Skip to content

meeting_observation_log.md

Latest commit

 

History

History
97 lines (80 loc) · 5.85 KB

meeting_observation_log.md

File metadata and controls

97 lines (80 loc) · 5.85 KB
Raw

Unit Meeting Observation Log

Project: Unit Meeting Procedures & Assessment
Team: Kefei Yan, Vanessa Zhou, Sean Khan
Purpose: Document what actually happens in unit meetings.

How to Use This Document

  1. Attend a unit meeting → copy the Observation Template below and fill it in

Key questions to keep in mind while observing:

  • Does the meeting have a clear structure, or is it unorganized?
  • Who drives the conversation — the facilitator, the advisor, or students?
  • Do researchers from different projects interact with each other?
  • What makes well-run meetings different from poorly-run ones?
  • What would a researcher need to prepare meaningfully for a meeting?

Observation Template

Copy this for each meeting attended:

### [Unit Name] — [Date]
- **Host/Facilitator:**
- **Attendees (approx):**
- **Duration:**
- **Meeting Structure:** (Was there an agenda? Slides? Clear flow?)
- **Discussion Quality (1-5):** (1 = silent, 5 = very active)
- **Cross-Project Discussion:** (Yes/No — if yes, describe)
- **Researcher Contributions:** (Demos, articles shared, questions raised?)
- **Advisor/Faculty Role:** (Present? How did they participate?)
- **Key Topics Discussed:**
- **Problems Raised:**
- **What Worked:**
- **What Didn't Work:**

Meeting Observations

Image Processing Unit — 2026-02-27

  • Host/Facilitator: Dr. Mussmann
  • Attendees (approx): 9
  • Duration: ~1 hr 19 min (originally scheduled for 1 hr; last ~20 min was unit manager + faculty discussing next steps)
  • Meeting Structure: Yes — two clear halves. First half: Dr. Mussmann presented a topic (Weak Supervision). Second half: one project group shared their update. Post-meeting: unit manager stayed with faculty to discuss logistics and next steps.
  • Discussion Quality (1-5): 4
  • Cross-Project Discussion: Yes — the topic presentation (Weak Supervision) was broadly relevant across projects. Covered incomplete supervision, noisy supervision, and semi-supervised learning with examples from medical imaging, remote sensing, NLP, and more. Multiple projects could relate their own work to the concepts discussed.
  • Researcher Contributions: One project group presented their update in the second half
  • Advisor/Faculty Role: Dr. Mussmann led the first half with a structured 19-slide presentation. Stayed after to discuss next steps with the unit manager.
  • Key Topics Discussed: Weak supervision methods including aggregated labels, censored labels, labeling rules, external systems, auxiliary signals, crowdsourced labels, self-training, representation learning, consistency-based methods, and co-training/multi-view learning.
  • Problems Raised: Not documented
  • What Worked: Faculty-led topic presentation gave the meeting real substance and educational value. The topic was chosen to be relevant across multiple projects. Having a clear two-part structure (topic + project update) kept the meeting organized. High attendance (9 people) compared to other units.
  • What Didn't Work: Meeting ran 19 minutes over scheduled time. The last 20 minutes was only unit manager + faculty, not the full group — unclear if this should be a separate meeting.

User-Interface Unit — 2026-02-26

  • Host/Facilitator: Unknown / no clear facilitator
  • Attendees (approx): 2
  • Duration: ~11 minutes
  • Meeting Structure: No agenda, no slides, no structured flow
  • Discussion Quality (1-5): 2
  • Cross-Project Discussion: No — too few attendees for any cross-project exchange
  • Researcher Contributions: None
  • Advisor/Faculty Role: Not present
  • Key Topics Discussed: None — meeting lasted only ~11 minutes
  • Problems Raised: None
  • What Worked: Nothing notable
  • What Didn't Work: Extremely low attendance (only 2 people). No agenda or facilitator to guide discussion. Meeting ended after 11 minutes with no meaningful content exchanged. Without a critical mass of attendees or a structured plan, the meeting had no purpose.

3-D Vision Unit — 2026-02-25

  • Host/Facilitator: None
  • Attendees (approx): 1 (only the person who opened the meeting)
  • Duration: ~15 minutes
  • Meeting Structure: No agenda, no slides
  • Discussion Quality (1-5): 1
  • Cross-Project Discussion: N/A — no other attendees
  • Researcher Contributions: None
  • Advisor/Faculty Role: Not present
  • Key Topics Discussed: None — no participants showed up
  • Problems Raised: None
  • What Worked: Nothing notable
  • What Didn't Work: Only 1 person joined. No other researchers or advisors attended. The person waited ~15 minutes and then ended the meeting themselves. This raises serious questions about whether researchers see value in unit meetings, whether reminders are being sent, and whether there is any accountability for attendance.

Notes & Emerging Patterns

  • Facilitator presence is the single biggest factor. Dr. Mussmann's Image Processing meeting had 9 attendees and ran 1hr 19min with active discussion. UI and 3-D Vision meetings without a faculty facilitator had 2 and 1 attendees respectively, lasting 11-15 minutes.
  • Structure matters. The Image Processing meeting had a clear two-part format (topic presentation + project update) that kept things moving. The other two meetings had no agenda at all.
  • Faculty-led topic presentations create cross-project value. Dr. Mussmann's Weak Supervision talk was relevant to multiple projects, giving researchers a shared reference point. This is exactly the kind of knowledge transfer Bree intended.
  • Meetings without a facilitator or agenda collapse almost immediately (11 min and 15 min)
  • No evidence of any pre-meeting reminders or accountability for attendance in the cold meetings
  • Open question: should the unit manager + faculty post-meeting discussion (last 20 min of Image Processing) be formalized as a separate recurring check-in?