vivid-node 0.5.583

Enterprise Machine-Learning and Predictive Analytics

Vivid Node

Vivid Node (alias Rian) is a free Python library for next generation machine learning- and data analysis applications, that implement cloud-assisted meta programming. Rian is part of the Vivid Code framework and actively developed at the Frootlab Organization.

The key goal of Rian is to automate and support collaborative data science. To achieve this goal Rian orchestrates established frameworks like TensorFlow® or Keras® and dynamically extends their capabilities by cloud based community algorithms. Thereby Rian allows the usage of abstract defined algorithms, that are specified with respect to their category, the used data type and an evaluation metric. During runtime these abstract specifications are resolved cloud-sided by currently best fitting algorithm, that match the specification. This allows the separation of engineering and data science, as well as simple collaborations between organizations without the requirement to share data.

Collaborative data science using the Vivid Code framework

For a given algorithm category and data type, the currently best fitting algorithms are determined by the used metric. Examples for such metrices would be the prediction accuracies within a fixed set of gold standard samples of the respective domain of application (e.g. latin handwriting, spoken words, TCGA gene expression data, etc.).

Current Development Status

Rian currently is in Pre-Alpha development stage, which immediately follows the Planning stage. This means, that at least some essential requirements of Rian are not yet implemented.

Installation

Comprehensive information and installation support is provided within the online manual. If you already have a Python environment configured on your computer, you can install the latest distributed version by using pip:

$ pip install rian

Or alternatively:

$ pip install vivid-node

Documentation

The documentation of the latest distributed version is available as an online manual and for download, given in the formats PDF, EPUB and HTML.

Contribution

Contributors are very welcome! Feel free to report bugs, ideas and feature requests to the issue tracker, provided by GitHub. Currently, as our team still is growing, we do not provide any Contribution Guide Lines. So, if you are interested to help or to join the team, we would be glad, to hear about you.

License

Rian is open source software and available free for any use under the GPLv3 license:

© 2019 The Frootlab Organization
© 2013-2019 Patrick Michl