Discretization
One word, taught differently in 3 fields. Each use is given apart.
In AI & ML
Casting model weights to a lower-bit precision format like 8-bit to reduce memory usage while preserving performance.
In Engineering
Converting a continuous-time linear dynamical system into a discrete-time representation using sampling intervals.
In Mathematics
The process of transferring continuous models and equations into discrete counterparts by creating a grid of points.
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.
Continuous-time linear dynamical systems1Floating point precision1Matrix exponential1Model weights1Poisson equation1
Opens up
Concepts that name this one as a prerequisite.
In AI & ML
1In Engineering
1In Mathematics
1Lectures that assume it
8- Chebfun: Numerical Computing with Functions via Chebyshev Polynomials (SIAM 2013)
- Yang-Mills and the Mass Gap
- Dynamic Programming I
- Inputs, Outputs, and Symmetric Matrices in Linear Dynamical Systems
- Controllability and State Transfer in Modern Electrical Engineering
- Imitation Learning III: Inverse RL & Generative Adversarial Imitation Learning (GAIL)
- Operator Learning in Theory and Practice
- Guest Lecture: Dorsa Sadigh (UC Berkeley CS 285)
Connections
18The lectures that define it or assume it, and the concepts next to it.
Concepts7
- Matrix exponential
- Poisson equation
- Autoregressive discretization
- Curse of dimensionality
- Discretization invariance
- GPTQ
Show 1 moreShow fewer
Lectures11
- Solutions via Laplace Transform and Matrix Exponentials
- A Hacker's Guide to Language Models
- Multigrid Methods for Laplace's Equation
- Chebfun: Numerical Computing with Functions via Chebyshev Polynomials (SIAM 2013)
- Yang-Mills and the Mass Gap
- Dynamic Programming I
Show 5 moreShow fewer
- Inputs, Outputs, and Symmetric Matrices in Linear Dynamical Systems
- Controllability and State Transfer in Modern Electrical Engineering
- Imitation Learning III: Inverse RL & Generative Adversarial Imitation Learning (GAIL)
- Operator Learning in Theory and Practice
- Guest Lecture: Dorsa Sadigh (UC Berkeley CS 285)