CS490U: Generative AI Tools & Applications

CS490U: Generative AI Tools & Applications (Fall 2023)

Introduction

Generative AI tools are becoming increasingly powerful and widely used in various industries, from art and entertainment to healthcare and finance. These tools have the ability to generate realistic images, sound, videos, and text, and can even create new and original contents. However, with great power comes great responsibility, and there are limits to what generative AI tools can and should do. In this course, students will explore the capabilities and limitations of generative AI tools, and learn how to use them responsibly and ethically. Throughout the course, students will have the opportunity to use various generative AI tools and work on a term project of their choosing, applying the principles and techniques they have learned to a real-world problem or application. They will present their project at the end of the term and engage in a critical discussion of the implications of their work. By the end of the course, students will have gained a deep understanding of the capabilities and limitations of generative AI tools, and will be equipped with the skills and knowledge necessary to use these tools to solve real world problems.

Instructor:

Prof. James Won-Ki Hong (PIAI 337) : 054-279-2244, jwkhong(@)postech.ac.kr

Lectures:

We will meet on Mon. & Thu. 17:00-18:55 in Room 306, Mueunjae Memorial Hall for lectures, Q&A, hands-on tutorials, discussions and student presentations. The lecture materials can be found in POSTECH LMS PLMS. We may occasionally use Vmeeting (https://vmeeting.postech.ac.kr/postech/GenAI) for online class when needs arise.

Course Schedule:

The following is a tentative course schedule for this course.

Teaching Assistant (TA):

Joonwon Jang (Graduate School of AI), 010-4502-8808, kaoara(@)postech.ac.kr
Jiho Ryu (Graduate School of AI), 010-5434-0271, johoryu(@)postech.ac.kr

Pre-requisites:

There is no pre-requisite for this course. All you need is an open mind and strong desire to learn and collaborate with peers.

Required Textbook:

There will not be a textbook for this course. My lecture slides and various materials (videos, papers, software, presentations, etc.) found from the Internet will be used for the course.

Suggested Materials:

Videos:
- ChatGPT: - Introduction, Tutorial 1, Tutorial 2
- GPT-4: Introduction
- The Evolution of ChatGPT From GPT-1 to GPT-4
- Promt Engineering: Tutorial

Articles & Books:

Evaluation:

Evaluation on each student will be done based on the following:
1. Assignments - 40%
2. Quizzes - 5%
3. Term Project - 45%
4. Class Participation - 10%
Note: the above evaluation scheme may change slightly during the course.

Term Project:

There will be a major term project (worth 45% of the final mark) on the use of Generative AI tools. The project topics will be discussed in class. Students will be asked to prepare, submit and present materials (Powerpoint & Word) related to the project throughout the course.
1. Term Project Ideation (worth 2%, out: Sept. 12, 2023, due: 5pm, Sept. 18, 2023) (submit project ideation assignment in PLMS Week #3 Board so that everyone can see other students submissions)
2. Project Proposal Preparation (worth 5%, out: Sept. 21, 2023, due: midnight Oct. 06, 2023 & class presentation: 5pm Oct. 09, 2023)
3. Project Requirements & High-Level Design Presentation (5%, Oct. 19 & Oct. 23, 2023)
  - Use this guideline to prepare a requirements document for your group project
4. Detailed Design (5%, due: midnight Nov. 12, presentations Nov. 13 , 16 & 20 , 2023). Please enter your preference on presentation time (served based on FCFS)
5. Final Term Project Deliverables:
  - Powerpoint Slides (10-20 pages) & Demo video (2 minutes or less) (Midnight, Dec. 15, 2023) & Final Presentations (Dec. 16, 2023, 9am-5pm, E2-102)
  - Final Report Document (4-6 page Technical Paper in English in the IEEE 2-column format -due: Midnight, Dec. 22, 2023). Click here for an example technical paper.

Assignments:

There will be a few assignments (worth 40% of the final mark). You should submit your assignment materials to PLMS.
- Assignment 1 (1%) (out: Sept. 4, due: 5pm, Sept. 14)
- Assignment 2 (7%) (to be done in pairs)(out: Sept. 14, due: 5pm, Sept. 25)
- Assignment 3 (7%) (to be done in pairs)(out: Sept. 25, due: 5pm, Oct. 16)
- Assignment 4 (11%) (to be done in pairs)(out: Oct. 23, due: 5pm, Nov. 11)
- Assignment 5 (7%) (to be done in pairs)(out: Nov. 6, due: midnight, Nov. 20)
- Assignment 6 (7%) (to be done in pairs)(out: Nov. 23, due: 5pm, Dec. 7)
Note: the above assignment schedule may change slightly during the course.
Late assignments may be handed in, but there will be a penalty of 20% of the mark for assignments turned in less than one day late, and an additional penalty of 10% for each day thereafter.
Cheating Policy -- Cheating will not be tolerated in this course. Students are encouraged discuss things related to courses and assignments but the materials handed in must be his/her own. The maximum penalty for the first offense is for the assignment in question. For subsequent offenses may result in an automatic failure of the course and possibly other academic punishments.

Class Participation:

Students are strongly encouraged to attend all lectures and do the quizzes. Students are also strongly encouraged to participate in discussions (questions & answers) during classes. 10% of the final mark is assigned for good and active class participation.

Topics Covered:

The following is a tentative list of lecture topics for the course.

Introduction
- Introduction to the course
- Introduction to Generative AI tools
Overview of Generative AI
- Introduction to Generative AI
- Introduction to Large Language Models
ChatGPT
- Introduction to ChatGPT
- Applications of ChatGPT
Google Bard
- Introduction to Google Bard
- Applications of Google Bard
Media generation AI & Prompt Engineering
- Introduction to Media generation AI & Applications
- Introduction to Prompt Engineering
Stable Diffusion
- Introduction to Stable Fusion for generating photo-realistic images given text input
- Applications of Stable Fusion
DALL-E 2 & 3
- Introduction to DALL-E 2 for creating realistic images and art given text input
- Applications of DALL-E 2
Reading Week
Midjourney
- Introduction to Midjourney for creating artwork given text input
- Applications of Midjourney
Runway
- Introduction to Runway for generating video clips based on text prompts
- Applications of Runway
Riffusion
- Introduction to Riffusion - generating music using images of sound rather than audio
- Applications of Riffusion
Application of Generative AI to Automotive Industry
Chatsonic
- Introduction to Chatsonic - alternative to ChatGPT, integrates with Google Search to create content with the latest information.
- Applications of Chatsonic
AdCreative.ai
- Introduction to AdCreative.ai for creating advertisements
- Applications of AdCreative.ai
Project Development and Preparation for Presentation, Report & Demo
Project Presentation & Demo

Note: Generative AI tools mentioned above may change during the course.

Dr. James Won-Ki Hong

Professor

Dept. of Computer Science and Engineering

Pohang University of Science and Technology (POSTECH)

Pohang, Korea

Tel: +82 54 279 2244

Fax: +82 54 279 5663

Email: jwkhong{@}postech.ac.kr

Last modified: Aug. 18, 2023 This page is maintained by J. W. Hong. If you have any questions or suggestions, please send email to jwkhong at postech.ac.kr.