Machine Tagging Challenge

The Seeker desires an algorithm to map customer-supplied tag names for data sources onto Seeker-defined Standard Tags. The algorithm must perform the mapping operation without human intervention and must be able to adapt to new data as it becomes available.

This is a Reduction-to-Practice Challenge that requires written documentation, output from the algorithm, and submission of working software and source code implementing the solution.

The Seeker receives sensor data from machines installed at numerous customer sites around the world. Each machine may produce hundreds of sensor data streams and each stream is tagged by the customer to indicate its source sensor. Unfortunately not all customers use the same tagging scheme so data from the same sensor on different machines may have different tags depending on the customer or even on the specific machine producing the data. The customer tags must be mapped to standard tags by the Seeker in order to analyze the data from different machines. The Seeker is in need of an algorithm that can do the mapping automatically with no human intervention. A large training set of data is provided for algorithm development and a smaller testing set of data is provided for Solvers to run their algorithm against and submit the output. The algorithm should be able to learn from incoming data.

A submission to the Challenge should include the following:

1. Working software that fulfill the required functionalities that map the data, including source code and executable with all dependencies as well as sufficient documentation to enable the Seeker to compile, execute, and validate the method using additional validation data sets.
2. A detailed description of the proposed Solution addressing specific Technical Requirements presented in the Detailed Description of the Challenge. This should also include a thorough description of the algorithm used in the Solution accompanied by a well-articulated rationale for the method employed.
3. Output from the proposed algorithm applied to a testing data set. Output must be in the form described in the Detailed Description of the Challenge.  Submissions will be ranked by the Seeker based on accuracy against the testing data set.

The Challenge award is contingent upon theoretical evaluation of the method/algorithm by the Seeker, and validation by the Seeker of the submitted reduction to practice Solution by testing the delivered software with additional validation data.

To receive an award, the Solvers will not have to transfer their exclusive IP rights to the Seeker. Instead, Solvers will grant to the Seeker a non-exclusive license to practice their solutions. 

Submissions to this Challenge must be received by 11:59 PM (US Eastern Time) on January 30, 2018 February 13, 2018.

Late submissions will not be considered.