to understand differences between them due to specific historical and application domain biases. Through lectures together with hands-on prototyping exercises, you will learn not only about a number of fundamental and useful PAMI techniques but also lessons on how to successfully conduct a research project.
Objectives:
· Learn the foundation of PAMI studies:
o Artificial intelligence
o Pattern recognition
o Machine learning
o Neural network and Deep Learning
· Learn the general concepts across various fields:
o Data and knowledge representation
o Problem formulation
o Problem solving
· Become familiarized with basic algorithms in PAMI practice.
· Become exposed to various application domains and basic literature in PAMI studies.
Prerequisites:
Graduate standing or consent of the instructor.
Office Hours:
Zoom links for office hours for both the instructor and the TA are given in the course Canvas page. Office hours will be available for online conversations during the chosen office hour time specified above. Students are requested to make yourself visible by turning the video on. Students will be met in first-come first-served basis. Please email the instructors ahead of time to reserve a specific time slot of 10-20 minutes, should you have important issues to consult with.
Final Project
Literature Survey Report & Presentation:
An independent literature survey project on deep learning is to be carried out by each student. This assignment provides you a hands-on exercise for conducting literature review and presentation toward preparing your own thesis and publication. Each student must choose appropriate representative articles (approved by the instructor before Mar 19), conduct short presentation (on May 14), and submit a survey report due on the last class meeting (on May 14). Your work will be graded based on the quality and completeness of your presentation and report. Late policy specified below will apply. Read the assignment in the above lecture plan table for more details.
In-Class Exercises
MATLAB Tutorials/Exercises:
These tutorials/exercises offer you opportunity to learn MATLAB: a popular numerical computing software tool which will be used during our fast-prototyping exercises. The instructors will introduce you to basics of the tool and guide you through some hands-on exercises. Please prepare your own laptop with an installation of MATLAB. A free copy of MATLAB is available for all SFSU students via https://athelp.sfsu.edu/hc/en-us/articles/360011475074-Getting-MATLAB-for-students. Please consult the instructor upon questions about this tool. Tutorials given via zoom will be recorded and made accessible as your study resource in the Canvas page when they become available.
Fast-Prototyping Exercises:
These in-class hands-on programming exercises are designed to help you learn how to implement and test an algorithm quickly, and to familiarize you with practical computer vision applications and their solutions. Three classic algorithms (PCA for face recognition, Mean Shift for clustering/segmentation, and LDA for classification) will be implemented by you. One reference paper will beb provided for each algorithm to familiarize you with their theory. The instructor will provide data, followed by a brief description of tasks to be tackled. Students are to complete a software prototype during class meetings with minimum preparation. Please bring your own laptop with an installation of MATLAB. Please refer to the above description on how to acquire the MATLAB copy.
Exams/Homework
Exams:
No midterm/final examinations. J
Homework:
Five homework assignments will be given as take-home exams. Their schedule is given in the above lecture plan table. You will be asked to answer questions and solve quantitative analytical problems. Some homework involves solving difficult pen&paper analytical problems. You are advised not to procrastinate. Unless specified otherwise, each homework is due in one week.
Canvas Usage:
Distribution and submission of your assignments will be handled through the Canvas course page. After the instructor made it available, click the specified Canvas link to view each homework assignment. Work on your problems and write down your answers on your own papers (make sure you write your name and SID# for each page). After you complete it, sort your answer papers according to the problem numbers and scan them by using a PDF-Scan tool (e.g., CamScanner, Scannable, GeniusScan, AdobeScan, Multifunction Printer, etc) and save it to a single PDF file, combining all pages into one file. After checking that all pages are present and readable, submit your PDF file by using the specified Canvas assignment submission link. Always double-check what you submitted for its completeness and readability.
Grading Policy
Numerical Grade Weights:
· 50%: Homework
· 25%: Final Report
· 10%: Final Presentation
· 15%: Fast Prototyping
Grading Policy:
· If you like to appeal your homework or test grades, you must do so within one week after graded assignments were made available to students. There will be no exception even if you miss those classes and/or announcements. You are responsible to find out your own grades.
· The default grade distribution is as follows: A (100% - 92.5%), A- (92.4%-90%), B+ (89.9% - 87.5%), B (87.4% - 82.5%), B- (82.4% - 80%), C+ (79.9% - 75%), C (74.9% - 65%), C- (64.9% - 60%), D+ (59.9% - 57.5%), D (57.4% - 52.5%), D- (52.4% - 50%), F (49.9% - 0%).
Late Policy:
· Late submission is not allowed in general unless an agreement due to documented special circumstances was reached with the instructor before the due date.
· Late submission for the project proposal and report will be penalized by 10% per day up to 40%.
· After 4 days, late final report receives zero credit.
Course Materials
Course Web Page:
https://bidal.sfsu.edu/~kazokada/csc872/
Text Book:
Artificial Intelligence: A Modern Approach (4th, 3rd, 2nd Ed), Russell SJ and Norvig P, Prentice Hall, 2020, 2009, 2002
Recommended Readers:
Pattern Classification (2nd Ed), Duda RO, Hart PE, Stork DG. Wiley-Interscience, 2000 (PR, ML, NN)
http://www.wiley.com/WileyCDA/WileyTitle/productCd-0471056693.html
Neural Network for Pattern Recognition, Bishop CM. Oxford University Press, 1996 (NN, PR, ML)
http://portal.acm.org/citation.cfm?id=525960
The Elements of Statistical Learning (2nd Ed), Hastie T, Tibshirani R, Friedman JH. Springer, 2016, 2003 (ML)
http://www.springer.com/computer/ai/book/978-0-387-84857-0
Digital Image Processing (4th, 3rd, 2nd Ed), Gonzalez RC, Woods RE. Prentice Hall, 2017, 2007, 2002 (Imaging)
Rules
Honor Code:
Your continued enrollment in this class indicates that you have carefully and entirely read and pledge to abide by the Honor Code published in the Canvas page of this course and accept the consequences to violations of its terms. You will be asked to acknowledge this Honor Code in Canvas page. For example, usage of online cheating sites, such as a use of Chegg.com and CourseHero.com, is strictly forbidden. Violation of this rule will be treated seriously without any leniency.
Syllabus is Subject to Change:
This syllabus and schedule are subject to change. The official syllabus will be maintained at the course website. It is your responsibility to check on the site frequently and check on announcements made while you were absent.
Absence:
Regular attendance is recommended. In the event of an absence, it is the student’s responsibility to learn of any material missed, including announcements made by the instructor. Lectures and demonstrations will not be repeated during office hours. In case of extraordinary circumstance, it is the student’s responsibility to inform the instructor and submit supporting documents as soon as s/he can.
In-Class Communication Device Rule:
Any ringing devices such as cell phones and must be turned-off during the lectures. Students with a ringing device will be asked to leave the class meeting immediately and not allowed to come back for the class meeting.
Academic Integrity & Plagiarism:
Academic integrity, the ethical presentation of one's own work in accordance with the rules established for this class, is required. Instances of academic misconduct will be reported to the College in which the course is housed, the Division of Graduate Studies (if a graduate student), and the Office of Student Conduct with the report being kept in those offices until a student earns his/her degree. Any instances of cheating, deceit, fabrication, forgery, plagiarism, unauthorized altering of records or submitting false documents, unauthorized collaboration, unauthorized submission of work previously given credit, or other forms of academic misconduct will be assigned a grade penalty, likely an F or a grade of zero. Failing one or more assignments or examinations for reasons of academic integrity violations may result in a final class grade of F. Students may not withdraw from classes in which they have committed academic misconduct. Consequences for violations of academic integrity may exceed an F on the assignment, examination, or class as determined by the Academic Integrity Review Committee.
Members of our academic community have a responsibility to develop an awareness of academic integrity, to cultivate skills to realize honesty in academic and community work, and to sustain actively academic honor as a core value of our community. Students are expected to engage in behaviors that reflect well upon the university. In addition to attending to one's own actions, the Standards for Student Conduct require that students who witness academic dishonesty notify their faculty/instructor, department chair, or the Office of Student Conduct. Supporting academic integrity enhances the reputation of the University and the value attributed to degrees awarded by the University.
I encourage discussion among students, but I expect each student to hand in their original work. You are responsible for doing your own work and for ensuring that your work is protected from copying. Usage of online cheating sites, such as a use of Chegg.com and CourseHero.com, is strictly forbidden. Violation of this rule will be treated seriously without any leniency. Your continued enrollment in this class indicates that you have carefully and entirely read and pledge to abide by the Honor Code published on the Canvas page of this course and accept the consequences to violations of its terms. Violation of the university and departmental rules (found in below links) is a serious offence and can result in severe penalties. It is your responsibility to familiarize yourself with the following rules:
· SFSU Code of Student Conduct: http://conduct.sfsu.edu/standards
· Academic Dishonesty: http://conduct.sfsu.edu/academic-dishonesty
· Plagiarism: http://conduct.sfsu.edu/plagiarism
· Computer Science Department Policy: https://cs.sfsu.edu/student-policies#accordion-collapse-accordion-824-2
Important Resources
Learning Assistances:
The Tutoring and Academic Support Center (TASC) is a new university-wide center that supports the academic success of all San Francisco State students. They are offering services online via Zoom. Please email tutoring@sfsu.edu from your SF State email address with your name, student ID, course for which you are seeking tutoring, and available days/times for an appointment. They will reply with details for your online appointment. More information can be found at https://ueap.sfsu.edu/tutoring.
· Phone Number: 415-405-5516
· Location: Tutoring is currently being offered online via Zoom
· Hours: Mon-Thr, 9:00 a.m. to 7:00 p.m. and Fri, 9:00 a.m. to 2:00 p.m.
· Email: tutoring@sfsu.edu
Religious Holidays:
Reasonable accommodations will be made for you to observe religious holidays when such observances require you to be absent from class activities. It is your responsibility to inform the instructor during the first two weeks of class, in writing, about such holidays.
Disability Access:
Students with disabilities who need reasonable accommodations are encouraged to contact the instructor. The Disability Programs and Resource Center (DPRC: https://access.sfsu.edu/) is available to facilitate the reasonable accommodations process. The DPRC is located in the Student Service Building and can be reached by telephone (voice/415-338-2472, video phone/415-335-7210) or by email (dprc@sfsu.edu).
Student Disclosures of Sexual Violence:
SF State fosters a campus free of sexual violence including sexual harassment, domestic violence, dating violence, stalking, and/or any form of sex or gender discrimination. If you disclose a personal experience as an SF State student, the course instructor is required to notify the Title IX Coordinator by completing the report form available at http://titleix.sfsu.edu, emailing equityprograms@sfsu.edu or calling 338-2032.
To disclose any such violence confidentially, contact:
· The SAFE Place - (415) 338-2208; https://psyservs.sfsu.edu/content/safe-place
· Counseling and Psychological Services Center - (415) 338-2208; http://psyservs.sfsu.edu/
· For more information on your rights and available resources: http://titleix.sfsu.edu
Kazunori Okada © 2024, All rights are reserved.