System for Natural Language Understanding

• US 20170011023A1
• Filed: 07/07/2015
• Published: 01/12/2017
• Est. Priority Date: 07/07/2015
• Status: Active Grant

• Current Assignee: Maan Ghannam, Rima Ghannam
• Original Assignee: Maan Ghannam, Rima Ghannam
• Inventors: GHANNAM, Rima, GHANNAM, Maan
• Granted Patent: US 9,824,083 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/90332 Natural language query formulation or dialogue systems
• CPC Class Codes: G06F 40/205 Parsing
• CPC Class Codes: G06F 40/211 Syntactic parsing, e.g. based on context-free grammar [CFG] or unification grammars
• CPC Class Codes: G06F 40/30 Semantic analysis

Abstract

A general-purpose apparatus for analyzing natural language text that allows for the implementation of a broad range of natural language understanding applications. The apparatus for natural language understanding analyzes a source text and transforms the source text into a semantically-interpretable syntactic representation (SISR), comprising a syntax template and semantic clause annotations. The general-purpose apparatus for natural language understanding is adaptable to various source text natural languages and is adaptable to various natural language understanding applications, such as query answering, translation, summarization, information extraction, disambiguation, and parsing. A natural language query answering apparatus for answering questions about a source text, whereby the query answering apparatus utilizes the general-purpose apparatus for transforming the natural language query into SISR format.

Claims

1. A machine implemented query answering method, for receiving a digitally encoded query in a semantically-interpretable syntactic representation (SISR), and a natural language source text digitally encoded in SISR and generating an application output,

comprising:

query classification;

interpretation; and

SISR decoding to convert the SISR into a natural language response;

wherein the SISR of a natural language (NL) text is an outcome of a reverse engineering process of a sentence construction out of initial independent clauses, and wherein the SISR is a structured representation of standard templates and fields, and wherein each existent initial independent clause in the said NL text is represented in a single SISR entry, wherein each clause of the said existent clauses is represented in its complete, independent and declarative form and in the active voice, and in a complement-free (unless obligatory) manner and without the linking expressions and the conjunctions external to the clause and represented also in terms of units, wherein a unit comprises one or more words able to be associated together to represent a grammatical function as noun, verb, preposition, verbal phrase, adjective, comparative adjective, superlative adjective, superlative adverb, comparative adverb, or copula,

and wherein each SISR entry comprises;

an entry identifier;

a clause of the said existent clauses represented in terms of the said units in a syntax template;

the complements of the clause of the said existent clauses, represented in terms of the said units; and

annotations;

wherein the said annotations are data comprising information related to the SISR entry and its components.

2. The query answering method according to claim 1, further comprising a lexicon of question prototype categories wherein the said lexicon consists of left side entries of the question prototypes and the right side value of either:

‘

cause’

, ‘

effect’

, ‘

goal’

, ‘

time’

, ‘

number’

, ‘

amount’

, ‘

subject’

, ‘

object’

, ‘

manner’

, ‘

location’

, ‘

proposition truth’

, ‘

preposition’

, ‘

adjective’

or ‘

the entire proposition’

.

3. The query answering method according to claim 1, further comprising a fact-act lexicon, wherein each left side entry of verb or verb occurrence prototype corresponds to a right side value of either ‘

fact’

or ‘

act’

depending on whether the said entry relates to an intentional act.

4. The query answering method according to claim 1, further comprising generating a semantically expanded SISR upon receiving the said digitally encoded query in a SISR and the said digitally encoded natural language source text in a SISR.

5. The semantically expanded SISR of claim 4, wherein the said semantically expanded SISR is a converged version, wherein a converged version could not be subject to further expansion.

6. The query answering method according to claim 1, wherein the query classification comprises the classification of the application request into one of the following query categories:

cause, effect, goal, time, number, amount, subject, object, manner, location, proposition truth, adjective, or relation, by matching the SISR application request to a query classification lexicon.

7. The query answering method according to claim 1, wherein the SISR interpretation comprises:

act-fact tagging;

clause subcategory implication;

transitivity processing;

variable resolving;

set processing;

time processing;

condition realization checking;

relation determining;

goal tracing;

arithmetic processing;

Boolean processing;

clause equivalence checking; and

clause components processing.

8. The SISR interpretation according to claim 7, wherein the act-fact tagging accesses an act-fact lexicon.

9. The query answering method according to claim 1, wherein the SISR interpretation comprises input of a target data, a question, a question prototype to output one or more arguments wherein the said arguments are values conform to parameters worked out by various interpretation operations, and wherein the said arguments are key values for a valid answer of the said question.

10. The query answering method according to claim 1, wherein the SISR decoding generates a natural language application output based upon the result of the SISR interpretation.

11. The query answering method according to claim 1, wherein the SISR decoding further comprises:

answer formulating, wherein an answer template to a question prototype is assigned;

reverse mapping processing, wherein a semantically augmented SISR is mapped back to its origin; and

a reverse SISR formatting processing, wherein SISR clauses are converted into natural language expressions.

12. A web server connected to a network comprising query answering according to the method of the claim 1 and its dependent claims herein,

wherein the web server is configured to receive requests over the network encoding natural language queries, and

wherein the said query answering according to the said method is configured to process said natural language queries, and

wherein the web server is configured to return the application output obtained through the query answering method as a response over the network.