Waterloo Robot Learning Reading Group - Reinforcement Learning for VLAs (Week 10)
Meetup to discuss state-of-the-art research on robot learning (humanoids, foundation models, RL, sim2real, etc.), similar to the Toronto ML/Systems Reading Group and Vector Institute's Machine Learning Lunches- list of topics & articles below - all are welcome! 🎉
Pizza sponsored by Palatial and BracketBot!
Topic - Reinforcement Learning for VLAs
7:00pm - Physical Intelligence., 2025. π*0.6: a VLA That Learns From Experience by Krish Mehta
Additional Reading List:
Kalashnikov, et al. 2018. QT-Opt: Scalable deep reinforcement learning for vision-based robotic manipulation.
Liu, et al., 2025. What Can RL Bring to VLA Generalization? An Empirical Study.
Li, et al., 2025. Reinforcement Learning with Action Chunking.
Zhou, et al., 2025. Efficient Online Reinforcement Learning Fine-Tuning Need Not Retain Offline Data.
Background knowledge:
Questions we should answer as we read these papers:
What are the authors trying to do? Articulate their objectives.
How was it done prior to their work, and what were the limits of current practice?
What is new in their approach, and why do they think it will be successful?
What are the mid-term and final “exams” to check for success? (i.e., How is the method evaluated?)
What are limitations that the author mentions (any that they omit)?