Time
Tue/Thu 11-12.20 in CSE 4140
Instructor:
Taylor Berg-Kirkpatrick
tberg@eng.ucsd.edu
Office hours Fri 12-1pm CSE 4106
Teaching assistants and tutors:
Kishore P Venkatswammy Reddy [Office hours Tues 2:30-3:30pm CSE B215, kvenkats@eng.ucsd.edu]
Daniel Spokoyny [Office hours Thurs 2:30-3:30pm CSE B215, dspokoyn@andrew.cmu.edu]
Announcements:
1/9
First discussion section is today in CSE 2154 at 1pm. The TAs will help you setup numpy on your computer and help you get started on a numpy tutorial. Please bring you laptop to discussion. The notebook is available here: numpy tutorial notebook
1/16
Homework 0 has been released. Download it here. This assignment covers some simple probability in the context of nearest neighbor classification and includes a simple programming assignment. Please use this homework to make sure you are up-to-speed on background material. We will grade based on effort. This homework is worth 10 points. The data for the last problem can be found here: hw0train.txt, hw0validate.txt, hw0test.txt. Homework 0 is due before midnight on Wednesday 1/23 and should be turned in on gradescope (gradescope invites to follow later this week).
1/16
As mentioned, in class tomorrow (1/17) we will have our first quiz! It will be graded on completion rather than accuracy and will cover basic probability and calculus. Again, this is a warm-up to help you check background material. If you struggle on the quiz, please come to office hours to discuss how to catch up. Please bring a pen or pencil to class. Paper will be provided.
1/23
Due to issues with setting up Gradescope and Piazza, we are postponing the deadline for HW0 by two days. The new deadline will be Friday January 25 at midnight. Submission instructions will be announced in lecture tomorrow, 1/24 and posted to the course website as an announcment.
1/23
Gradescope is finally up! Link is here. Sign up and join the course via the following course entry code: MVDB7Y. Once you're logged in, you will be able to find the assignment named "Homework 0" and turn in your PDF. The system will not allow late submissions, so make sure you submit by Friday at midnight!
1/24
Piazza is finally up! Please join the class by follow this link.
1/24
Due to scheduling conflict, Taylor's office hours have moved to 12-1pm on Fridays starting this week.
1/29
Piazza is now working for all email addresses! If you don't have an eng.ucsd email address, please use this link with the following access code to sign up: hrtnh. Please post questions about course content to Piazza. Instructors will try to answer in a timely fashion -- students are also welcome to answer!
1/30
Homework 1 has been released. Download it here. This assignment covers some more basic probability, as well as generative classification and multivariate Gaussians. It includes a generative classification programming assignment and is worth 20 points. Homework 1 is due before midnight on Thursday 2/7 and should be turned in on gradescope.
1/31
We are aiming to have Q0, Q1, and HW0 graded and viewable on gradescope by tomorrow morning. (Just a reminder: the drop deadline for all courses is tomorrow, Fri 2/1.) If you're planning on staying in the course, please make sure you've signed up for both Gradescope and Piazza -- we're starting to see more activity on Piazza and will use it's emailing feature for urgent announcements.
2/1
Please check your current quiz and homework grades on Gradescope. If you don't see any grades entered it is because we couldn't find your account. If so, please come talk to us after class on Tuesday.
2/7
Taylor's office hours tomorrow have been moved to 2pm due to faculty meeting.
2/13
Link to the PyTorch tutorial mentioned in discussion section.
2/15
Homework 2 has been released. Download it here. This assignment covers linear regression, convexity, optimization, and requires you to implement multi-class logistic regression. This homework is more intensive that the last one. We encourage you to get started early and ask questions. Discussion this coming week will focus on homework assistance and debugging. It will be worth 25 points and is due before midnight on Friday 2/22 and should be turned in on gradescope.
2/22
Deadline for HW2 has been postponed to Tues, Feb 26, at midnight.
2/24
We are holding an extra office hours Monday (tomorrow) Feb 25 at 2pm to answer HW2 questions. Both the TAs and myself (Taylor) will attend. Room is TBA -- defaulting to my office if all rooms are taken.
2/28
Homework 3 has been released. Download it here. This assignment covers perceptron, SVM, and neural nets, and requires you to implement a neural digit classifier. Again, this homework is more intensive than the last one. We encourage you to get started early and ask questions. Discussion this coming week will focus on homework assistance and debugging. It will be worth 30 points and will be due before midnight on Tuesday 3/12 and should be turned in on gradescope.
2/28
Homework 4 has been released. Download it here. This assignment is open-ended and gives you the choice of exploring supervised or semi-supervised digit classification. We encourage you to get started early and ask questions. This final discussion section this week will focus on assistance with PyTorch to help you build on your HW3 system to complete the final homework. It will be worth 30 points and will be due before midnight on Friday 3/22 and should be turned in on gradescope.
Syllabus, dates, other administrative stuff
Schedule of lectures and homeworks