Compiler-based Automatic Differentiation (CompAD) Automatic differentiation has been proven extremely useful in the context of numerous applications of computational science and engineering requiring numerical methods that are based on derivative information. Refer to [M1-M4] for details. Integration of AD into an industrial-strength compiler has the potential of making this technology more robust and efficient as well as user-friendly. The automatic generation of adjoint code is of particular interest to scientists and engineers aiming at a transition from the pure numerical simulation to optimization of the underlying models and/or their parameters.
Keywords for this software
References in zbMATH (referenced in 7 articles )
Showing results 1 to 7 of 7.
- Baydin, Atılım Güneş; Pearlmutter, Barak A.; Radul, Alexey Andreyevich; Siskind, Jeffrey Mark: Automatic differentiation in machine learning: a survey (2018)
- Rauser, Florian; Marotzke, Jochem; Korn, Peter: Ensemble-type numerical uncertainty information from single model integrations (2015)
- Stumm, Philipp; Walther, Andrea: New algorithms for optimal online checkpointing (2010)
- Naumann, Uwe: DAG reversal is NP-complete (2009)
- Stumm, Philipp; Walther, Andrea: Multistage approaches for optimal offline checkpointing (2009)
- Gay, David M.: Semiautomatic differentiation for efficient gradient computations (2006)
- Naumann, Uwe; Riehme, Jan: Computing adjoints with the NAGWare Fortran 95 compiler (2006)