Intelligent question answering method and device based on enterprise entity

An intelligent question answering and enterprise technology, applied in the field of knowledge question answering, can solve problems such as lack of knowledge and low accuracy of knowledge question answering, and achieve the effect of improving accuracy

Inactive Publication Date: 2021-06-08
北京海致星图科技有限公司
6 Cites 1 Cited by

AI-Extracted Technical Summary

Problems solved by technology

[0003] Existing question answering systems are basically knowledge question answering systems in the field of general knowledge, and question answering methods only solve knowledge question answering problems by matching quest...
View more

Method used

The intelligent question answering device based on the enterprise entity provided by the present embodiment adopts the entity disambiguation technology of the PageRank graph model, so that the abbreviation and the brand name of the enterprise entity can be accurately linked to the corresponding enterprise entity in the enterprise knowledge base, further improving The accuracy rate of enterprise knowledge question answering.
The intelligent question answering method based on the enterprise entity provided by the present embodiment adopts the entity disambiguation technology of the PageRank graph model, so that the abbreviation and the brand name of the enterprise entity can be accurately linked to the corresponding enterprise entity in the enterprise knowledge base, further improving The accuracy rate of enterprise knowledge question answering.
The intelligent question-and-answer device based on the enterprise entity provided by the present embodiment adopts the complete matching method to retrieve the corresponding enterprise entity of the target enterprise entity, stores the full name of the enterprise entity in advance, improves the retrieval matching speed, and then improves the speed of enterprise knowledge question-and-answer , the matching dictionary of abbreviations and brand names is pre-stored in the enterprise knowledge base, which improves the accuracy of the link between the target enterprise entity and the enterprise knowledge base, and further improves the accuracy of enterprise knowledge questions and answers.
The intelligent question-and-answer device based on the enterprise entity provided by the present embodiment adopts the intelligent question-and-answer method based on the enterprise knowledge base and deep learning technology. The generation stage provides strong technical support and can play an important role in solving the question-and-answer scenario surrounding enterprise entities. The present invention adopts the entity recognition technology based on the BiLSTM-CRF model in the problem understanding process, and can accurately identify the abbreviation and brand name of the enterprise entity. The MV-LSTM model is used in the matching process. This model can perform deep semantic matching between the core question of the question and the attributes of the enterprise entity in the enterprise knowledge base from multiple semantic dimensions, which greatly improves the accuracy of matching the answer to the question.
The intelligent question-and-answer method based on the enterprise entity provided by the present embodiment adopts the enterprise entity corresponding to the target enterprise entity to be retrieved by the complete matching method, stores the full name of the en...
View more

Abstract

The embodiment of the invention provides an intelligent question answering method and device based on enterprise entities, which are applied to a server. The method comprises the steps: recognizing a target enterprise entity in a question input by a user; obtaining a final enterprise entity corresponding to the target enterprise entity based on a corresponding relationship between the enterprise entity and an enterprise knowledge base; determining an attribute set of the final enterprise entity, and taking the attribute set as a candidate answer set of the target enterprise entity; and based on an MV-LSTM model, determining a matching value between the question input by the user and each attribute in the attribute set, and taking the attribute corresponding to the maximum matching value as a final answer to the question input by the user. By applying the scheme provided by the embodiment of the invention, the accuracy of enterprise knowledge questions and answers can be improved.

Application Domain

Digital data information retrievalCharacter and pattern recognition +3

Technology Topic

EngineeringData mining +6

Image

  • Intelligent question answering method and device based on enterprise entity
  • Intelligent question answering method and device based on enterprise entity
  • Intelligent question answering method and device based on enterprise entity

Examples

  • Experimental program(6)

Example Embodiment

[0082] Example one
[0083] See figure 1 , figure 1 A first flow diagram of an intelligent question and answer method based on an enterprise entity is applied to the server, such as figure 1 As shown, the method includes the following steps:
[0084] S110, identify the target enterprise entity in the user input question;
[0085] The process is to obtain target enterprise entities within the user input, and link to enterprise-enriched enterprise entities. The business entity in the question may be full name, referred to as a brand name, such as "Bank of China Co., Ltd.", "Baidu", "Hungry", "Bank of China Co., Ltd.", "Hungry", "Bank of China Co., Ltd.", "Hungry" Enterprise entity.
[0086] Alternatively, the present embodiment identifies the target enterprise entity in the user input question based on the BILSTM-CRF model.
[0087] The BILSTM_CRF model is a two-way LSTM (long-term memory network, long short-term memory) model and CRF (condition random, Condensional Random Fields) model. The LSTM model solves the defects of the original circulatory neural network, which has become the current most popular neural network model, successful application in many fields such as speech recognition, picture description, natural language processing, and the CRF model is a discriminant probability model, which is an airport. One, is often used in labeling or analyzing sequence information, such as natural language or biological sequence.
[0088] In the Bilstm_CRF model training stage, the training data set adopts the People's Daily 1998 annotation, and the data enhancement strategy is used to expand the corporate. It is to replace some companies in the people's daily in accordance with certain proportion of enterprise knowledge base. Entities, experiments show that the strategy can improve the identification accuracy of enterprise entities. The model training phase uses the Bies and O labeling system to mark the corners, namely Begin, Inner, End, Single and Other, the loss function is the cross entropy loss function, and the Momentum Optimization Algorithm is used for model parameters learning.
[0089] When data training, each word in each training statement is quantified by Embedding mode. After entering the two-way LSTM model, get the hidden layer output of each time T, and then give the two-way hidden layers of each time. Finally, use the full connection layer Mapped to n × k dimension R n×k The vectors P obtained in the space, where K is all the number of named entity tabs. After entering the vector P to the CRF layer, the final score function Where A is the label from Y i Y i+1 Transfer matrix, then use the SoftMax function to make globally normalized score functions.
[0090] Specifically, based on the BILSTM-CRF model, identify the target enterprise entity in the user input question, including:
[0091] When each single word in the user input question is quantified, entered into the BILSTM-CRF model, obtain the label of each single word;
[0092] The maximum probability of label sequence is calculated using the Viteri algorithm;
[0093] According to the label sequence, the target enterprise entity in the user input question is determined.
[0094] S120, based on the correspondence between enterprise entities and corporate knowledge base, the final enterprise entity corresponding to the target enterprise entity;
[0095] In order to further improve the accuracy of the enterprise entity in the corporate knowledge base, the company's abbreviated as a commonly known company's referreditter ornament is used to supplement the company's knowledge base, such as common Internet companies, financial technology companies, etc.
[0096] In question, the company may have a company, such as "Baidu", "Tencent" and company brand name "Hungry", "US Mission" and company have used names, "Baidu", "Baidu", "" Baidu "correspondence Baidu Online Network Technology (Beijing) Co., Ltd. "," Tencent "correspondence is" Shenzhen Tencent Computer System Co., Ltd. "," Hungry, "Correspondence Enterprise Entity is" Shanghai Razas Information Technology " Co., Ltd., "US Mission", the "US Mission" corresponds to the "Beijing Third Express Technology Co., Ltd." and so on. In order to solve the above-mentioned company, the brand name and the name of the name are identical to the same enterprise entity. This embodiment uses the PageRank diagram model, the BM25 algorithm, and the alias dictionary, and link to the same enterprise entity of the corporate knowledge base, and then The final enterprise entity corresponding to the target enterprise entity.
[0097] S130 determines the attribute set of the final enterprise entity, and collects the attribute set as a collection of candidate answers to the target enterprise entity;
[0098] After obtaining the final enterprise entity corresponding to the user entry, the target enterprise entity is obtained from the enterprise knowledge base, and all attributes of the corresponding target enterprise entity can be obtained, and the properties of the enterprise entity include, but are not limited to, one of the following properties. Several types:
[0099] The company type, corporate address, business scope, industry, registered capital, shareholder information, foreign investment and consistent actors.
[0100] S140, based on the MV-LSTM model, determine the matching value of the user input question and the attribute set, and the attribute corresponding to the maximum match value is the final answer of the user input question.
[0101]After searching for the attribute collection of the final enterprise entity, it needs to be determined which attribute is consistent with the user input. Here is the matching value of the user input question and attribute, and then the maximum match value is highest. The attribute is the final answer to the user input question.
[0102] Specifically, the MV-LSTM model is used during the matching calculation, and the MV-LSTM model is a multi-semant model based on the bidirectional LSTM network. The two-way LSTM is used to process two sentences, and then the LSTM hidden layer output is used to calculate matching value. To confirm the properties of the target enterprise entity in the user input question, you can think that this is a MULTI-View (MV) process, which can examine the meaning of each word in different contexts.
[0103] Specifically, the training processes of the MV-LSTM model include: training data, training data, and training data, industry classification, business scope, registered capital, executives, legal persons, and foreign investment. Each training sample includes a question and an attribute, while the negative sample of training data is obtained by random sampling. Hinge loss function (Hinge Loss):
[0104]
[0105] among them Match score for the right sample, Match score for negative samples.
[0106] After the model prediction phase, after the words and candidate properties are embedded, first get the hidden layer representation of each word, ask the hidden layer of the candidate attribute two times indicating the Cosine function, the double linear function, and the Tensor function calculation. After getting three similarity matrices, then enter K-Max Pooling (Pihua) and MLP (all connected neural networks, multilayer perceptron), and finally after a linear conversion got the final match score, the maximum candidate attribute with the score is used as a candidate property. The final answer of question.
[0107] After determining the question of the final enterprise entity and the most matching attribute of the corporate knowledge base, the most matching entity property is directly returned to the user as a question.
[0108] Specifically, based on the MV-LSTM model, it is determined that the user input is a matching value of each attribute in the attribute set, and the attribute corresponding to the maximum match is determined as the final answer of the user input question, including:
[0109] The user input question and the attribute set respectively perform EmbedDing to quantization;
[0110] When the quantized user input question and attribute set input into the two-way LSTM network structure, obtain the hidden layer output of each time T and the two-way hidden layer output of the time is given
[0111] The similarity value between the two time steps of the corresponding two time steps of the corresponding hidden layer output of the COSINE function, the double linear function, and the TENSOR function, respectively, the similar values ​​calculated by the three similarity functions constitute three similar matrices. M m×n.
[0112] The similar matrix M extracts the similar matrix M using K-MAX POOLING for the three similar matrices. m×n Key matching feature;
[0113] The key matching feature of the extracted three similarity matrices uses a full connection layer connection and linear conversion calculation score, and the score of the calculation is determined as the user input question and the matching value of each attribute in the attribute set. ;
[0114] The attribute corresponding to the maximum match value is the final answer of the user input question.
[0115] In this embodiment, the intelligent question and answer method based on enterprise-based entity is adopted by enterprise knowledge base and deep learning technology. Enterprise Knowledge Library will provide a basis for the question of question, and deep learning technology is provided in question understanding and answering phase. Powerful technical guarantees, can play an important role in solving the question and answer scene around the enterprise entity. The present invention uses the physical recognition technology based on the BILSTM-CRF model in the problem, and the abbreviation and brand name of the enterprise entity can be accurately identified. The MV-LSTM model is used in the matching process. The model can ask the core issue and the corporate entity property in the enterprise knowledge base, which greatly enhances the accuracy of the problematic answer.

Example Embodiment

[0116] Example 2
[0117] See figure 2 , figure 2 A second flow chart of an intelligent question and answer method according to an embodiment of the present invention, which is applied to the server, and in the method according to the method of the embodiment, based on the embodiment, based on enterprise entity For the correspondence between the corporate knowledge base, the final enterprise entity corresponding to the target enterprise entity, including:
[0118] S2201, entity disambiguation processing of the target enterprise entity, based on the correspondence between enterprise entities and corporate knowledge base, and obtains the first candidate enterprise entity corresponding to the target enterprise entity;
[0119] For example, it is assumed that the target enterprise entity is Baidu, and the corresponding first candidate enterprise entity set can be {Baidu cloud, Baidu online network technology, Baidu translation, Baidu music}.
[0120] S2202, determine how popular each enterprise entity in the first candidate enterprise entity;
[0121] Specifically, based on the PAGERANK diagram model, calculate the popularity of each enterprise entity in the candidate enterprise entity, the proportion of investment relationships between enterprises in the enterprise's entity is used as the weight of the investment side in the figure, and each enterprise entity is initialized. The flow is a default value of a certain equality, such as 1 or 0.25, and gets the popularity of each enterprise entity after the PageRank algorithm is stabilized, and the calculated intersity is stored in advance in the corporate knowledge base.
[0122] S2203, determine the similarity of each enterprise entity and the target enterprise entity in the first candidate enterprise entity collection, and the similarity of the similarity is large to small, determine K enterprise entities that are large in the similarity result as the second Candidate enterprise entity collection;
[0123] Specifically, the BM25 algorithm is used to calculate the literal similarity of each enterprise entity in the target enterprise entity and the first candidate enterprise entity collection, and sequencing the literally similarity, the highest-to-peer entity, the highest-to-peer entity, as the second of the target enterprise entity Candidate enterprise entity collection.
[0124] S2204, the similarity of each of the enterprise entities in the second candidate enterprise entity is horizontally weighted to determine the ultimate enterprise entity corresponding to the target enterprise entity.
[0125] Then, the forwardness of the pre-calculated entities and the value of the literal similarity value linear weighting calculated as the final similarity value of the candidate enterprise entity and the target enterprise entity in the second candidate enterprise entity collection, and the K enterprise entity is calculated. The highest value of the resulting final similarity value corresponds to the final enterprise entity of the target enterprise entity.
[0126] It should be noted that the present embodiment does not limit the order of the calculation similarity and the interaction, that is, the process of the entity disambiguation processing and the physical link by the PAGERANK diagram model and the BM25 algorithm are not as defined, that is, the embodiment. In the case of the correspondence between enterprise entities and corporate knowledge base, the final enterprise entity corresponding to the target enterprise entity can also be specifically included:
[0127] Based on the correspondence between enterprise entity and corporate knowledge base, the collection of candidate enterprise entities corresponding to the target enterprise entity is obtained, using the BM25 algorithm to form a target enterprise entity in the full name of enterprise entities within the candidate enterprise entity, obtain a company The similarity of the entity and the target enterprise entity, and based on the similarity of the similar degree, the K candidate enterprise entity is selected, and then the K pythrough of candidate corporate entities is calculated based on the PageRank diagram model, and the PageRank diagram model is adopted. The ratio of investment relationship between enterprise entities in corporate knowledge base is the weight of investment in the figure. After the PageRank diagram is converged, the importance ranking value of each enterprise entity in the enterprise knowledge base is used as the popularity of each enterprise entity. The similarity calculated by the BM25 algorithm and the popularity based on the PageRANK map model are used to calculate a new value, and finally the enterprise entity corresponding to the highest value corresponds to the final enterprise entity corresponding to the target enterprise entity.
[0128] In this embodiment, the intelligent question and answer method based on the enterprise entity is used, which uses the PageRank diagram model, which enables the referriticity of the enterprise entity and the brand name to accurately link the corresponding enterprise entities within the corporate knowledge base, further improve corporate knowledge. The accuracy of the question and answer.

Example Embodiment

[0129] Example three
[0130] See image 3 , image 3 A third flow chart of an intelligent question and answer method according to an embodiment of the present invention, which is applied to the server, and on the basis of the method described in Example 2, entity is carried out on the target enterprise entity. Before the discrimination, the present embodiment also includes:
[0131] S30A, using a fully matched method to determine if there is a business entity that matches the target enterprise entity in the enterprise knowledge base. If so, the S30C is executed, the matching enterprise entity is used as the target enterprise entity. Final business entity, if no, execute S30B;
[0132] Specifically, first link the entity in the target enterprise entity and enterprise knowledge base, which is a fully matched method. Since enterprise companies have uniqueness, if you can retrieve corporate entities in the corresponding corporate knowledge base, you can map the only business entity in the retrieved corporate knowledge base; this process is based on enterprise knowledge base. The enterprise entity is full, so this step may also be called entity full name matching. In order to speed up the retrieval matching speed, you can presets the enterprise-wide enterprise entities in the REDIS library and the Elastic Search library, so that the retrieval matching speed can be greatly improved in engineering practice. If there is a business entity that fully matches the target enterprise entity in the corporate knowledge base, the S30C is executed, and the corresponding enterprise entity is performed as the final enterprise entity corresponding to the target enterprise entity, and if there is no full match with the company, the S30B is executed.
[0133] S30b, according to the correspondence between the brief or brand name and the company, it is judged whether or not the enterprise's knowledge base has a corporate entity corresponding to the target enterprise entity. If so, the corresponding enterprise entity is executed as the corresponding enterprise entity The final enterprise entity corresponding to the target enterprise entity; if not, S2201 is executed, the target enterprise entity is subjected to entity disambiguation steps.
[0134]In order to enhance the accuracy of the target enterprise entity and corporate knowledge base, the enterprise knowledge base also includes the briefing and brand name matching dictionary. The matching dictionary here is mainly collected by manual way, including the company's referred to as the brand name and the company's full name. It is the key value of the company's abbreviation or brand name to the corporate knowledge base, including common Internet companies, financial technology companies and other companies referred to as entities in the corporate knowledge base. In this embodiment, the matching word can be placed in the REDIS library and the ELASTIC Search library accelerates the retrieval speed. Then, according to the relevant relationship between the referral or brand name and the company, it is judged whether or not the enterprise's knowledge base has a corporate entity corresponding to the target enterprise entity. If yes, execute S30C, if no, execute S2201.
[0135] In this embodiment, the intelligent question and answer method of the enterprise entity is used. It is used to retrieve the corporate entity corresponding to the target enterprise entity. The pre-stored enterprise entity is fully reproduced, and the speed of retrieval matching, which enhances the speed of corporate knowledge, corporate knowledge The library pre-stores the matching dictionary of the brief and brand name, improving the accuracy of the linkage of the target enterprise entity and corporate knowledge base, further improving the accuracy of the company's knowledge question and answer.

PUM

no PUM

Description & Claims & Application Information

We can also present the details of the Description, Claims and Application information to help users get a comprehensive understanding of the technical details of the patent, such as background art, summary of invention, brief description of drawings, description of embodiments, and other original content. On the other hand, users can also determine the specific scope of protection of the technology through the list of claims; as well as understand the changes in the life cycle of the technology with the presentation of the patent timeline. Login to view more.

Similar technology patents

Network for formance measuring method

InactiveCN101026504Aimprove accuracy
Owner:HUAWEI TECH CO LTD

Intelligent protocol parsing method and device

InactiveCN101035111Aimprove accuracyHigh protocol recognition efficiency
Owner:BEIJING VENUS INFORMATION TECH

TR309 - portable otoscope video viewer

InactiveUS20050171399A1easily attainableimprove accuracy
Owner:RICH TONY C +1

Communication terminal apparatus and communication system

InactiveUS20060003776A1fluctuation of be improveimprove accuracy
Owner:SONY ERICSSON MOBILE COMM JAPAN INC

Classification and recommendation of technical efficacy words

  • improve accuracy

Golf club head with adjustable vibration-absorbing capacity

InactiveUS20050277485A1improve grip comfortimprove accuracy
Owner:FUSHENG IND CO LTD

Direct fabrication of aligners for arch expansion

ActiveUS20170007366A1improve accuracyimproved strength , accuracy
Owner:ALIGN TECH

Stent delivery system with securement and deployment accuracy

ActiveUS7473271B2improve accuracyreduces occurrence and/or severity
Owner:BOSTON SCI SCIMED INC

Method for improving an HS-DSCH transport format allocation

InactiveUS20060089104A1improve accuracyincrease benefit
Owner:NOKIA SOLUTIONS & NETWORKS OY

Catheter systems

ActiveUS20120059255A1increase selectivityimprove accuracy
Owner:ST JUDE MEDICAL ATRIAL FIBRILLATION DIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products