https://github.com/darioizzo/audi
Header only, C++ library (and the Python pyaudi package) implementing the algebra of Taylor truncated polynomials and a few algorithms useful for its applications (Differential Intelligence, automatic differentiation, Taylor Models, etc.)
Science Score: 23.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
○CITATION.cff file
-
✓codemeta.json file
Found codemeta.json file -
○.zenodo.json file
-
○DOI references
-
✓Academic publication links
Links to: zenodo.org -
○Committers with academic emails
-
○Institutional organization owner
-
○JOSS paper metadata
-
○Scientific vocabulary similarity
Low similarity (14.8%) to scientific vocabulary
Keywords from Contributors
Repository
Header only, C++ library (and the Python pyaudi package) implementing the algebra of Taylor truncated polynomials and a few algorithms useful for its applications (Differential Intelligence, automatic differentiation, Taylor Models, etc.)
Basic Info
- Host: GitHub
- Owner: darioizzo
- License: gpl-2.0
- Language: C++
- Default Branch: master
- Homepage: http://darioizzo.github.io/audi/
- Size: 5.67 MB
Statistics
- Stars: 75
- Watchers: 8
- Forks: 19
- Open Issues: 6
- Releases: 16
Metadata Files
README.md
Audi
Audi (not the car, rather from latin: “listen!”) is an open source, header only, C++ library (exposed to python in the pyaudi package) that implements the algebra of Taylor truncated polynomials and a few algorithms useful for its applications (Differential Intelligence, high-order automatic differentiation, Taylor Models, etc.)
The underlying truncated Taylor polynomial algebra (a differential algebra since integration and derivations are defined too) is dealt with using obake and can deal with high orders and many variables without eating up the whole system memory.
The polynomial multiplication algorithm used in obake (original with the software author Francesco Biscani) takes advantage of sparsity, multiple-threads and cache efficiency allowing a modest memory usage also at high orders.
AuDi was developed with the goal to surpass the capabilities of existing automated differentiation libraries enabling high order differentiation in many variables.
While other automated differentiation codes may be more efficient on some targeted application requesting a specific order and sparsity, AuDi was built to be fast and efficient across all application ranges (low orders, high orders, one variable, many variables, sparse and dense).
Documentation (preliminary) can be found here
Comparison with existing code
Alternative projects that have similar capabilities to AuDi are libtaylor and COSY infinity. Unlike libtaylor AuDi can be used in a dynamic library and can compute at high orders with greater efficiency. Unlike COSY infinity AuDi is entirely open source and has vectorization capabilities.
From the point of view of efficiency, the main difference of AuDi w.r.t. existing codes is in the polinomial multiplication algorithm. AuDi uses the third party obake code and thus gets all the pros and cons of that particular algebraic manipulation system which is still actively developed and was born to deal with massively large polynomial manipulations typically encountered in celestial mechanics perturbation theory. To cut a long story short, AuDi will be "unbeatable" for high orders and many variables (n>=11, m>=11). Below this orders AuDi will still be incredibly memory efficient and fast when used in a machine where multiple threading capabilities are possible.
NEW: Audi also allows for making computations using complex numbers and a vectorized type. This last type allows to compute the derivatives (Taylor polynomial) in multiple points at once, making audi the fastest code of its kind for application such as Machine Learning where this has use. See the folloiwng paper for details on the speed-up w.r.t COSY:
Izzo, Dario, Francesco Biscani, and Alessio Mereta. "Differentiable Genetic Programming." arXiv preprint arXiv:1611.04766 (2016).
pyaudi
Pre-compiled pyaudi binaries are available both from the Pyhton Package Index (PyPi) and from conda-forge. Not all architectures are supported, namely only win64 (PyPi), linux 64 (PyPi and conda) and osx (only conda). The best is to try the following:
conda config --add channels conda-forge
conda install pyaudi
or
pip install pyaudi --user
Owner
- Name: Dario Izzo
- Login: darioizzo
- Kind: user
- Location: Noordwijk
- Company: European Space Agency
- Website: www.esa.int/act
- Repositories: 15
- Profile: https://github.com/darioizzo
GitHub Events
Total
- Issues event: 1
- Watch event: 5
- Push event: 1
- Fork event: 1
Last Year
- Issues event: 1
- Watch event: 5
- Push event: 1
- Fork event: 1
Committers
Last synced: about 1 year ago
Top Committers
| Name | Commits | |
|---|---|---|
| Dario Izzo | d****o@g****m | 558 |
| Francesco Biscani | b****i@g****m | 89 |
| cesans | c****z@g****m | 4 |
| Joerg H. Mueller | n****n@g****m | 2 |
| mikeheddes | m****s@g****m | 1 |
| Dario Izzo | d****o@g****m | 1 |
Issues and Pull Requests
Last synced: 11 months ago
All Time
- Total issues: 36
- Total pull requests: 27
- Average time to close issues: 4 months
- Average time to close pull requests: 17 days
- Total issue authors: 14
- Total pull request authors: 5
- Average comments per issue: 3.03
- Average comments per pull request: 0.52
- Merged pull requests: 23
- Bot issues: 0
- Bot pull requests: 0
Past Year
- Issues: 1
- Pull requests: 0
- Average time to close issues: N/A
- Average time to close pull requests: N/A
- Issue authors: 1
- Pull request authors: 0
- Average comments per issue: 0.0
- Average comments per pull request: 0
- Merged pull requests: 0
- Bot issues: 0
- Bot pull requests: 0
Top Authors
Issue Authors
- darioizzo (9)
- bluescarni (8)
- cesans (2)
- delvillarfr (2)
- Serrof (2)
- neXyon (2)
- yurivict (2)
- miaodi (2)
- Microno95 (2)
- echo66 (1)
- AlOa (1)
- mikeheddes (1)
- CoolRunning (1)
- ianhbell (1)
Pull Request Authors
- darioizzo (16)
- bluescarni (7)
- cesans (2)
- neXyon (1)
- mikeheddes (1)
Top Labels
Issue Labels
Pull Request Labels
Packages
- Total packages: 3
-
Total downloads:
- pypi 248 last-month
-
Total dependent packages: 2
(may contain duplicates) -
Total dependent repositories: 1
(may contain duplicates) - Total versions: 39
- Total maintainers: 2
pypi.org: pyaudi
Implementation of a high-order automated differentiation system using generalized dual numbers. Implementation of a differential algebra.
- Homepage: https://github.com/darioizzo/audi
- Documentation: https://pyaudi.readthedocs.io/
- License: GPLv3+/LGPL3+
-
Latest release: 1.9.1
published almost 6 years ago
Rankings
conda-forge.org: pyaudi
- Homepage: http://darioizzo.github.io/audi/
- License: GPL-2.0-only OR GPL-3.0-only OR LGPL-3.0-only
-
Latest release: 1.9.2
published almost 4 years ago
Rankings
conda-forge.org: audi
- Homepage: http://darioizzo.github.io/audi/
- License: GPL-2.0-only OR GPL-3.0-only OR LGPL-3.0-only
-
Latest release: 1.9.2
published about 4 years ago
Rankings
Dependencies
- actions/checkout v2 composite
- conda-incubator/setup-miniconda v2 composite
- microsoft/setup-msbuild v1.0.2 composite