Signatory: differentiable computations of the signature and logsignature transforms, on both CPU and GPU
Signatory is a library for calculating and performing functionality related to the signature and logsignature transforms. The focus is on machine learning, and as such includes features such as CPU parallelism, GPU support, and backpropagation. To our knowledge it is the first GPU-capable library fo...
| Main Authors: | , |
|---|---|
| Format: | Conference item |
| Language: | English |
| Published: |
OpenReview
2021
|
| _version_ | 1826312210058575872 |
|---|---|
| author | Kidger, P Lyons, TJ |
| author_facet | Kidger, P Lyons, TJ |
| author_sort | Kidger, P |
| collection | OXFORD |
| description | Signatory is a library for calculating and performing functionality related to the signature and logsignature transforms. The focus is on machine learning, and as such includes features such as CPU parallelism, GPU support, and backpropagation. To our knowledge it is the first GPU-capable library for these operations. Signatory implements new features not available in previous libraries, such as efficient precomputation strategies. Furthermore, several novel algorithmic improvements are introduced, producing substantial real-world speedups even on the CPU without parallelism. The library operates as a Python wrapper around C++, and is compatible with the PyTorch ecosystem. It may be installed directly via \texttt{pip}. Source code, documentation, examples, benchmarks and tests may be found at \texttt{\url{https://github.com/patrick-kidger/signatory}}. The license is Apache-2.0. |
| first_indexed | 2024-03-07T08:25:39Z |
| format | Conference item |
| id | oxford-uuid:1065ed58-1984-40a8-a8d1-85ff83f74d5a |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-03-07T08:25:39Z |
| publishDate | 2021 |
| publisher | OpenReview |
| record_format | dspace |
| spelling | oxford-uuid:1065ed58-1984-40a8-a8d1-85ff83f74d5a2024-02-16T11:39:52ZSignatory: differentiable computations of the signature and logsignature transforms, on both CPU and GPUConference itemhttp://purl.org/coar/resource_type/c_5794uuid:1065ed58-1984-40a8-a8d1-85ff83f74d5aEnglishSymplectic ElementsOpenReview2021Kidger, PLyons, TJSignatory is a library for calculating and performing functionality related to the signature and logsignature transforms. The focus is on machine learning, and as such includes features such as CPU parallelism, GPU support, and backpropagation. To our knowledge it is the first GPU-capable library for these operations. Signatory implements new features not available in previous libraries, such as efficient precomputation strategies. Furthermore, several novel algorithmic improvements are introduced, producing substantial real-world speedups even on the CPU without parallelism. The library operates as a Python wrapper around C++, and is compatible with the PyTorch ecosystem. It may be installed directly via \texttt{pip}. Source code, documentation, examples, benchmarks and tests may be found at \texttt{\url{https://github.com/patrick-kidger/signatory}}. The license is Apache-2.0. |
| spellingShingle | Kidger, P Lyons, TJ Signatory: differentiable computations of the signature and logsignature transforms, on both CPU and GPU |
| title | Signatory: differentiable computations of the signature and logsignature transforms, on both CPU and GPU |
| title_full | Signatory: differentiable computations of the signature and logsignature transforms, on both CPU and GPU |
| title_fullStr | Signatory: differentiable computations of the signature and logsignature transforms, on both CPU and GPU |
| title_full_unstemmed | Signatory: differentiable computations of the signature and logsignature transforms, on both CPU and GPU |
| title_short | Signatory: differentiable computations of the signature and logsignature transforms, on both CPU and GPU |
| title_sort | signatory differentiable computations of the signature and logsignature transforms on both cpu and gpu |
| work_keys_str_mv | AT kidgerp signatorydifferentiablecomputationsofthesignatureandlogsignaturetransformsonbothcpuandgpu AT lyonstj signatorydifferentiablecomputationsofthesignatureandlogsignaturetransformsonbothcpuandgpu |