ConceptAI & ML
Function approximation
Using parametric functions to generalize value or policy estimates across large or continuous state and action spaces.
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.
Approximate policy evaluation for action-value functions1Approximation error1Coarse coding1Contraction in Q-learning1Convergence of control algorithms1Deep learning1Differentiable function approximators1Discretization invariance1Doubly robust estimator1Epistemic uncertainty1Kernel methods1Monte Carlo evaluation with function approximation1
Lecturers
2Each in their own terms, from the easiest level.
Brian Yu
Using a function to estimate values based on features rather than storing a discrete value for every state-action pair.
Lectures that assume it
15The first 12 are shown.
- Value Function Approximation & Deep Q-Networks (DQN)
- Natural Policy Gradients, Covariant Updates & The Fisher Information Matrix
- Computer-Assisted Proofs in Fluid PDEs using Interval Arithmetic
- Accuracy of Mathematical Functions in Julia
- Dynamic Programming III
- Deep Reinforcement Learning
- Approximation Theory
- Representation Learning: Theory
- Generative Control, Action Chunking, and Moravec’s Paradox
- Offline Reinforcement Learning
- Challenges and Open Problems in Deep Reinforcement Learning (CS 285, Lecture 23, Part 1)
- Questions for Theory in the New Age of Machine Learning
Connections
27The lectures that define it or assume it, and the concepts next to it.
Concepts10
Lectures17
- Harvard CS50’s Artificial Intelligence with Python – Full University Course
- Towards a Statistical Foundation for Reinforcement Learning
- Value Function Approximation & Deep Q-Networks (DQN)
- Natural Policy Gradients, Covariant Updates & The Fisher Information Matrix
- Computer-Assisted Proofs in Fluid PDEs using Interval Arithmetic
- Accuracy of Mathematical Functions in Julia
Show 11 moreShow fewer
- Dynamic Programming III
- Deep Reinforcement Learning
- Approximation Theory
- Representation Learning: Theory
- Generative Control, Action Chunking, and Moravec’s Paradox
- Offline Reinforcement Learning
- Challenges and Open Problems in Deep Reinforcement Learning (CS 285, Lecture 23, Part 1)
- Questions for Theory in the New Age of Machine Learning
- Operator Learning in Theory and Practice
- From Large Language Models to General Artificial Intelligence?
- RL Theory