Instructor: Neeraja J. Yadwadkar
neeraja@austin.utexas.edu
Pronouns: she/her/hers
Lectures:
When? MW 9-10:30am
Where? ECJ 1.318.
Office Hours:
When? 15 minutes immediately after every lecture, and weekly one hour (the weekly office hour time will be communicated close to the beginning of the semester).
Where? On zoom (Please find the link on Canvas)
Course Description:
The objective of this new class is to empower research at the intersection of Computer Systems and Machine Learning (ML) – ML for Systems and Systems for ML.
Today, ML is powering key applications ranging from face recognition to autonomous vehicles. ML is powerful but there are many challenges in deploying these models in the real-world while achieving the desired performance. Recent advancements in software and hardware platforms have made rapid progress in ML algorithms. In this class, we will study the latest developments in computer systems keeping in mind emerging new ML algorithms, and applications that are driven by these ML models. We will also discuss the use of ML techniques for optimizing various aspects of systems. Some essential background on ML and systems will be covered in class. We will be reading and discussing papers from recent conferences that focus on (a) using ML for various optimizations across the systems stack, and (b) building systems for enabling emerging ML algorithms/models to be deployed in the real-world.
We will cover a variety of topics, including (tentative list):
- Big data systems
- Domain-specific architectures and hardware for ML
- DL frameworks/compilers
- Scheduling and resource management for DL workloads
- Challenges in ML deployment
- Deep RL and challenges in deploying in real-world
- ML in Cloud/Databases/Networking
- Systems challenges in DL training and inference serving
The format of the class will be a mix of lectures, and paper presentations by students followed by discussions. The initial few classes will be lectures that will focus on understanding fundamentals in Systems and ML. Then the class format will turn into a mix of lectures, and paper presentations by students followed by discussions. Students will be required to complete the reading assignments and submit a short review answering questions such as “What is the problem solved in the paper? What is the key finding? What is the key result? What are the limitations? How can we build on this work?”
The key component of this class will be a research project in a group of 2-3 students. Ideally, project groups should contain both systems and ML students. I encourage students to bring their current research to the class as their research projects. The aim for these projects is to produce a short (5 pages) report and optionally a demo. An ideal project should potentially lead to a workshop or conference publication at a good venue. I will provide guidance in selecting the idea and preparing the findings. I strongly encourage starting early on the projects! We will have two project presentations by students: one mid-session (describing the problem, related work, and initial results/measurements) and one at the end of the session (discussing preliminary results, current status, and next steps). We will also hold a short poster session for these projects where we will invite students and faculty members interested and/or working in relevant areas of research. This will be a good opportunity for students to get feedback on their projects.
Prerequisites:
- Proficiency in coding (Python, C/C++, or the language of your choice for building a system or a component using ML as your class project)
- Basic knowledge of ML (at least one introductory class in ML/Statistics/probability/optimization/data science)
- Basic knowledge of Computer systems (at least one introductory class on OS/Databases/Networking/Cloud computing/Distributed Systems/Architecture/Embedded systems).
If you haven’t already done so, please fill out this form providing information about prereqs.
Grading:
- Class participation: 15%
- Paper reviews: 10%
- Paper presentation: 15%
- Project proposal: 10%
- Mid-session project presentation: 15%
- Final project presentation + report: 35%
There will be no exams for this class.
Course Website: We will use the web-based course management system “Canvas”. The students are responsible for regularly checking the course web page for announcements and postings at https://utexas.instructure.com/
Drop Policy: The last day to drop this course without permission from the Dean is the 4th class day. After this day, drops are approved only in the case of health or personal problems. An engineering student should make an appointment with his/her departmental advisor to discuss adding or dropping any course if the change will alter the classes that were originally approved by the departmental advisor. If the add or drop requires the approval of the Dean, then the student will need to schedule an appointment with an Academic Advisor in the Office of Student Affairs, ECJ 2.200 (471-4321) to discuss the request. Additional information can be found at: http://www.engr.utexas.edu/current/policies/pol_add- drop-wdraw.cfm
Academic Dishonesty: Cheating will not be tolerated and will be dealt with according to the policy established by the office of the Dean of Students.
Disability and Access: The university is committed to creating an accessible and inclusive learning environment consistent with university policy and federal and state law. Please let me know if you experience any barriers to learning so I can work with you to ensure you have equal opportunity to participate fully in this course. If you are a student with a disability, or think you may have a disability, and need accommodations please contact Disability & Access (formerly Services for Students with Disabilities). Here are some examples of the types of diagnoses and conditions that can be considered disabilities: Attention-Deficit/Hyperactivity Disorders (ADHD), Autism, Blind & Visually Impaired, Brain Injuries, Deaf & Hard of Hearing, Learning Disabilities, Medical Disabilities, Physical Disabilities, Psychological Disabilities and Temporary Disabilities. Please refer to the D&A website for contact and more information. If you are already registered with D&A, please deliver your Accommodation Letter to me as early as possible in the semester so we can discuss your approved accommodations and needs in this course.
Religious Holy Days: In accordance with section 51.911 of the Texas Education code and University policies on class attendance, a student who misses classes or other required activities, including examinations, for the observance of a religious holy day should inform the instructor as far in advance of the absence as possible so that arrangements can be made to complete an assignment within a reasonable period after the absence. A reasonable accommodation does not include substantial modification to academic standards, or adjustments of requirements essential to any program of instruction. Students and instructors who have questions or concerns about academic accommodations for religious observance or religious beliefs may contact the Office for Inclusion and Equity. The University does not maintain a list of religious holy days.
COVID-19 Guidance: To help preserve our in-person learning environment, UT Austin recommends the following:
- Adhere to university mask guidance.
- Vaccinations are available, free and not billed to health insurance.
- Self-test kits are available and free.
- Visit Protect Texas Together for more information
Title IX Reporting: Title IX is a federal law that protects against sex and gender-based discrimination, sexual harassment, sexual assault, sexual misconduct, dating/domestic violence, and stalking at federally funded educational institutions. UT Austin is committed to fostering a learning and working environment free from discrimination in all its forms where all students, faculty, and staff can learn, work, and thrive. When sexual misconduct occurs in our community, the university can:
- Intervene to prevent harmful behavior from continuing or escalating.
- Provide support and remedies to students and employees who have experienced harm or have become involved in a Title IX investigation.
- Investigate and discipline violations of the university’s relevant policies.
Faculty members and certain staff members are considered “Responsible Employees” or “Mandatory Reporters,” which means that they are required to report violations of Title IX to the Title IX Coordinator at UT Austin. I am a Responsible Employee and must report any Title IX related incidents that are disclosed in writing, discussion, or one-on-one. Before talking with me, or with any faculty or staff member about a Title IX related incident, be sure to ask whether they are a responsible employee. If you want to speak with someone for support or remedies without making an official report to the university, email advocate@austin.utexas.edu. For more info about reporting options and resources, visit the campus resources page or e-mail the Title IX Office at titleix@austin.utexas.edu.
Mental Health Counseling: College can be stressful and sometimes we need a little help. Luckily, we have a wealth of resources and dedicated people ready to assist you, and treatment does work. The Counseling and Mental Health Center provides counseling, psychiatric, consultation, and prevention services that facilitate academic and life goals and enhance personal growth and well-being. Counselors are available Monday-Friday 8am-5pm by phone (512-471-3515) and Zoom.Alternatively, you can talk to Mr. Armando Sanchez, LCSW, right here in the College of Engineering. Mr. Sanchez is our CARE Counselor and he can be reached at 512-471-3741.
If you are experiencing a mental health crisis (e.g. depression or anxiety), please call the Mental Health Center Crisis line at 512-471-CALL(2255). Call even if you aren’t sure you’re in a full-blown crisis, but sincerely need help. Staff are there to help you.
Student Rights and Responsibilities:
- You have a right to a learning environment that supports mental and physical wellness.
- You have a right to respect.
- You have a right to be assessed and graded fairly.
- You have a right to freedom of opinion and expression.
- You have a right to privacy and confidentiality.
- You have a right to meaningful and equal participation, to self-organize groups to improve your learning environment.
- You have a right to learn in an environment that is welcoming to all people. No student shall be isolated, excluded, or diminished in any way.
With these rights come responsibilities, you are responsible for
- taking care of yourself, managing your time, and communicating with the teaching team and others if things start to feel out of control or overwhelming.
- acting in a way worthy of respect and respectful of others.
- creating an inclusive environment and speaking up when someone is excluded.
- holding yourself accountable to these standards, holding each other to these standards, and holding the teaching team accountable as well.
Your experience with this course is directly related to the quality of the energy that you bring to it, and your energy shapes the quality of your peers’ experiences.
Personal Pronoun Use: Professional courtesy and sensitivity are especially important with respect to individuals and topics dealing with differences of race, culture, religion, politics, sexual orientation, gender, gender expression, gender variance, and nationalities. Class rosters are provided to the instructor with the student’s legal name, unless they have added a “preferred name” with the Gender and Sexuality Center. Canvas provides an opportunity to select a pronoun preference. I will gladly honor your request to address you by a name that is different from what is on the roster, and by the gender pronouns you use (she/he/they/ze, etc.).
Official Correspondence: UT Austin considers e-mail as an official mode of university correspondence. You are responsible for following course-related information on the course Canvas site.
Safety Information: If you have concerns about the safety or behavior of students, TAs, Professors, or others, call the Behavioral Concerns Advice Line at 512-232-5050. Your call can be anonymous. If something doesn’t feel right, it probably isn’t. Trust your instincts and share your concerns. Occupants of buildings are required to evacuate buildings when a fire alarm is activated. Alarm activation or announcement requires exiting and assembling outside.
- Familiarize yourself with all exit doors of each classroom and building you may occupy. The nearest exit door may not be the one you used when entering the building.
- Students requiring assistance in evacuation shall inform their instructor in writing during the first week of class.
- In the event of an evacuation, follow the instruction of faculty or class instructors. Do not re-enter a building unless given instructions by the following: Austin Fire Department, UT Austin Police Department, or Fire Prevention Services.
- Information regarding emergency evacuation routes and emergency procedures.
- More safety information.
Sanger Learning Center: More than one-third of undergraduates use the Sanger Learning Center each year to improve their academic performance. All students are welcome to join their classes and workshops and make appointments for their private learning specialists, peer academic coaches, and tutors. For more information, see the Sanger Web site or call 512-471-3614 (JES A332).