Search-based natural language intent determination

An Intelligent Personal Assistant (IPA)-enabled system, the IPA-enabled system comprising:

• a non-transitory memory;
• and
• one or more hardware processors coupled to the non-transitory memory and configured to read instructions from the non-transitory memory to cause the IPA-enabled system to perform operations comprising:
• receiving, by a smart device having the IPA-enabled system, a command having one or more words from a user;
• parsing, using a natural language processor, the one or more words in the command for one or more relationships between the one or more words;
• determining, using at least one machine learning (ML) model of the natural language processor, a user intent for the command based on the parsing and a previous command by the user, wherein the ML model of the natural language processor is trained based on a plurality of past commands to at least one of the smart device or one or more corresponding smart devices, and wherein the ML model is further trained based on a plurality of command variants for the command and additional commands by a plurality of users including the user;
• determining, using the at least one ML model, at least one of the plurality of command variants of the command;
• determining a plurality of search results each having a corresponding relevance score based on a search using the command, the user intent, and the at least one of the plurality of command variants;
• ranking the plurality of search results based on the corresponding relevance score of each of the plurality of search results;
• determining at least one of an available component or an available communication mechanism for use by the smart device;
• determining an action to execute based on the ranked plurality of search results and the at least one of the available component or the available communication mechanism;
• and
• executing, by the smart device, the action using the at least one of the available component or the available communication mechanism.

A non-transitory computer-readable medium having stored thereon instructions executable by a computer system to cause the computer system to perform operations comprising:

• receiving, by a smart device, a command having one or more words from a user;
• parsing, using a natural language processor, the one or more words in the command for one or more relationships between the one or more words;
• determining, using at least one machine learning (ML) model of the natural language processor, a user intent for the command based on the parsing and a previous command by the user, wherein the ML model of the natural language processor is trained based on a plurality of past commands to at least one of the smart device or one or more corresponding smart devices, and wherein the ML model is further trained based on a plurality of command variants for the command and additional commands by a plurality of users including the user;
• determining, using the at least one ML model, at least one of the plurality of command variants of the command;
• determining a plurality of search results each having a corresponding relevance score based on a search using the command, the user intent, and the at least one of the plurality of command variants;
• ranking the plurality of search results based on the corresponding relevance score of each of the plurality of search results;
• determining at least one of an available component or an available communication mechanism for use by the smart device;
• determining an action to execute based on the ranked plurality of search results and the at least one of the available component or the available communication mechanism;
• and
• executing, by the smart device, the action using the at least one of the available component or the available communication mechanism.