KR20200061877A - System and method for generating paraphrase sentence based on ontology - Google Patents

Abstract

The system for generating a paraphrase sentence acquires semantic resources including each of ontology components corresponding to tokens extracted from the query sentence, according to an exemplary embodiment of the present invention, and semantic A feature extraction unit for generating a frame having a structure in which tokens are replaceable in a pattern and query sentence, which are a combination of resources, a pattern index unit that correlates patterns and frames, and stores the corresponding patterns and frames in a pattern storage unit, and Based on the frames corresponding to the pattern stored in the pattern storage unit, it may include a sentence generation unit for generating a paraphrase sentence of the query sentence.

Description

System and method for generating ontology-based paraphrase sentences {SYSTEM AND METHOD FOR GENERATING PARAPHRASE SENTENCE BASED ON ONTOLOGY}

The technical idea of the present invention relates to the generation of paraphrase sentences, and more particularly, to a system and method for generating ontology-based paraphrase sentences.

The present invention is derived from a study conducted by the ILO Corporation and jointly researched by Saltlux Co., Ltd. as part of the Robot Industry Core Technology Development Project-Artificial Intelligence Convergence Robot System Technology by the Ministry of Commerce, Industry and Energy. [Research period: 2018.01.01~2018.12.31, Research management agency: Korea Institute of Industrial Technology Evaluation and Management, Research project name: Household social robot and service development system, task identification number: 10077633]

Recognizing and processing natural language so that machines understand human language can be referred to as natural language understanding, and natural language understanding can be used in various fields. For example, natural language understanding may be used in a question and answering system that automatically provides an answer to a query by recognizing the user's query.

In spite of the same meaning, since human language can be expressed in various ways, understanding of natural language may require understanding the same meaning from variously expressed sentences. Accordingly, preparing multiple sentences corresponding to the same meaning, that is, paraphrase sentences may be an important basis for understanding natural language.

The technical idea of the present invention provides a system and method for automatically generating paraphrase sentences based on ontology.

In order to achieve the above object, the system for generating a paraphrase sentence according to the technical idea of the present invention, the semantics including each of the ontology (ontology) components corresponding to the tokens extracted from the query sentence Acquire resources, and correlate feature extraction units, patterns and frames that generate frames having token-replaceable structures in patterns and query sentences that are combinations of semantic resources, and store the corresponding patterns and frames in the pattern storage unit It may include a pattern index unit, and a sentence generator for generating a paraphrase sentence of a query sentence based on frames corresponding to a pattern stored in the pattern storage unit.

According to an exemplary embodiment of the present invention, the feature extraction unit may obtain a plurality of query sentences through a network.

According to an exemplary embodiment of the present invention, the feature extraction unit receives a natural language processing unit that extracts tokens by performing natural language processing on a query sentence, and receives semantic resources corresponding to tokens from a knowledge base, and generates patterns and frames It may include a pattern generator.

According to an exemplary embodiment of the present invention, the pattern index unit may store patterns including each of semantic resources in different orders as different patterns in the pattern storage unit.

According to an exemplary embodiment of the present invention, the sentence generation unit applies at least one preliminary paradigm by applying the patterns corresponding to the pattern generated by the feature extraction unit in the pattern storage unit and the tokens to the searched frames. It may include a vocabulary substitution unit for generating a phrase sentence, and a sentence post-processing unit for generating a paraphrase sentence by modifying at least one preliminary paraphrase sentence according to a natural language rule.

According to an exemplary embodiment of the present invention, the lexical substitution unit may generate at least one preliminary paraphrase sentence by applying synonyms of tokens to searched frames based on semantic resources.

According to an exemplary embodiment of the present invention, the vocabulary substitution unit may generate at least one preliminary paraphrase sentence by changing the structure of the searched frame.

According to an exemplary embodiment of the present invention, the lexical substitution unit may generate an additional preparatory paraphrase sentence by changing the position of the adverb phrase in the searched frame.

According to an exemplary embodiment of the present invention, when the searched frame ends with a noun phrase, the lexical substitution unit may generate an additional preliminary paraphrase sentence by adding a query termination phrase.

According to the system and method according to the technical idea of the present invention, accurate and rich paraphrase sentences can be automatically generated by analyzing sentences based on a proven ontology.

In addition, according to the system and method according to the technical concept of the present invention, due to the abundantly provided paraphrase sentences, tasks based on natural language understanding, such as a query response system, knowledge extraction, etc., are significantly improved. Can be.

In addition, according to the system and method according to the technical idea of the present invention, quality data for machine learning may be prepared due to the abundantly provided paraphrase sentences, thereby significantly expanding the utilization of machine learning. Can be.

The effects obtainable in the exemplary embodiments of the present invention are not limited to the above-mentioned effects, and other effects not mentioned are common knowledge in the art to which the exemplary embodiments of the present invention belong from the following description. It can be clearly drawn and understood by those who have it. That is, unintended effects of implementing exemplary embodiments of the present invention can also be derived by those skilled in the art from the exemplary embodiments of the present invention.

Fig. 1 is a block diagram showing a system and its input/output relationship according to an exemplary embodiment of the present invention.
Fig. 2 is a block diagram showing an example of the characteristic extraction unit of Fig. 1 according to an exemplary embodiment of the present invention.
3A and 3B show examples of query sentences, patterns, and frames according to exemplary embodiments of the present invention.
4 shows an example of patterns and frames stored in the pattern storage unit of FIG. 1 according to an exemplary embodiment of the present invention.
5 is a block diagram illustrating an example of a sentence generation unit of FIG. 1 according to an exemplary embodiment of the present invention.
6A and 6B are diagrams illustrating examples of the operation of the vocabulary substitution unit of FIG. 5 according to exemplary embodiments of the present invention.
Fig. 7 is a flow chart showing a method for generating a paraphrase sentence according to an exemplary embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. The embodiments of the present invention are provided to more fully describe the present invention to those skilled in the art. The present invention can be applied to various changes and may have various forms, and specific embodiments will be illustrated in the drawings and described in detail. However, this is not intended to limit the present invention to a specific disclosure form, and it should be understood that it includes all modifications, equivalents, and substitutes included in the spirit and scope of the present invention. In describing each drawing, similar reference numerals are used for similar components. In the accompanying drawings, the dimensions of the structures are enlarged or reduced than actual ones for clarity of the present invention.

Terms used in the present application are only used to describe specific embodiments, and are not intended to limit the present invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this application, terms such as “include” or “have” are intended to indicate the presence of features, numbers, steps, actions, elements, parts or combinations thereof described in the specification, but one or more other features. It should be understood that the presence or addition possibilities of fields or numbers, steps, actions, components, parts or combinations thereof are not excluded in advance.

Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by a person skilled in the art to which the present invention pertains. Terms, such as those defined in a dictionary used in general, should be interpreted as having meanings consistent with meanings in the context of related technologies, and are interpreted as ideal or excessively formal meanings unless explicitly defined in the present application. Does not work.

In the following drawings and descriptions, components represented or described as one block may be hardware blocks or software blocks. For example, each of the components may be an independent hardware block that communicates with each other, or may be a software block executed in one processor. Also, as used herein, “system” or “database” may refer to a computing system that includes at least one processor and memory accessed by the processor.

Fig. 1 is a block diagram showing a system and its input/output relationship according to an exemplary embodiment of the present invention. As shown in FIG. 1, the paraphrase generation system 100 may be communicatively connected to the knowledge base 300 and the pattern storage unit 500. In addition, the paraphrase generation system 100 may receive a query sentence as a natural language, and generate a paraphrase sentence having the same meaning as a query sentence by communicating with the knowledge base 300 and the pattern storage unit 500. As will be described later, the paraphrase generation system 100 may extract a pattern from a query sentence with reference to the knowledge base 300 and store a plurality of frames corresponding to the pattern in the pattern storage unit 500 together with the pattern. Can be. The blocks 100, 300, and 500 shown in FIG. 1 may communicate with each other through a network, or may communicate with each other through a dedicated channel for one-to-one communication. Also, two or more of the blocks 100, 300, and 500 shown in FIG. 1 may be included in one system (eg, a computing system), and in some embodiments, the pattern storage unit 500 generates a paraphrase It may be included in the system 100.

The paraphrase generation system 100 may receive query sentences in various ways. In some embodiments, the paraphrase generation system 100 may include a user interface, and receive a query sentence from a person, ie, a user, through the user interface. In some embodiments, the paraphrase generation system 100 may connect to a network, such as the Internet, and may actively collect multiple query sentences from the network. As shown in FIG. 1, the paraphrase generation system 100 may include a feature extraction unit 120, a pattern index unit 140, and a sentence generation unit 160.

The feature extraction unit 120 may obtain semantic resources including each of ontology components corresponding to tokens extracted from the query sentence. Ontology is an expression that exists or can be recognized by human beings in a form that can be handled by a computer. ; P), value (V). Additionally, ontology components may further include relationships (inter-entity attributes or inter-class attributes), function terms, restrictions, rules, events, and the like. The knowledge base 300 may store vast amount of knowledge data based on an ontology. For example, the knowledge base 300 may include knowledge data expressed using a Resource Description Framework (RDF), and a triple ( triple) can be used. Knowledge base 300 may return a triple in response to a query, such as a SPARQL (SPARQL Protocol and RDF Query Language) query. In some embodiments, the feature extraction unit 120 may deliver a query sentence to a natural language recognition system (not shown) outside the paraphrase generation system 100, and may extract tokens extracted from the query sentence from the natural language recognition system. I can receive it. In some embodiments, the feature extraction unit 120 may generate semantic resources by directly processing a query statement, as described below with reference to FIG. 2.

The feature extraction unit 120 may generate patterns and frames from semantic resources obtained from query sentences based on the ontology. As will be described later with reference to FIGS. 3A and 3B, the pattern may correspond to the meaning indicated by the query sentence as a combination of semantic resources. That is, sentences corresponding to the same pattern may be referred to as paraphrase sentences. In addition, as described below with reference to FIGS. 3A and 3B, the frame may have a structure in which tokens extracted from a query sentence are replaceable in the query sentence. As described below with reference to FIG. 4, one pattern may correspond to at least one frame, and patterns and frames corresponding to each other may be stored in the pattern storage unit 500 by the pattern index unit 140. The feature extraction unit 120 may provide a pattern and a frame to the pattern indexing unit 140, and may provide a token list and pattern including the extracted tokens to the sentence generation unit 160.

The pattern indexer 140 may receive patterns and frames generated from the query sentence from the feature extractor 120. The pattern index unit 140 may correlate patterns and frames, and may store patterns and frames corresponding to each other in the pattern storage unit 500. Accordingly, when a pattern stored in the pattern storage unit 500 is searched (eg, by the sentence generation unit 160), corresponding frames may be searched together. In some embodiments, the pattern indexing unit 140 may search for a pattern received from the characteristic extraction unit 120 in the pattern storage unit 500, and the pattern received from the characteristic extraction unit 120 may be a pattern storage unit When the frame already received in the 500 and the frame received from the feature extraction unit 120 is also stored in the pattern storage unit 500, the storage of the pattern and the frame received from the feature extraction unit 120 may be omitted. have. The structure of the patterns and frames stored in the pattern storage unit 500 by the pattern indexing unit 140 will be described later with reference to FIG. 4.

The sentence generation unit 160 may receive a pattern and a token list corresponding to the query sentence from the feature extraction unit 120 and may generate a paraphrase sentence of the query sentence. The sentence generation unit 160 may search for a pattern in the pattern storage unit 500 and obtain the searched pattern and frames corresponding thereto from the pattern storage unit 500. For example, the sentence generation unit 160 may search for a pattern received from the feature extraction unit 120 in the pattern storage unit 500, and based on the token list and frames corresponding to the searched pattern, query sentences and A sentence having the same meaning, that is, a paraphrase sentence, can be generated. An example of the sentence generation unit 160 will be described later with reference to FIG. 5.

As described below with reference to the drawings, the paraphrase generation system 100 may generate a paraphrase sentence of a query sentence using the ontology-based knowledge base 300. Accordingly, the query sentence can be analyzed based on the verified ontology, and accurate and rich paraphrase sentences can be automatically generated. In addition, due to the paraphrase sentences generated by the paraphrase generation system 100, performances based on natural language understanding, such as a query response system and knowledge extraction, can be significantly improved, and machine learning Quality data for this can be properly prepared.

2 is a block diagram illustrating an example of the feature extraction unit 120 of FIG. 1 according to an exemplary embodiment of the present invention, and FIGS. 3A and 3B are query sentences, patterns, and patterns according to exemplary embodiments of the present invention. Examples of frames are shown. As described above with reference to FIG. 1, the feature extraction unit 120 ′ of FIG. 2 may receive a query sentence, and generate a pattern and a frame corresponding to the query sentence with reference to the knowledge base 300. . As shown in FIG. 2, the characteristic extraction unit 120 ′ may include a natural language processing unit 122 and a pattern generation unit 124, and FIGS. 2, 3A, and 3B will be described with reference to FIG. 1 below. Will be explained.

In some embodiments, the feature extraction unit 120 ′ may extract tokens by natural processing the query sentence, and obtain semantic resources corresponding to the tokens from the knowledge base 300. As described above with reference to FIG. 1, in some embodiments, unlike the example of FIG. 2, the feature extraction unit 120 of FIG. 1 is external to the paraphrase generation system 100 instead of directly processing the natural language of the query sentence. You can also send a query sentence to the natural language processing system of, and receive tokens from it.

The natural language processing unit 122 may extract tokens from the query sentence. For example, as shown in FIG. 3A, with respect to the first query sentence Q1, "What is the population of Seoul?", the natural language processing unit 122 "Seoul", "Ui", "Population", "A" , Token list including "?" can be generated. In some embodiments, the token extracted by the natural language processing unit 122 may generate characteristics as well as words. For example, the token may include a part-of-speech (eg, noun, adjective adverb, etc.) or a sentence component (eg, subject, object, adverb, etc.) in a query sentence. The natural language processing unit 122 can extract tokens in any manner. For example, the natural language processing unit 122 may extract tokens with reference to the knowledge base 300. For example, the natural language processing unit 122 may determine the part of speech of the word included in the query sentence based on the entities, attributes, and the like of the knowledge base 300. Further, in some embodiments, as described below with reference to FIGS. 6A and 6B, the extracted tokens may have a tree structure.

The pattern generation unit 124 may receive tokens from the natural language processing unit 122 and obtain semantic resources corresponding to the received tokens from the knowledge base 300. For example, as shown in FIG. 3A, for the first query sentence Q1, "What is the population of Seoul?", it is possible to obtain the entity "<Entity-City>" as a semantic resource of "Seoul". , As a semantic resource of "population", property "<Property_Population>" can be obtained. Accordingly, the first pattern P1 is semantic resources, and may include "<Entity-City>" and "<Property_Population>". In addition, as shown in FIG. 3B, for the second query sentence Q2 "Please tell the population of Seoul?", it is possible to acquire the entity "<Entity_City>" as a semantic resource of "LA"," The attribute "<Peoperty_Population>" can be obtained as a semantic resource of the population. The second pattern P2 may include “<Entity-City>” and “<Property_Population>”, and thus may be the same as the first pattern P1.

In some embodiments, a pattern can define a combination of semantic resources as well as a sequence of semantic resources. That is, even if patterns having the same semantic resources have different order of semantic resources, the patterns may be stored in the pattern storage unit 500 by the pattern indexer 140 as different patterns.

Also, the pattern generator 124 may generate a frame based on the query sentence and pattern. For example, as illustrated in FIG. 3A, based on the first query sentence Q1 and the first pattern P1, the pattern generation unit 124 is configured to generate a first frame F11, that is, "<Entity_City> <Property_Population>?". In the first frame F11, an instance such as “Seoul”, “LA”, etc. may be applied (or assigned) to the semantic resource corresponding to the entity, that is, “<Entity_City>”, and the semantic resource corresponding to the attribute, That is, "<Property_Population>" may be one of synonyms such as "population" and "population number". Accordingly, when applying the tokens included in the query sentence analyzed by the first pattern P1 in the first frame F11 as the query sentence different from the first query sentence Q1, in the first frame F11, corresponding A paraphrase sentence of the query sentence may be generated. Similarly, as illustrated in FIG. 3B, based on the second query sentence Q2 and the second pattern P2, the pattern generation unit 124 generates a second frame F12, that is, <<Entity_City>. Property_Population>." Similar to the first frame F11, in the second frame F12, an instance such as “Seoul”, “LA”, etc. can be applied to the semantic resource corresponding to the entity, that is, “<Entity_City>”, and the attribute One of the synonyms such as "population" and "population number" may be applied to the corresponding semantic resource, that is, "<Property_Population>". Accordingly, in the case where the tokens included in the query sentence analyzed by the first pattern P1 in the second frame F12 as the query sentence different from the second query sentence Q2 are applied to the second frame F12, the corresponding A paraphrase sentence of the query sentence may be generated.

In the example of FIGS. 3A and 3B, the first query sentence Q1 and the second query sentence Q2 correspond to the same pattern (that is, because the first pattern P1 and the second pattern P2 match) , By applying the tokens of the first query sentence Q1 to the second frame F12, a paraphrase sentence of the first query sentence Q1, that is, "Please tell me the population of Seoul." Similarly, by applying the tokens of the second query sentence Q2 to the first frame F11, a paraphrase sentence of the second query sentence Q2, that is, “What is the population of LA?” can be generated.

4 illustrates an example of patterns and frames stored in the pattern storage unit 500 of FIG. 1 according to an exemplary embodiment of the present invention. As described above with reference to FIG. 1, the pattern index unit 140 may correlate patterns and frames, and may store patterns and frames corresponding to each other in the pattern storage unit 500. In the following, FIG. 4 will be described with reference to FIG. 1.

As described above with reference to FIGS. 3A and 3B, multiple frames may correspond to a single pattern, and after analyzing a pattern corresponding to a given query sentence, tokens of the query sentence may be assigned to multiple frames corresponding to the pattern Multiple paraphrase sentences can be generated by applying. Accordingly, as illustrated in FIG. 4, the pattern storage unit 500 may include a plurality of groups G1 to Gn corresponding to each of the plurality of patterns P1 to Pn (n is less than 1). Large integer). Each of the plurality of groups G1 to Gn may include one pattern and a plurality of frames corresponding to the pattern. For example, the first group G1 may include the first pattern P1 and a plurality of frames F11 to F1x corresponding thereto (x is an integer greater than 1), and the second group G2. May include the second pattern P2 and a plurality of frames F21 to F2y corresponding thereto (y is an integer greater than 1), and the third group G3 corresponds to the third pattern P3 and the same. It may include a plurality of frames (Fn1 to Fnz) (z is an integer greater than 1).

The pattern index unit 140 may store patterns and frames in the pattern storage unit 500 to have a structure as shown in FIG. 4, while the sentence generation unit 160 is due to the structure shown in FIG. 4 Thus, a plurality of frames corresponding to the searched pattern can be obtained. For example, when the pattern received from the sentence generation unit 160 corresponds to the second pattern P2, the sentence generation unit 160 stores the second pattern P2 as a pattern storage unit. It is possible to search in (500), it is possible to obtain a plurality of frames (F21 to F2y) corresponding to the second pattern (P2). It will be appreciated that the example shown in FIG. 4 is only a pattern and a correspondence between a plurality of frames, and that any data structures can be used to implement the structure shown in FIG. 4. Further, in some embodiments, the frame may have a tree structure as described below with reference to FIG. 6A.

5 is a block diagram illustrating an example of the sentence generator 160 of FIG. 1 according to an exemplary embodiment of the present invention. As described above with reference to FIG. 1, the sentence generation unit 160 ′ of FIG. 5 may receive a pattern and a frame corresponding to a query sentence from the characteristic extraction unit 120 and stored in the pattern storage unit 500 A paraphrase sentence of a query sentence can be generated by referring to a pattern and a frame. As illustrated in FIG. 5, the sentence generation unit 160 ′ may include a pattern search unit 162, a vocabulary substitution unit 164, and a sentence post-processing unit 166. Will be explained with reference.

The pattern search unit 162 may receive a pattern from the characteristic extraction unit 120, and may search for frames corresponding to the pattern received from the characteristic extraction unit 120 in the pattern storage unit 500. As described above with reference to FIG. 4, since the pattern storage unit 500 may be structured to include patterns and frames corresponding to each other, the pattern search unit 162 searches the received pattern in the pattern storage unit 500 By doing so, frames corresponding to the pattern may be received from the pattern storage unit 500.

The vocabulary substitution unit 164 may receive a token list from the feature extraction unit 120 and receive frames from the pattern search unit 162. As described above with reference to FIG. 1, the token list may include tokens extracted from a query sentence, and the vocabulary substitution unit 164 may include the tokens included in the token list from the pattern search unit 162 By applying to, preliminary paraphrase sentences can be generated. For example, if the token list extracted from the first query sentence Q1 of FIG. 3A is received, and the second frame F12 of FIG. 3B is received from the pattern search unit 162, the vocabulary replacement unit 164 May generate a preliminary paraphrase sentence "Please tell me the population of Seoul." by applying "Seoul" and "Population" included in the token list to the second frame F12. Similarly, if the token list extracted from the second query sentence Q2 of FIG. 3B is received, and the first frame F11 of FIG. 3A is received from the pattern search unit 162, the vocabulary substitution unit 164 By applying "LA" and "population number" included in the token list to the first frame F11, a preliminary paraphrase sentence "What is the population of LA?" can be generated.

In some embodiments, the vocabulary substitution unit 164 may generate a preliminary paraphrase sentence by applying a synonym of the word corresponding to the token to the frame. For example, if the token list extracted from the second query sentence Q2 of FIG. 3B is received, and the first frame F11 of FIG. 3A is received from the pattern search unit 162, the vocabulary substitution unit 164 Is a preliminary paraphrase sentence, "What is the population of LA?" In addition, it is also possible to generate "What is the population of Los Angeles?" or "Please tell me the population of Los Angeles." based on the second frame (F12). Also, in some embodiments, the lexical substitution unit 164 may generate a preliminary paraphrase sentence by changing the structure of the frame. Examples of the vocabulary substitution unit 164 generating a preliminary paraphrase sentence by changing the structure of the frame will be described later with reference to FIGS. 6A and 6B.

The sentence post-processing unit 166 may receive a preliminary paraphrase sentence from the vocabulary substitution unit 164, and may generate a paraphrase sentence by correcting the preliminary paraphrase sentence according to natural language rules. The preliminary paraphrase sentence in which tokens are applied to the frame by the vocabulary substitution unit 164 may include a part in which the natural language rules are violated. For example, the preliminary paraphrase sentence may include an inappropriate investigation and/or ending, and the sentence post-processing unit 166 may generate a paraphrase sentence as a natural sentence by modifying the investigation and/or ending based on natural language rules. Can be.

6A and 6B are diagrams illustrating examples of operation of the vocabulary substitution unit 164 of FIG. 5 according to exemplary embodiments of the present invention. Specifically, FIGS. 6A and 6B show examples of the lexical substitution unit 164 that generates preliminary paraphrase sentences by changing the structure of the frame. The vocabulary substitution unit 164 may change the structure of the frame, or may change the structure of a preliminary paraphrase sentence in which tokens are applied to the frame. Hereinafter, FIGS. 6A and 6B will be described with reference to FIG. 5, and overlapping contents of the descriptions of FIGS. 6A and 6B will be omitted. For convenience of description, FIGS. 6A and 6B show examples of changing the structure of a preliminary paraphrase sentence in which tokens are applied to a frame.

Referring to FIG. 6A, the vocabulary substitution unit 164 may generate an additional preliminary paraphrase sentence by changing the position of the adverb phrase in the frame. For example, a preliminary paraphrase of "Which Smithsonian Art Museum is in the United States?" can be expressed in a tree structure as shown at the top of FIG. 6A. In FIG. 6A, reference numerals indicated below tokens are characteristics of tokens, and may indicate parts of speech and sentence components. For example, “VNP” may indicate “noun+is” as a positive designating phrase, “NP_SBJ” may indicate that a body (eg, noun, pronoun, rhetoric) is used as a subject, and “NP_AJT” It is possible to indicate that the body is used as an adverb (alternative modifier). In addition, "NP_MOD" may indicate that the body is used as a tubular language (a body modifier), and "DP" may indicate a tubular phrase.

The vocabulary substitution unit 164 may change the position of the sub-tree corresponding to the adverb. For example, as shown at the bottom of FIG. 6B, the vocabulary substitution unit 164 may change the location of the sub-tree corresponding to “to any city in the United States” and “Smithsonian Art Museum”. Accordingly, a preliminary paraphrase sentence, "Which city in the United States has the Smithsonian Art Museum?" can be generated. The sentence post-processing unit 166 of FIG. 5 can finally generate a paraphrase sentence, that is, "Which city in the United States has the Smithsonian Art Museum?"

Referring to FIG. 6B, the vocabulary substitution unit 164 may generate additional preparatory paraphrase sentences for one or more phrases. For example, a preliminary paraphrase sentence, "Which city has the Smithsonian Art Museum?" can be structured as shown on the left side of FIG. 6B. When the searched frame ends with a noun phrase, the vocabulary replacement unit 164 may generate an additional preliminary paraphrase sentence by adding a query termination phrase. For example, as shown on the right side of FIG. 6B, the vocabulary replacement unit 164 adds a "where" as the query terminator in the frame ending with the noun phrase "city", that is, a preliminary paraphrase sentence, that is, "Smith "Where is the city where the Sony Museum of Art is located?"

Fig. 7 is a flow chart showing a method for generating a paraphrase sentence according to an exemplary embodiment of the present invention. For example, the method of FIG. 7 may be performed by the paraphrase generation system 100 of FIG. 1. Specifically, steps S10, S20, and S30 may be performed by the feature extraction unit 120 of FIG. 1, and steps S40, S50, and S60 may be performed by the sentence generation unit 160. In the following, FIG. 7 will be described with reference to FIG. 1.

Referring to FIG. 7, in step S10, an operation of receiving a query sentence may be performed. For example, the feature extraction unit 120 may receive a query sentence from a user through a user interface or collect query sentences from a network.

In step S20, an operation of acquiring semantic resources may be performed. For example, the feature extraction unit 120 may acquire semantic resources corresponding to the extracted tokens from the knowledge base 300 by processing the query sentence in natural language. As described above with reference to the drawings, the knowledge base 300 may be constructed based on ontology and may include verified knowledge data, and thus may be utilized to generate a paraphrase sentence of a query sentence. The natural language processing of the query sentence may be performed outside the paraphrase generation system 100 in some embodiments, or may be performed by the feature extraction unit 120 in some embodiments.

In step S30, an operation for generating a pattern and a frame may be performed. For example, the feature extraction unit 120 may generate a pattern as a combination of semantic resources obtained based on a query sentence, and may generate a frame having a structure in which tokens are replaceable in a branch sentence. Also, the feature extraction unit 120 may generate a token list including the extracted tokens.

In step S40, an operation of searching for a pattern may be performed. For example, the pattern storage unit 500 may store patterns and frames corresponding to each other, and one pattern may correspond to a plurality of frames. The sentence generation unit 160 may search for the pattern generated in step S30, and the pattern storage unit 500 may obtain frames corresponding to the searched pattern from the pattern storage unit 500.

In step S50, an operation of applying tokens to the frame may be performed. For example, the sentence generation unit 160 may generate preliminary paraphrase sentences by applying tokens included in the token list to a frame obtained from the pattern storage unit 500. In some embodiments, the sentence generator 160 may generate an additional preparatory paraphrase sentence using a synonym of the token, or may generate an additional preparatory paraphrase sentence by changing a frame or a structure of the preliminary paraphrase sentence. have.

In step S60, post-processing of the sentence may be performed. For example, the preliminary paraphrase sentence generated in step S50 may include a part in which the natural language rule is violated, and the sentence generation unit 160 corrects the investigation and/or ending of the preliminary paraphrase sentence according to the natural language rule. Paraphrase sentences can be finally generated.

As described above, exemplary embodiments have been disclosed in the drawings and the specification. Although the embodiments have been described using specific terminology in this specification, they are only used for the purpose of explaining the technical spirit of the present invention and are not used to limit the scope of the present invention as defined in the claims or the claims. . Therefore, those skilled in the art will understand that various modifications and other equivalent embodiments are possible therefrom. Therefore, the true technical protection scope of the present invention should be determined by the technical spirit of the appended claims.

Claims (9)

A system for generating paraphrase sentences,
Semantic resources including each of ontology components corresponding to tokens extracted from a query sentence are obtained, and a pattern having a combination of the semantic resources and a frame having a structure in which the tokens are replaceable in the query sentence are generated. Characteristic extraction unit configured to;
A pattern index unit configured to correlate patterns and frames with each other, and to store mutually corresponding patterns and frames in a pattern storage unit; And
And a sentence generation unit configured to generate a paraphrase sentence of the query sentence based on frames corresponding to the pattern stored in the pattern storage unit. The method according to claim 1,
The feature extraction unit, characterized in that the system is configured to obtain a plurality of query sentences through the network. The method according to claim 1,
The characteristic extraction unit,
A natural language processor configured to extract the tokens by performing natural language processing on the query sentence; And
And a pattern generator configured to receive the semantic resources corresponding to the tokens from the knowledge base, and to generate the pattern and the frame. The method according to claim 1,
And the pattern index unit is configured to store patterns each containing semantic resources of different order as the different patterns in the pattern storage unit. The method according to claim 1,
The sentence generation unit,
A pattern search unit configured to search frames corresponding to the pattern generated by the feature extraction unit in the pattern storage unit;
A vocabulary substitution unit configured to generate at least one preliminary paraphrase sentence by applying the tokens to the searched frames; And
And a sentence post-processing unit configured to generate the paraphrase sentence by modifying the at least one preliminary paraphrase sentence according to a natural language rule. The method according to claim 5,
And the vocabulary replacement unit is configured to generate the at least one preliminary paraphrase sentence by applying synonyms of the tokens to the searched frames based on the semantic resources. The method according to claim 5,
And the vocabulary replacement unit is configured to generate the at least one preliminary paraphrase sentence by changing the structure of the searched frame. The method according to claim 7,
And the vocabulary replacement unit is configured to generate an additional preliminary paraphrase sentence by changing the position of the adverb phrase in the searched frame. The method according to claim 7,
The vocabulary replacement unit is configured to generate an additional preliminary paraphrase sentence by adding a query terminator when the searched frame ends with a noun phrase.

Images