Metadata-Version: 2.1
Name: camd
Version: 2021.6.11.post7
Summary: camd is software designed to support autonomous materials research and sequential learning
Home-page: https://github.com/TRI-AMDD/CAMD
Author: AMDD - Toyota Research Institute
Author-email: murat.aykol@tri.global
Maintainer: Murat Aykol, Joseph Montoya
Maintainer-email: murat.aykol@tri.global
License: Apache
Keywords: materials,battery,chemistry,science,density functional theory,energy,AI,artificial intelligence,sequential learning,active learning
Platform: UNKNOWN
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: Apache Software License
Classifier: Operating System :: OS Independent
Description-Content-Type: text/markdown
Requires-Dist: python-dateutil (==2.8.1)
Requires-Dist: networkx (==2.5.1)
Requires-Dist: matplotlib (==3.4.1)
Requires-Dist: pandas (==1.2.4)
Requires-Dist: matminer (==0.6.5)
Requires-Dist: autologging
Requires-Dist: awscli (==1.19.44)
Requires-Dist: boto3 (==1.17.44)
Requires-Dist: docopt (==0.6.2)
Requires-Dist: scikit-learn (==0.24.1)
Requires-Dist: taburu (==2020.5.9)
Requires-Dist: GPy (==1.9.9)
Requires-Dist: watchtower (==1.0.6)
Requires-Dist: qmpy-tri (==2021.6.11)
Provides-Extra: proto_dft
Requires-Dist: protosearch (==2020.5.10) ; extra == 'proto_dft'
Provides-Extra: tests
Requires-Dist: pytest ; extra == 'tests'
Requires-Dist: pytest-cov ; extra == 'tests'
Requires-Dist: coveralls ; extra == 'tests'


camd is software designed to support Computational Autonomy for Materials Discovery
based on ongoing work led by the
[Toyota Research Institute](http://www.tri.global/accelerated-materials-design-and-discovery/).

camd enables the construction of sequential learning pipelines using a set of
abstractions that include
* Agents - decision making entities which select experiments to run from pre-determined
    candidate sets
* Experiments - experimental procedures which augment candidate data in a way that
    facilitates further experiment selection
* Analyzers - Post-processing procedures which frame experimental results in the context
    of candidate or seed datasets

In addition to these abstractions, camd provides a loop construct which executes
the sequence of hypothesize-experiment-analyze by the Agent, Experiment, and Analyzer,
respectively.  Simulations of agent performance can also be conducted using
after the fact sampling of known data.


