Common · Knowledge
Differentiable programming
Programming paradigm in which a numeric computer program can be differentiated throughout via automatic differentiation, allowing for machine learning based on gradient descent etc.
Differentiable programming is a programming paradigm in which a numeric computer program can be differentiated throughout via automatic differentiation. This allows for gradient-based optimization of parameters in the program, often via gradient descent, as well as other learning approaches that are based on higher-order derivative information.
From Wikipedia
Differentiable programming is a programming paradigm in which a numeric computer program can be differentiated throughout via automatic differentiation. This allows for gradient-based optimization of parameters in the program, often via gradient descent, as well as other learning approaches that are based on higher-order derivative information. Differentiable programming has found use in a wide variety of areas, particularly scientific computing and machine learning. One of the early proposals to adopt such a framework in a systematic fashion to improve upon learning algorithms was made by the Advanced Concepts Team at the European Space Agency in early 2016.
Text: Wikipédia, CC BY-SA 4.0. ·
Related cards
-
★★
Parallel computing
Programming paradigm in which many calculations or the execution of processes are carried out simultaneously
-
★★★
Gradient
Multi-variable generalization of the derivative
-
V★★
Variety (cybernetics)
Number of states of a cybernetic system
-
★★
Conjugate gradient method
Method to compute systems of linear equations whose matrix is symmetric positive-definite
-
A★★
Aspect-oriented programming
Programming paradigm
-
P★
Preimage theorem
Mathematical theorem (differential topology)
-
★★
Logic programming
Programming paradigm based on formal logic
-
★★★
High-level programming language
Programming language with strong abstraction from details of hardware
-
★★★
Convolutional neural network
Regularized type of feed-forward neural network that learns features by itself via filter (or kernel) optimization
-
F★
Functional derivative
Concept in calculus of variation
-
★★★
Gradient descent
Optimization algorithm
-
N★
Numerical stability
Property of a numerical algorithm which describes how errors and approximations in the input data or intermediate values propagate through the algorithm without diverging its results
-
★
Computational engineering
Field of algorithmic training
-
★
Pushforward (differential)
Linear approximation of smooth maps on tangent spaces
-
D★
Data-flow analysis
Technique for gathering information about the possible set of values calculated at various points in a computer program
-
P★★
Predictive coding
Psychological term
-
L★
Learning rate
Tuning parameter (hyperparameter) in optimization
-
G★★
Generic programming
Way of designing and writing programs where algorithms are written in terms of parametric types enabling easy reuse