USER INTENT AND CONTEXT BASED SEARCH RESULTS
A user statement associated with a natural query is received. A syntactic parse of the user statement is performed to generate a parsed user statement. The parsed user statement is matched against a set of one or more interpretations determined to have meaning in a context of a knowledge base with which the user statement is associated. A user intent is determined based at least in part on said one or more interpretations. A determined query is performed based on said user intent.
1 . A system, comprising:
a communication interface; and
a processor coupled to the communication interface and configured to:
receive via the communication interface a user statement associated with a natural query;
perform a syntactic parse of the user statement to generate a parsed user statement;
match the parsed user statement against a set of one or more interpretations determined to have meaning in a context of a knowledge base with which the user statement is associated;
determine a user intent based at least in part on said one or more interpretations; and
perform a determined query based on said user intent.
2 . The system of claim 1 , wherein the syntactic parse comprises mapping raw bytes of user input to low-level parts of natural language.
3 . The system of claim 2 , wherein the mapping comprises at least one of the following:
normalization of encoding systems; recognition of intentional and unintentional variations of is terms; detection of non-alphabetical data; labelling of terms according to natural language models; and detection of spans.
4 . The system of claim 3 , wherein the recognition of intentional and unintentional variations of terms comprises at least one of the following: spelling errors, alternate spellings, abbreviations, shortcuts, and emoji.
5 . The system of claim 3 , wherein labelling of terms comprises labelling at least one of the following: adjective, noun, preposition, conjugation, and declension.
6 . The system of claim 3 , wherein detection of spans comprises detection of one or more terms that represent a discrete concept in a mind of a user.
7 . The system of claim 6 , wherein the detection of spans comprises a domain and a probability.
8 . The system of claim 2 , wherein the mapping comprises multiple incompatible segmentations and parses of user input.
9 . The system of claim 1 , wherein matching the parsed user statement comprises a semantic and grammatical parse.
10 . The system of claim 9 , wherein the semantic and grammatical parse comprises at least one of the following: adjectival filters; categorical filters; prepositional entity relationships; target domain inference; grammatical relationships; implicative grammatical relationships; discourse state concepts; and discourse state objects.
11 . The system of claim 9 , wherein the semantic and grammatical parse comprises at least one of the following: a Viterbi search algorithm and a domain pruning.
12 . The system of claim 1 , wherein an interpretation of the set of one or more interpretations comprises a grammatical tree representing an understanding of the user statement.
13 . The system of claim 12 , wherein a node on the grammatical tree is tagged with at least one of the following: its syntactic role; its grammatical role; and its semantic role.
14 . The system of claim 1 , wherein the processor is further configured to generate a machine readable query at least in part by resolving an unbound concept in the interpretation, wherein the determined query is the machine readable query.
15 . The system of claim 14 , wherein resolving an unbound concept in the interpretation comprises binding it to an object associated with a search.
16 . The system of claim 14 , wherein binding comprises determining based at least in part on a user context, wherein the user context comprises a user location.
17 . The system of claim 14 , wherein binding comprises determining based at least in part on a user conversation state, wherein the user conversation state comprises a conversation vector.
18 . The system of claim 1 , wherein the processor is further configured to generate a clarifying question in the event the parsed user statement matches a plurality of interpretations.
19 . A method, comprising:
receiving a user statement associated with a natural query;
performing a syntactic parse of the user statement to generate a parsed user statement;
matching the parsed user statement against a set of one or more interpretations determined to have meaning in a context of a knowledge base with which the user statement is associated;
determining a user intent based at least in part on said one or more interpretations; and
performing a determined query based on said user intent.
20 . A computer program product, the computer program product being embodied in a tangible computer readable storage medium and comprising computer instructions for:
receiving a user statement associated with a natural query;
performing a syntactic parse of the user statement to generate a parsed user statement;
matching the parsed user statement against a set of one or more interpretations determined to have meaning in a context of a knowledge base with which the user statement is associated;
determining a user intent based at least in part on said one or more interpretations; and
performing a determined query based on said user intent.