Syllabus

The syllabus is still being finalized and is subject to major changes.

Table of contents

Prerequisites

This is a class that deals with advanced topics in the field of systems and AI. Therefore, it is very important that you have the necessary background to succeed in this class. For this course, the prerequisites are:

  • CS 6210 - Advanced Operating Systems (strict)
  • Strong system building skills and proficiency with C/C++/Java/Scala/Python programming (strict)
  • Experience with machine learning training/inference frameworks and tools (e.g., PyTorch/TensorFlow/vLLM/Triton) (strict)
  • Ability to work on medium to large code bases (strict)
  • CS 8803 SMR - Systems for Machine Learning (highly recommended)
  • CS 7210 - Distributed Systems (recommended)

Textbook & Exams

This course has no textbooks or exams. The course content will be based on recent papers from top venues in the areas of systems and AI.

Format

This course will be offered in a graduate seminar format, where students will lead paper discussions. Being a graduate seminar, the majority of the grade will be determined by a research project. Students will be expected to work on a substantial, original research project that combines systems and LLMs. The ultimate goal of this project is to produce high quality work that can potentially be published in a top-tier systems conference such as SOSP, OSDI and NSDI.

To foster a deeper understanding of the papers and encourage critical thinking, we will utilize the following format:

Required Reading

Each lecture will have a certain number of required readings. Everyone in the class is expected to read them. There will be one or more optional related reading(s) that the presenter(s) should be familiar with. They are optional for the rest of the class.

Student Lectures

The course will be conducted as a seminar. Each student will be assigned at least one paper to present over the course of the semester. Only the students assigned will present in each class. Since there will be more than one required reading, there will be more than one student presenter assigned each day. Your task is to collaborate with the other person(s) assigned for the day and come up with a cohesive presentation that should last at most 45 minutes without interruption. Note that you must expect questions and interruptions during your presentation. The idea is to leave enough room to have class discussions.

You are free to come up with a single presentation, or separate presentations. However, the goal of your presentation must be the following:

  • Provide the necessary background and motivate the problem. Note that each lecture has a “theme” such that the papers are connected in some way. For instance, perhaps they are trying to solve the same problem using different approaches, or maybe one is building on top of the other. Your presentation should try to make this connection.
  • Present the high level idea, approach, and/or insight (using examples, whenever appropriate) in the required reading as well as the additional reading.
  • Discuss technical details so that one can understand key details without carefully reading.
  • Explain the differences between related works.
  • Identify strengths and weaknesses of the required reading and propose directions of future research.

The slides must be provided to the instructor team at least 24 hours prior to the corresponding class. Late submissions will not be accepted.

If you miss a lecture where you are required to present, it is your responsibility to find a replacement before the day of the lecture.

Role-based Discussion

Each lecture will be followed by a panel discussion. The goal of this discussion is to simulate a top conference’s program committee meeting, but with a twist. In real program committee meetings, the authors are not present. However, in our class, we will simulate the presence of the authors who can defend their work by using roles:

Authors

The students who present and the students who scribe each day will play the role of the authors.

Reviewers

Just like in a typical systems conference, you will be assigned as a reviewer (i.e., you will play the role of a Program Committee member) for several of the required papers in the reading list. Reviewers critically assess the paper, posing challenging questions and highlighting potential weaknesses or areas for further investigation. Your goal is to engage in a constructive critique of the paper, simulating a peer review scenario.

Each reviewer will review their pile of assigned papers, and write a detailed review for each. The review must be comprehensive, pointing out the strengths, the weaknesses, constructive feedback on how to improve the paper on the weaknesses you pointed out and areas for further improvement. You should submit your reviews before the lecture. Late submissions will not be entertained.

In addition to writing the reviews, you must come to the class when a paper you reviewed is being presented, and participate in a simulated PC meeting. If you miss a lecture where you are required to be a reviewer, it is your responsibility to find a replacement before the day of the lecture.

Others

The rest of the class must participate in the paper discussions. You are expected to actively ask questions and engage.

Participation & Attendance

Given the discussion-based nature of this course, participation is required both for your own understanding and to improve the overall quality of the course. You are required to attend all lectures (if you have a legitimate reason to miss a class, contact us in advance for approval), and more importantly, participate in class discussions. If you fail to attend all the lectures (except the ones where you received our permission to miss) and contribute towards the class discussions, you will forfeit the entire grade allocated for attendance & participation. This semester, we are giving every student two (2) permissions by default. You can use these two permissions to miss any two student-led lectures (you cannot use it for (a) lectures where you are assigned some task, (b) staff-led lectures or (c) mid-point or final project presentations) if you like; you don’t have to explicitly reach out to us. In addition to in-class participation, you are also expected to engage and contribute towards the online discussions in Ed forums.

Project

You will have to complete substantive work on an instructor-approved problem and have original contribution. Projects will be done in groups of 1, 2 or 3 students. Please note that the default is groups of 2 students; therefore groups of 1 and 3 must be approved by the instruction team. Typically, groups of 1 and 3 are for PhD/MS thesis students working towards their thesis where this class project becomes a part.

Although they are important research directions, surveys, hardware design, algorithmic/theoretical work and simulations are not permitted as projects due to the Systems for AI focus of this class. Your project proposal must be approved by the instructor before it can become your official project. The instructor’s decision in this regard will be final.

You must meet the following milestones (unless otherwise specified in future announcements) to ensure a high-quality project at the end of the semester:

  • Form a group of 1, 2 or 3 members and declare your group’s membership by August 29. After this date, we will form groups from the remaining students.
  • Turn in a 2-page draft proposal by September 5. Remember to include the names and email addresses of the group members.
  • Each group must present mid-semester progress during class hours on October 1 and October 8.
  • Each group must turn in an 8-page final report and your code on or before December 8 23:59ET. The report must be submitted as a PDF file, with formatting similar to that of the papers you’ve read in the class. This format can be achieved using USENIX templates.
  • In addition to your report, you must submit a working version of your code. Your code submission must contain a README file with step-by-step instructions on how to compile and run the provided code and replicate the key result(s) in your report. If this requires specific datasets, you must include the datasets in your submission (or link them if they are open-source). You must submit a working version of your code to receive the allocated grade.

Tentative Grading

Course grades will be determined according to the following criteria:

  • Paper presentation: 10%
  • Paper reviews: 15%
  • Participation (including attendance): 15%
  • Project report & working code: 45%
    • Proposal: 5%
    • Final report: 20%
    • Working code with instructions: 20%
  • Project presentation: 15%
    • Mid-semester presentation: 5%
    • Final presentation: 10%

Subject to Change

Due to the highly dynamic situation (e.g., global phenomena outside of the instructor’s control, conference travel, etc), the syllabus and course schedule may be subject to change. It is the responsibility of students to check Canvas, Ed Discussion, email messages, and course announcements (through course canvas and Ed Discussion) to stay up-to-date with any course logistics changes. We will make every effort to communicate changes via these mechanisms. The course is held IN RESIDENCE until and unless announced otherwise on course Ed Discussion or Canvas. Virtual option will not be available this semester.

Communication

We will be using Ed Discussion for asynchronous communication. ALL communication regarding this course must be via Ed forums. This includes questions, discussions, as well as private messages.

Auditing Policy

This class does not offer an audit option. To be fair to the large waitlist of interested students, you will be removed from the class if you enroll in the audit option.

Special Needs and Services

If you are a student with learning needs that require special accommodation, contact the Office of Disability Services at (404) 894-2563 or http://disabilityservices.gatech.edu/, as soon as possible, to make an appointment to discuss your special needs and to obtain an accommodations letter. Please also e-mail me as soon as possible in order to set up a time to discuss your learning needs.

Your mental health and well being is important to us. If you are having a difficult time or just need to talk, please make use of campus counseling services. These services are available in the Smithgall Student Services building (Flag Building) between 8:00 am and 4:00 pm, Monday through Friday. There is also a 24/7 service for students in crisis. Again, do not wait till the end of the semester to get help.

Conduct

Be respectful of one another as well as the instructors both in class and out. Be considerate and professional in online Ed Discussion posts, emails, and other conversation related to the course.

Please help to provide a distraction-free learning environment in class, and be considerate to your teachers and other students by minimizing in-class distractions such as cellphone use, eating, etc.

Please plan to attend the lectures. Students are responsible for all material covered in class as well as any announcements made in class (as well as on Ed Discussion and Canvas). We make our best effort to put information on Ed Discussion or Canvas but cannot promise everything said in class will make it on there.

Academic Integrity

Georgia Tech aims to cultivate a community based on trust, academic integrity, and honor. Students are expected to act according to the highest ethical standards. For information on Georgia Tech’s Academic Honor Code, please visit http://www.catalog.gatech.edu/policies/honor-code/ or http://www.catalog.gatech.edu/rules/18/.

Any student suspected of cheating or plagiarizing will be reported to the Office of Student Integrity, who will investigate the incident and identify the appropriate penalty for violations.

Statement of Intent for Inclusivity

As a member of the Georgia Tech community, I am committed to creating a learning environment in which all of my students feel safe and included. Because we are individuals with varying needs, I am reliant on your feedback to achieve this goal. To that end, I invite you to enter into dialogue with me about the things I can stop, start, and continue doing to make my classroom an environment in which every student feels valued and can engage actively in our learning community.