Correctness of Automatic Differentiation via Diffeologies and Categorical Gluing
Correctness of Automatic Differentiation via Diffeologies and Categorical Gluing
复制标题
DOI:
10.1007/978-3-030-45231-5_17
复制
发表时间:
2020-04-17
期刊:
影响因子:
--
通讯作者:
Vákár M
中科院分区:
文献类型:
--
作者:
Huot M;Staton S;Vákár M
We present semantic correctness proofs of Automatic Differentiation (AD). We consider a forward-mode AD method on a higher order language with algebraic data types, and we characterise it as the unique structure preserving macro given a choice of derivatives for basic operations. We describe a rich semantics for differentiable programming, based on diffeological spaces. We show that it interprets our language, and we phrase what it means for the AD method to be correct with respect to this semantics. We show that our characterisation of AD gives rise to an elegant semantic proof of its correctness based on a gluing construction on diffeological spaces. We explain how this is, in essence, a logical relations argument. Finally, we sketch how the analysis extends to other AD methods by considering a continuation-based method.
登录
查看更多内容
影响因子:
1.8
作者:
Brunel, Alois;Mazza, Damiano;Pagani, Michele
通讯作者:
Pagani, Michele
影响因子:
2.7
作者:
LIU, DC;NOCEDAL, J
通讯作者:
NOCEDAL, J
DOI:
10.1007/978-3-030-45231-5_17
发表时间:
2020-04-17
期刊:
Foundations of Software Science and Computation Structures
影响因子:
--
作者:
Huot M;Staton S;Vákár M
通讯作者:
Vákár M
DOI:
10.1090/s0002-9947-2011-05107-x
发表时间:
2011-11-01
影响因子:
1.3
作者:
Baez, John C.;Hoffnung, Alexander E.
通讯作者:
Hoffnung, Alexander E.
影响因子:
1.1
作者:
Ehrhard, T;Regnier, L
通讯作者:
Regnier, L