Your search
Differentiable Causal Computations via Delayed Trace
Resource type
Authors/contributors
- Sprunger, David (Author)
- Katsumata, Shin-ya (Author)
Title
Differentiable Causal Computations via Delayed Trace
Abstract
We investigate causal computations taking sequences of inputs to sequences of outputs where the nth output depends on the first n inputs only. We model these in category theory via a construction taking a Cartesian category C to another category St(C) with a novel trace-like operation called “delayed trace”, which misses yanking and dinaturality axioms of the usual trace. The delayed trace operation provides a feedback mechanism in St(C) with an implicit guardedness guarantee.
Date
6/2019
Proceedings Title
2019 34th Annual ACM/IEEE Symposium on Logic in Computer Science (LICS)
Conference Name
2019 34th Annual ACM/IEEE Symposium on Logic in Computer Science (LICS)
Place
Vancouver, BC, Canada
Publisher
IEEE
Pages
1-12
Language
en
DOI
10/ggdf98
ISBN
978-1-72813-608-0
Accessed
2019-11-23T16:57:38Z
Library Catalog
Crossref
Extra
ZSCC: 0000002
Citation
Sprunger, D., & Katsumata, S. (2019). Differentiable Causal Computations via Delayed Trace. In 2019 34th Annual ACM/IEEE Symposium on Logic in Computer Science (LICS) (pp. 1–12). Vancouver, BC, Canada: IEEE. https://doi.org/10/ggdf98
DIFFERENTIAL CALCULUS
MACHINE LEARNING
Link to this record