Robot Learning Micro-Hackathon / Research Sprint
Going to do this with New Systems. But going to start with a Robot Learning Reading Group.
Similar initiatives
Ideas:
- Robot Learning Reading Group
- Call for collaborators
- 1-day hackathons every weekend to learn about a topic
Vision
This recurring event will serve as a central place for technical robot learning people to gather and collectively think about how to advance the robot learning field in Toronto. A place for researchers and research-interested people to meet.
Target audience: robot learning researchers, people interested in robot learning and want to genuinely get deep in the topic and publish research. Opportunity to find collaborators and share interesting research.
Really I want to start this as an unconference.
Topics:
- Robot Foundation Models
- Reinforcement Learning
- RL from scratch
- Offline-to-online finetuning
- goal-conditioned RL
- Sim2Real
- Scaling laws for robot learning
- World Model
- Learning from Human Videos
- Real2Sim
- In-Context Learning
Understanding Researcher Incentives.
Prototype
A weekly lunch talk series on topics in robot learning, hosted at the New Systems.
Structure 1: Non-hackathon, purely cowork vibe-based for robot learning
- 10:30am-11:15am - Robot Foundation Models intro workshop
- 11:15am-11:30am - Call for collaborators
- 11:30am - 12:45pm - Work session 1
- 12:45pm-1:30pm - Free Lunch / networking
- 1pm-3:00pm - Coworking session 2
- 3:00pm - 3:30pm — Wrap-up + Show-and-Tell
- I don’t like this because each person just works on their own thing, I might as well stay home to work on things
During work session:
- You can read papers
- Or work on a robot learning project (continue working on your research). Also good chance to find people
Structure 2: Mini-Hackathon based
- 10:00–10:30am — Welcome and schedule for the day
- 11:00–11:15 — Team finding
- 10:30–11:30am —
<INSERT TOPIC>workshop- 11:15 - 7pm - Hacking on topic
- 7pm - 8pm - presentation and wrap up
- Goal is to work on things that push the limits of what is happening.
Structure 3: Have fun / learn in morning, work in afternoon
- 10:30am-11:15am - Robot Foundation Models intro workshop
- 11:15am-11:30am - Call for collaborators
- 12:45pm-1:30pm - Free Lunch / networking
- 1:30pm - 5:00pm - Work session with cracked people (Split into deep work / shallow work)
Structure 4: Unconference
- 10:30am-11:15am - Session 1
- 11:15am-12:30pm - Session 2
- 12:30-1:00pm - Free Lunch / networking
- 1:00pm - 5:00pm - Work session with cracked people (Split into deep work / shallow work)