Todo Summary
TODO
2026 - February
-
02/02/2026 10:20 AM: Ask jason if undergrads will come to lab meetings.
-
02/02/2026 10:20 AM: Has done SURE before. Criteria on website saying SURE is for exploring research.
-
04/02/2026 9:14 AM: 504 Homework
-
04/02/2026 9:14 AM: Implement diverse sampling
-
04/02/2026 9:14 AM: Read this article Jason sent https://ai.engin.umich.edu/2023/08/17/eight-lessons-learned-in-two-years-of-ph-d/
-
16/02/2026 10:45 AM: Also should I change the system that selects which clusters to use to select those with the most samples in them? Or sample in proportion to the number of samples in each? Or is it good to force diversity by sampling randomly overly the clusters?
-
18/02/2026 10:11 AM - One-on-one: Read papers on sequential testing. It relates to the work I am doing now.
-
19/02/2026 10:28 AM - SURE Applicants: Ask jason if undergrads will come to lab meetings.
-
19/02/2026 10:28 AM - SURE Applicants: Has done SURE before. Criteria on website saying SURE is for exploring research.
-
25/02/2026 1:23 PM - One-on-one: How do we scale the dataset creation? Founder of duolingo has papers with insights into this.
-
25/02/2026 1:23 PM - One-on-one: Respond to undergrad email about dataset creation.
2026 - January
- 15/01/2026 11:43 AM: Do in person seminar things
2026 - March
- 20/03/2026 12:09 PM: Send this email and cc
ECE 567
-
17/02/2026 12:00 PM - Lecture: Look into distributional RL for research.
-
24/02/2026 12:02 PM - Lecture: Find a paper for this.
EECS 504
-
11/03/2026 4:33 PM - Lecture: Study ransac for exam. Specifically the table on the slide.
-
14/01/2026 4:33 PM - Lecture: Do the survey
-
14/01/2026 4:33 PM - Lecture: Do the vision in the wild thing
-
21/01/2026 4:28 PM - Lecture: Consider the fact that removing the set of answers from a constrained VQA could make the question ambiguous.
-
21/01/2026 4:28 PM - Lecture: Think further about how to collect VQA questions from humans given the fact that humans generally default to easy questions.
-
25/02/2026 4:53 PM - Lecture: Email Bernadette about these. Or slack her I don't know.
Lab Meetings
-
08/01/2026 2:13 PM: Look at NEURIPS papers for multi-turn clarification dialog.
-
19/02/2026 1:50 PM: Join departmental calendars for CSE and EECS
Reading List
-
Multi-Turn Multi-Modal Question Clarification for Enhanced Conversational Understanding: Check if the dataset from this paper is useful
-
Multi-Turn Multi-Modal Question Clarification for Enhanced Conversational Understanding: Read this paper
-
TPO: Aligning Large Language Models with Multi-branch & Multi-step Preference Tr: Read this paper
RGSC
-
Bringup TODO: Check if this is still in use. If not note that in a readme. If so update.
-
Bringup TODO: Create a readme that outlines tasks for secratary, when to complete them, how to compete them, what resources are needed, and what files are relevant/should be created.
-
Bringup TODO: Develop new protocol for keeping active members list up to date. Sheets script? Something simpler?
-
Bringup TODO: Talk to Connor and Jessi about changing how the TVs in the grad lounge work. Pull from RGSC calendar? QR code that lets people add their own announcements (subject to approval)? Perhaps integrate the systems so that anybody can fill in a form to get stuff put on the calendar (subject to approval) and that gets pulled to the TVs? Maybe you fill out a google form that leads to a sheet that we can approve on and when that happens it goes out as emails and onto the TVs? Always have up a way to subscribe to the calendar and a way to subscribe to email notification for specific categories of event.
-
Bringup TODO: Update officer list archive for 2026
RGSC Minutes
- 20/01/2026 5:29 PM: Suggestion box?