Logistics/Notes
Logistics/Notes#
Monday 10/3
Lecture: Neural networks
HW5 posted, should be straightforward (similar to HW4, but new methods)
Project poster and notebook due next week
In-class practice today, gradescope open!
If you are installing custom conda packages (e.g. graph networks for TSP), let me know and I can up your disk quota to make some extra room
Monday 3/26
Lecture: Train/validation splits, and non-parametric models
HW4 posted; should be able to get started after today’s lecture
Project feedback posted on gradescope, follow up if you have questions!
Monday 9/19
Lecture: Overfitting, model capacity, simple feature selection
Feedback from ML guest lectures was excellent. Quick discussion about class schedule.
Feedback from mid-term evals
Only ~10/50 responses
Quite bimodal (either people thought things are going really well, or it was too fast)
Common feedback:
Lectures move fast (please tell me when it’s moving too fast or you want me to slow down!)
Homeworks can take a long time
Lots of code/functions/definitions/etc to understand
HW3 posted (should be fairly straightforward based on lecture from Wednesday, please ask question on piazza and come to office hours!)
Project proposal due noon 9/23 on gradescope
Office hours as normal this week here: https://ulissigroup.cheme.cmu.edu/F22-06-325/syllabus.html#personnel
jupyterhub: I haven’t seen any questions/etc in the past two weeks, so I think everything has stabilized and is working well!
Wednesday 9/7
Lecture: Project overview, and intro to supervised regression
-
Email link sent to CATME to help with group/project assignment, please fill out by Friday!
Groups will be assigned over the weekend
You will have ~9/23 to investigate the project info, dataset, and write a short proposal on what you want to try. Feel free to ask for suggestions or clarifications in office hours!
Please upload the practice to gradescope at the end of class!
HW2 is posted, due Monday
Office hours, updated on Syllabus
Ulissi office hours Thurs 530-630 on zoom
Optional python lab Friday 330-430 DH 1102
jupyterhub: Updated save tips at Software Environments
jupyterhub: Save-to-pdf with chemical equations fixed (hopefully that works cleanly now!) https://github.com/ulissigroup/F22-06-325/blob/main/Dockerfile#L66
jupyterhub: There appears to be a problem with JupyterLab RTC that might be causing the problems with saving files (https://github.com/jupyterlab/jupyterlab/issues/12154). It’s disabled now, so please let me know if there’s still any problems