Leslie Kaelbling: Reinforcement Learning, Planning, and Robotics
Core Takeaways
Leslie Kaelbling transitioned from philosophy to AI, leveraging symbolic systems to bridge the two fields.
Why it matters
Kaelbling's interdisciplinary approach highlights the importance of diverse academic foundations in AI innovation.
Abstraction and Markov Decision Processes (MDPs) are crucial for managing complexity and uncertainty in AI systems.
▶ 6:00
Why it matters
MDPs and abstraction are essential for developing AI that can effectively operate in uncertain environments.
Hierarchical planning in AI allows for breaking down long-term goals into manageable tasks, aiding robotic reasoning.
▶ 9:00
Why it matters
Hierarchical planning is key to advancing robotic capabilities, enabling more autonomous and efficient operations.
Leslie founded the Journal of Machine Learning Research to address high costs and access issues in academic publishing.
▶ 12:00
Why it matters
This shift towards open access publishing democratizes knowledge and encourages deeper, long-term research.
Engineering intelligent robots requires balancing introspection programming with training large neural networks.
▶ 15:00
Why it matters
This balance is crucial for creating robots that are both adaptable and efficient, impacting future AI development.
Ask this episode Deep
A preview of how Deep chat answers, grounded in this episode with citations and timestamps:
Cite this episode
For papers, blog posts, anywhere.
Related episodes
Where to go next from this conversation.
More on these ideas
Pieter Abbeel: Deep Reinforcement Learning
Shares reinforcement learning, robotics
42m
Gavin Miller: Adobe Research
Shares robotics
1h 9m
Gary Marcus: Toward a Hybrid of Deep Learning and Symbolic AI
More like this
1h 25m
Sam Altman: OpenAI, GPT-5, Sora, Board Saga, Elon Musk, Ilya, Power & AGI
More like this
1h 55m
AI-generated summary · last refreshed 2026-06-08 20:28:30 · how we make these
Quotes are matched verbatim against the source transcript; references are checked to resolve to real URLs. Even so, AI can misread structure or attribute claims imperfectly. If you spot an error, please let us know.