Unfortunately, due to the large number of similar e-mails that I receive I can no longer respond to each e-mail individually. I've provided information below on the most common requests:
Supervising PhD or MSc
Thank you for your interest in studying under my supervision. I may be looking to take 1 or 2 new students next year, but these spots will be very competitive as machine learning and optimization are popular fields (last year we had over 700 applications and I only accepted 3 students, so acceptance rate is less than 1%). On the other hand, we are planning to hire new machine learning faculty so the acceptance rate may be higher next year.
But in any case, faculty members in computer science do not directly admit students. If you are interested then you need to apply through the department of computer science, which usually has a deadline in December.
If you are contacting me because of an interest in convex optimization, you may also want to consider applying to UBC's Okanagan campus. They have a strong group there, and this requires a separate application.
Updates on Grad School Applications
Due to the large number of applicants, I can't provide updates on the status of individual grad student applications. The department sends out all decisions by April 15th. Similarly, I don't have time to do Skype calls with each applicant to discuss potential research topics.
Unfortunately, the application fee is imposed by the university on us and we don't have control over it.
CPSC 340 or 532M Registration
Please see the webpage for registration details.
CPSC 340 Prerequisite Exceptions
I do not control prereq checking for undergrad courses.
CPSC 540 Registration
I put the registration information for CPSC 540 here.
Also please note that CPSC 540 is a followup to CPSC 340 (which is listed as CPSC 532M for graduate students), and for 99% of students it makes more sense to take CPSC 340. Also, please note that in the all sections I've taught that we were able to accommodate every student on the CPSC 340 waiting list.
Internship Requests (Non-UBC Students)
Due to the high interest from UBC students, I'm not planning on taking any interns that aren't UBC students in the near future. You may want to consider the Mitacs Globalink program, which matches prospective interns with supervisors that are looking for students.
Internship Requests (UBC Students)
I will not have any internship positions available until summer 2020 at the earliest. There positions tend to be competitive, and I prefer taking people that have completed both CPSC 340 and CPSC 540. Contact me in December 2019 if you are interested.
Thanks for your interest, but due to the large number of applications I'm currently only able to accept postdocs that have already published in the places that I normally publish like ICML and NeurIPS.
Due to the high demand for machine learning reviewers/editors/PCs and the availability of excellent not-for-profit publication venues, I will not accepting any reviewing/editing/PC duties for for-profit journals/conferences unless they are publication venues that my students want to submit to.
Companies Seeking Interns
If you are looking for student interns, you can send me a description of what you are looking for and I can forward it to my students.
(My students tend to get more offers than they can accept, and usually ask for competitive salaries.)
I can also forward it to the larger machine learning group at UBC if you like (these students have a higher variance in experience, but you are more likely to find someone going this route).
You may also have better luck contacting Mike Gelbart, who co-directs the Master of Data Science program where all students must complete a capstone project.
Usually, these are the best people first people to contact regarding setting up industry collaborations:
Raymond Ng (director of the Data Science Institute)
Kevin Leyton-Brown (director of the Centre for Artificial Intelligence Decision-making and Action)
Collaborations with me are most useful if you have already (i) collected a dataset and (ii) tried out existing "off the shelf" methods.