Parse arbitrator for arbitrating between candidate descriptive parses generated from descriptive queries

A non-transitory computer-readable medium, storing program instructions that, when executed by one or more processors of a computing system, cause the computing system to perform operations comprising:

• receiving, from a descriptive classifier, an indication of whether a textual input query is non-descriptive or descriptive, wherein the textual input query was obtained by way of a client device and includes one or more terms, wherein non-descriptive text identifies a specific known media content item, and wherein descriptive text identifies a category of media content items;
• receiving, from a rule-based parser, a first parse output of the textual input query, wherein the rule-based parser tags the one or more terms based on rules to identify the first parse output;
• receiving, from a machine-learning-based parser, a second parse output, wherein the machine-learning-based parser tags the one or more terms based on one or more trained models to identify the second parse output;
• based on at least (i) the indication from the descriptive classifier of whether the textual input query is non-descriptive or descriptive, and (ii) the first parse output from the rule-based parser, selecting a search system from either a rule-based descriptive search system or a machine-learning-based non-descriptive search system;
• routing one of the first parse output or the second parse output to the search system, wherein the first parse output is routed when the rule-based descriptive search system is selected, and wherein the second parse output is routed when the machine-learning-based non-descriptive search system is selected;
• receiving, from the search system, a representation of one or more media content items;
• and
• providing, to the client device, the representation of the one or more media content items.

A computing system comprising:

• one or more processors;
• memory;
• and
• program instructions, stored in the memory, that upon execution by the one or more processors cause the computing system to perform operations comprising:
• receiving, from a descriptive classifier, an indication of whether a textual input query is non-descriptive or descriptive, wherein the textual input query was obtained by way of a client device and includes one or more terms, wherein non-descriptive text identifies a specific known media content item, and wherein descriptive text identifies a category of media content items;
• receiving, from a rule-based parser, a first parse output of the textual input query, wherein the rule-based parser tags the one or more terms based on rules to identify the first parse output;
• receiving, from a machine-learning-based parser, a second parse output, wherein the machine-learning-based parser tags the one or more terms based on one or more trained models to identify the second parse output;
• based on at least (i) the indication from the descriptive classifier of whether the textual input query is non-descriptive or descriptive, and (ii) the first parse output from the rule-based parser, selecting a search system from either a rule-based descriptive search system or a machine-learning-based non-descriptive search system;
• routing one of the first parse output or the second parse output to the search system, wherein the first parse output is routed when the rule-based descriptive search system is selected, and wherein the second parse output is routed when the machine-learning-based non-descriptive search system is selected;
• receiving, from the search system, a representation of one or more media content items;
• and
• providing, to the client device, the representation of the one or more media content items.