General Information
- Instructor: Majid Komeili
- Email: majid.komeili@carleton.ca
- Time:
Thursday 10:05 am - 12:55 pmMondays 2:35 pm - 5:25 pm - Location: Zoom
- Office hours: By Appointment
Announcements
- Assignment 2 is posted on Brightspace.
- The schedule for paper presentations is posted here.
- A list of topics and related papers is posted here. Please submit your prefered papers by Monday September 16, 2024.
- Slides from lecture 1 are posted in the course schedule here.
- Welcome to COMP5801! We look forward to meeting you on
Thursday September 5 at 10:05 amMonday September 9 at 2:35 pm over Zoom. - Subsequent to the previous announcment sent out through Brightspace, the class time has changed as noted above.
- You can find previous offering of the course COMP 5900 here. Note that the course code has changed from COMP 5900 to COMP 5801.
Description
Machine learning (ML) is the scientific study of algorithms and statistical models that computers use in order to perform a specific task effectively without using explicit instructions, relying on patterns and inference instead. It is seen as a subset of artificial intelligence. This course will cover advanced topics in machine learning such as deep learning including CNNs, RNNs, GANs, Transformers, large language models, retrieval augmented generation, deep clustering, transfer learning, domain adaptation, few-shot learning, zero-shot learning, self-supervised learning and Interpretability of ML methods. The format of the course will be a mix of lectures and paper presentations.
Course Information
Prerequisites
You are expected to have a reasonable background in machine learning and be familiar with probability, statistics, linear algebra, calculus and Python.Evaluation
- Assignments: (25%).
- Paper presentations: (20%).
- Class participation and discussion(5%).
- Final Project: (5% proposal, 5% in class presentation, 40% Report).
Course Materials
There is no required text for this course. Notes will be posted periodically in the course schedule.Paper presentations
- Submit your prefered papers for approval before the Deadline. Late submissions will be penalized 10% per day.
- This includes choosing two papers from ML journals and conferences and presenting them in class. Please consult this page for topics and a list of related papers.
- Ideally these are the same as the key papers that you are going to build your course project on. Though, they can be different from your project. They can be on different topics.
- You will present one paper throughout the term. If your first preferred paper was already taken or was not qualified, I’ll go with your second preferred paper.
- Presentation should be around 20 minutes followed by about 15 minutes discussion. You should come up with a set of questions to foster a 15-minute discussion session that you will guide and facilitate after your presentation. Note that I may adjust the durations for some of the presentation as needs arise.
- You should record your presentation in advance. Your recorded video along with your slides must be submitted via Brightspace at least six days in advance of your scheduled presentation day. Late submissions will be penalized 10% per day.
Project
- Before you undertake your project you will need to submit (via Brightspace) a proposal for approval. The proposal must be one page + an additional page for references. See Deadlines.
- Please consult this page for topics and a list of related papers. The project should be done individually.
- You will present your project in class near the end of term. See Deadlines
- You will submit a report on your final projetc. The report should be:
Assignments
- There are two assignments which will be posted on the course website. They will be a mix of coding (PyTorch) and analytical parts.
- Assignments should be submitted through Brightspace. See Deadlines.
- Late submission of assignments will be penalized 10% per day and will be accepted up to five days past the deadline.