Machine learning model repository

Abstract

Embodiments are directed towards a machine learning repository for managing machine learning (ML) model envelopes, ML models, model objects, or the like. Questions and model objects may be received by a ML model answer engine. Machine learning (ML) model envelopes may be received based on the questions. The model objects may be compared to parameter models associated with the ML model envelopes. ML model envelopes may be selected based on the comparison such that the model objects satisfy the parameter models of each of the selected ML model envelopes. ML models included in each selected ML model envelope may be executed to provide score values for the model objects and the score values may be included in a report.

What is claimed as new and desired to be protected by Letters Patent of the United States is: 
     
         1 . A method for managing data over a network using one or more processors, included in one or more network computers, to perform actions, comprising:
 instantiating an answer engine to perform further actions, including:
 receiving one or more questions and one or more model objects, wherein the one or more model objects are part of a data model that conforms to a model schema; 
 receiving a plurality of machine learning (ML) model envelopes based on the one or more questions; 
 comparing the data model to parameter models that are associated with each of the plurality of ML model envelopes, wherein the comparison includes a traversal of the data model and one or more of the parameter models; 
 selecting one or more of the plurality of ML model envelopes based on the comparison, wherein one or more traversal paths corresponding to the one or more model objects satisfy the parameter models of each of the selected one or more ML model envelopes; 
 executing one or more ML models included in each selected ML model envelope to provide score values for the one or more model objects, wherein the score values are included in a report; and 
 providing selective optimization of one or more of performance or storage size for one or more ML model repositories based on one or more characteristics including one or more of model object usage frequency, number of model objects in a query result, model object size, or model object data type, wherein the one or more selective optimizations include one or more of indices to improve identification of each model object to be omitted from the one or more traversal paths, or storing a portion of the one or more model objects in a database or a fast data store.