Skip to content

Chapter 5. Forward Mode Automatic Differentiation

Forward mode automatic differentiation computes derivatives by carrying two values through a program at the same time: the ordinary value and its tangent. The ordinary value...

SectionTitle
1Chapter 5. Forward Mode Automatic Differentiation
2Dual Numbers
3Forward Evaluation Rules
4Jacobian-Vector Products
5Complexity Analysis
6Higher-Dimensional Tangent Spaces
7Efficient Seeding Strategies
8Sparse Forward Methods
9Case Studies
Chapter 5. Forward Mode Automatic DifferentiationForward mode automatic differentiation computes derivatives by carrying two values through a program at the same time: the ordinary value and its tangent. The ordinary value...
6 min
Dual NumbersDual numbers give forward mode automatic differentiation a compact algebraic form. Instead of storing a value and a tangent as two unrelated fields, we package them into one...
7 min
Forward Evaluation RulesForward mode automatic differentiation works by replacing each primitive operation with an extended operation on pairs:
7 min
Jacobian-Vector ProductsThe natural output of forward mode automatic differentiation is a Jacobian-vector product. Instead of constructing the full Jacobian matrix explicitly, forward mode computes...
7 min
Complexity AnalysisForward mode automatic differentiation has a simple cost model. It evaluates the original program and, at the same time, evaluates the tangent program. Each primitive...
8 min
Higher-Dimensional Tangent SpacesSo far, forward mode has propagated a single tangent direction:
7 min
Efficient Seeding StrategiesForward mode automatic differentiation computes Jacobian-vector products:
7 min
Sparse Forward MethodsMany real-world Jacobians are sparse. Most derivative entries are zero because outputs depend only on small subsets of inputs.
7 min
Case StudiesForward mode automatic differentiation appears in many numerical systems where directional derivatives, local sensitivities, or small parameter sets are important. This...
6 min