Query classification and processing using neural network based machine learning

A method comprising:

• training a neural network using a training set of queries to predict query execution metrics, wherein the neural network includes a memory layer that is trained based at least in part on how tokens are ordered within at least a subset of queries in the training set of queries;
• determining a sequence of tokens for at least a portion of a query;
• determining a vector representation for each token in the sequence of tokens;
• generating, using the neural network based on the vector representation for each token in the sequence of tokens, a prediction for one or more metrics associated with executing at least the portion of the query, wherein the neural network determines at least a first weight for at least a first token in the sequence of tokens and a second weight for at least a second token in the sequence of tokens, based at least in part on at least a first vector representation for the first token and a second vector representation for the second token that preceded the first token in the sequence of tokens, wherein the neural network generates the prediction based at least in part on the first weight and the second weight;
• and
• triggering a responsive action based at least in part on the prediction for one or more metrics associated with executing at least the portion of the query.

A system comprising:

• one or more hardware processors;
• one or more non-transitory computer-readable media storing instructions, which when executed by the one or more hardware processors, cause:
• training a neural network using a training set of queries to predict query execution metrics, wherein the neural network includes a memory layer that is trained based at least in part on how tokens are ordered within at least a subset of queries in the training set of queries;
• determining a sequence of tokens for at least a portion of a query;
• determining a vector representation for each token in the sequence of tokens;
• generating, using the neural network based on the vector representation for each token in the sequence of tokens, a prediction for one or more metrics associated with executing at least the portion of the query, wherein the neural network determines at least a first weight for at least a first token in the sequence of tokens and a second weight for at least a second token in the sequence of tokens, based at least in part on at least a first vector representation for the first token and a second vector representation for the second token that preceded the first token in the sequence of tokens, wherein the neural network generates the prediction based at least in part on the first weight and the second weight;
• and
• triggering a responsive action based at least in part on the prediction for one or more metrics associated with executing at least the portion of the query.