1:1 Meekail + Shannon

1:1 Meekail + Shannon === ###### tags: `Templates` `Meeting` :::info - **Location:** Zoom - **Date:** April 26, 2022 2:00 PM (UTC-4) - **Agenda** 1. Review current semester progress `10min` - 5 SciPy submissions (4 cilia, 1 OrNet) - 2 Masters theses (Nathan completed, Curtis in progress) - Improved SE aspect of codebase - Developed potential basis for dissertation 2. Review state of Cilia grant overall `10min` 3. Decide template for Mentor-Mentee Compact (MMC)`5min` - [University of Alabama-Birmingham](https://grad.uga.edu/wp-content/uploads/2021/02/University-of-Alabama-Birmingham.pdf) - [University of Minnesota](https://grad.uga.edu/wp-content/uploads/2018/01/mentor-mentee_compact_UMN.pdf) - [University of Wisconsin-Madison](http://grad.uga.edu/wp-content/uploads/2019/01/MentorshipAgreementTemplate.pdf) 4. Fill out MMC `15min` 5. Discuss future direction (summer/fall) - Discuss long-term (e.g. yearly) goals/markers ::: ## Current semester deliverables - [Github Repo](https://github.com/quinngroup/metric-vae) :books: Active Papers/Manuscripts --- - Evolving Mixtures Prior (Bella, Sara) - Metric Learning in VAEs (Nathan) - Active Learning and Categorical Encoding (Curtis) - Low-Level Feature Extraction for Cilia (Meekail, Eric) - Nathan's Thesis - Curtis's Thesis - Benchmarks for OrNet (Rachel, Victoria, Delaney) :mag: Spring Retro --- ### What went right - New paper-reading format where we tackle papers one caliber above what we would normally feel comfortable reading, then break it down as a group - Wildly positive reviews from team - Helped provide perspective and technical practice - Flawed in that 9/10 times the team wasn't able to discuss and answer questions themselves, ended up relying on me to answer and derive equations - Stricter SE principles in github repo - Encouraged unit tests (helped a lot w/ design) - Mandated PR workflow (best decision) - Focused more on design principles (e.g. inheritence vs composition) - Focused on *guiding* people to the right answers and solutions to technical problems, rather than outright *solving* them for them (mainly for Nathan/Curtis) - Guided the masters students to develop common API for their projects so that they could benefit from each others code contributions - Provided rigorous review for most aspects of projects: timelines, abstracts, manuscripts, code, etc. ### What went wrong - Found myself stretched a bit thin, especially in the beginning of the semester - Attention split between two masters students and their elevated needs - Worked on allowing each student a unique project with a sense of ownership, but that also meant managing the technical and theoretical aspects of 3 projects - Worked on helping OrNet establish...any infrastructure - Worked on helping OrNet establish research direction (felt vindicated when our bio collaborators mentioned mitochondrial structure is edge-based, not blob-based) - Worked on training(?) the postdoc - Miscommunications and unalligned expectations with the postdoc burned **a lot** of energy - Scheduling was difficult due to part-time job and leading: - Group project meeting (a lot of cross-module coordination was needed) - Group paper reading meeting - Individual module meetings (Nathan, Curtis, Bella/Sara) - Individual meetings with Sara - Group meetings for OrNet (Dr. Ojeme should have been running instead) - Group paper reading meetings (Dr. Ojeme should have been running instead) ### What should/will change - Team composition: - Bella, Sara, Rachel, Victoria, Delaney - Meeting logistics - I will be in Kennesaw so I can only commute ~1 day/week :closed_book: Tasks --- ==Importance== (1 - 5) ### Development Team: - [ ] ==5== Email invite - [ ] ==4== Email registration page - [ ] ==5== Email invitees - [ ] ==4== Setup e2e test in production ### Design Team: - [ ] ==4== Interview users - [ ] ==5== Build roll-up display content - [ ] ==5== Help user discover new features :dart: Goals for Fall --- - Need to find potential venues (submissions ideally due in Spring) - [IJCAI](https://ijcai-22.org/calls-papers/): paper submission due ~Jan. - [ICML](https://icml.cc/Conferences/2022): paper submission due ~Jan. - [COLT](http://www.learningtheory.org/colt2021/index.html#important-dates): paper submission due ~Jan. - [CVPR](https://cvpr2022.thecvf.com/submission-timeline): paper submission due ~Nov. - [ICLR](https://iclr.cc/Conferences/2022/Dates): paper submission due ~Oct. - Potential Papers - EMP with Riemannian metric - **EMP with human-in-the-loop training** - EMP with Categorical encoding - **Unitary-regularization** ## Notes  - Status of Web-Dev? - SciPy poster deadline? - [ ] Shannon - find full cilia data and put on Roci - spec out some way of interactively labeling data - Have Alireza build a napari interface?