One key step towards machine learning scenarios is the reproducibility of an experiment as well as the interchanging of machine learning metadata. A notorious existing problem on different machine learning architectures is either the interchangeability of measures generated by executions of an algorithm and general provenance information for the experiment configuration. This demand tends to bring forth a cumbersome task of redefining schemas in order to facilitate the exchanging of information over different system implementations. This scenario is due to the missing of a standard specification. In this paper, we address this gap by presenting a built upon on a flexible and lightweight vocabulary dubbed MEX. We benefit from the linked data technologies to provide a public format in order to achieve a higher level of interoperability over different architectures.
[1]
Luís Torgo,et al.
OpenML: networked science in machine learning
,
2014,
SKDD.
[2]
Jens Lehmann,et al.
MEX vocabulary: a lightweight interchange format for machine learning experiments
,
2015,
SEMANTICS.
[3]
Paul T. Groth,et al.
Wings: Intelligent Workflow-Based Design of Computational Experiments
,
2011,
IEEE Intelligent Systems.
[4]
Geoff Holmes,et al.
Experiment databases
,
2012,
Machine Learning.