ConceptAI & ML
KL divergence
A measure of how one probability distribution differs from a second, reference probability distribution.
Know first
Named as a prerequisite by the lectures that define this. The number is how many say so. A grey one has no page yet.
Opens up
Concepts that name this one as a prerequisite.
Bayesian surprise2Evidence lower bound2Fisher information matrix2Natural policy gradient2Constrained optimization of entropy1Covariant/natural policy gradient1Epistemic affordance1Expected free energy1Explicit policy constraint methods1Intrinsic value1KL divergence between Gaussians1KL divergence penalty in RL1
Lecturers
3Each in their own terms, from the easiest level.
Phillip Isola
Sergey Levine
A standard divergence measure on probability distributions that is parameter-independent.Ryan O'Donnell
A measure of the distance between two probability distributions, defined as the expected excess surprise of using one distribution over another.
Lectures that assume it
17The first 12 are shown.
- Exploration and Exploitation
- Local Node Differential Privacy
- Offline Reinforcement Learning II
- Generative Models
- AI Seminar: Actionable Interpretability for Language Model Reasoning
- Machine Text Detectors are Membership Inference Attacks
- Deep Reinforcement Learning
- Reinforcement Learning from Human Feedback: Algorithms & Applications
- Unifying the Cognitive Sciences: The Free Energy Principle
- Active Inference: Bridging Perception and Decision-Making
- The Physics of Motivation: An Active Inference Perspective
- Nonequilibrium Thermodynamics of Uncertain Stochastic Processes
Connections
32The lectures that define it or assume it, and the concepts next to it.
Concepts10
Lectures22
- Variational Inference, Part 2
- Generative Models: Representation Learning Meets Generative Modeling
- Quantum "Supremacy"
- Generative Models: Basics
- Exploration of deep reinforcement learning architectures and algorithms
- Exploration and Exploitation
Show 16 moreShow fewer
- Local Node Differential Privacy
- Offline Reinforcement Learning II
- Generative Models
- AI Seminar: Actionable Interpretability for Language Model Reasoning
- Machine Text Detectors are Membership Inference Attacks
- Deep Reinforcement Learning
- Reinforcement Learning from Human Feedback: Algorithms & Applications
- Unifying the Cognitive Sciences: The Free Energy Principle
- Active Inference: Bridging Perception and Decision-Making
- The Physics of Motivation: An Active Inference Perspective
- Nonequilibrium Thermodynamics of Uncertain Stochastic Processes
- Retrieval Augmented Language Models
- Exploration of advanced topics in deep reinforcement learning
- Deep Inference with Karl J. Friston
- Statistical Thermodynamics as a Paradigm for Data Science
- Exploration of advanced topics in deep reinforcement learning