Discovery Foundation
The purpose of this foundation package is to provide a common platform through a set of abstractions that support the core objectives of the Accelerated Machine Learning initiative. it applies the concepts of Parameterised Intent and its Separation of Concerns (SoC), based around advanced OOD patterns, to provide a common foundation to differing needs of data scientist and productisation coders while sharing common ideas and their implementation.
Parametrized Intent is a unique technique extracting the ideas and thinking of the data scientist or development specialist from their discovery code and capturing it as intent with parameters that can be replayed in a productionized environment. This decoupling and Separation of Concern between data, code and intended actions considerably improves transparancy of ideas, code reuse and reduced time to market.
Accelerated Machine Learning is a unique approach around machine learning that innovates the data science discovery vertical and productization of the data science delivery model. More specifically, it is an incubator project that shadowed a team of Ph.D. data scientists in connection with the development and delivery of machine learning initiatives to define measurable benefit propositions for customer success. To accomplish this, the project developed specific and unique knowledge regarding transition and preparation of data sets for algorithmic execution and augmented knowledge, which is at the core of the projects services offerings. From this the project developed a new approach to data science discovery and productization dubbed “Accelerated Machine Learning”.
Contents
1 Installation
1.1 package install
The best way to install this package is directly from the Python Package Index repository using pip
$ pip install discovery-foundation
if you want to upgrade your current version then using pip
$ pip install --upgrade discovery-foundation
2 Package Overview
2.1 Class Diagram
From the Class Diagram below the key abstarctions are the Abstracted Handlers
and the
Abstracted Cleaners
2.2 DataManager
the DataManager
class is a concrete implementation of the AbstractPropertiesManager
and a built in
extension of the foundation package to facilitate the management of contract properties of the connector handlers and
parameterised intent
2.3 AugmentedManager
the AugmentedManager
class is a concrete implementation of the AbstractPropertiesManager
and a built in
extension of the foundation package to facilitate the management of Augmented Knowledge.
** Augmented Knowledge** is the capture of information on data, activities and the rich stream of subject matter expertise, injected into the machine learning discovery vertical to provide an augmented n-view of the discovery journey. This might include security, sensitivity, data value scaling, data dictionary, terms of reference, observations, performance, optimization, bias, etc. This enriched view of data allows, amongst other things, improved knowledge share, AI explainability, feature transparency, and accountability that feeds into AI ethics, and insight analysis.
3 Reference
3.1 Python version
Python 2.6 and 2.7 are not supported. Although Python 3.5 is supported, it is recommended to install
discovery-foundation
against the latest Python 3.7.x whenever possible.
Python 3 is the default for Homebrew installations starting with version 0.9.4.
3.2 GitHub Project
Discovery-Transitioning-Utils: https://github.com/Gigas64/discovery-foundation.
3.3 Change log
See CHANGELOG.
3.4 Licence
BSD-3-Clause: LICENSE.