Learning an efficient range-string metalabel query system

A computer-implemented method to compute queries for data associated with heterogenous hierarchical range metalabels using a data processor for operating a neural network, the method comprising:

• receiving a query from a user, wherein the query is specified in natural language;
• using the neural network to analyze the heterogenous hierarchical range metalabels assigned to the data to determine corresponding structure, wherein the hierarchical metalabel is an ordered sequence of atomic labels;
• using the neural network to convert the user query to an optimized query on the database that represents and stores the ordered sequence of attributes representing one or more metalabels in a database that supports conjunctive and disjunctive queries, wherein the neural network uses the attributes and a loss function to optimize a metalabel range query;
• and
• returning the optimized query.

A computer-implemented method to compute queries for data associated with heterogenous hierarchical metalabels using a data processor for operating a neural network, the method comprising:

• receiving a part-query from a user, wherein the part-query is specified in natural language;
• using the neural network to analyze the heterogenous hierarchical metalabels assigned to the data to determine corresponding structure;
• wherein the hierarchical metalabel is an ordered sequence of atomic labels;
• determining and displaying to the user a list of possible extensions to the part-query;
• iterating over receiving a part-query and determining for the user a list of part-queries to generate a final query
• using the neural network to convert the final query to an optimized query on the database that stores the ordered sequence of attributes representing one or more metalabels;
• wherein the neural network uses the attributes and a loss function to optimize a metalabel range query;
• and
• returning the optimized query.