CN109460462B - Chinese similarity problem generation system and method - Google Patents
Chinese similarity problem generation system and method Download PDFInfo
- Publication number
- CN109460462B CN109460462B CN201811360413.5A CN201811360413A CN109460462B CN 109460462 B CN109460462 B CN 109460462B CN 201811360413 A CN201811360413 A CN 201811360413A CN 109460462 B CN109460462 B CN 109460462B
- Authority
- CN
- China
- Prior art keywords
- sales
- questions
- common
- chinese
- similarity
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Electrically Operated Instructional Devices (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention belongs to the technical field of natural language processing, and particularly relates to a system and a method for generating Chinese similarity problems. The invention provides a new Chinese similar question generation system and a method, the Chinese similar question generation system and the method can position corresponding answers in a semantic knowledge base according to similarity, and sequence the answers according to the relevance of each sales mark, so as to answer a user, and the positioned answers of the same common question with the sales mark attribute in the semantic knowledge base may be more than one, so that the answer can be more intelligently guided instead of simple question-answer, and benefits can be obtained among a platform, a merchant and a client after the question-answer.
Description
Technical Field
The invention belongs to the technical field of natural language processing, and particularly relates to a system and a method for generating Chinese similarity problems.
Background
The prior art provides a system and a method for generating similar problems, which can improve the matching degree and rationality of generated question sentences and original problems by adopting a mode of combining rules and statistics; however, in business, the nature of question answering is not in the accuracy of answers, and such an accurate one-to-one answer is not necessarily desirable to consumers.
Disclosure of Invention
Aiming at the problems, the invention provides a new system and a method for generating the Chinese similar questions, which can more intelligently answer the questions proposed by the user and effectively ensure the benefits among the platform, the merchants and the user.
The specific technical scheme of the invention is as follows:
the invention provides a Chinese similarity problem generation method, which comprises the following steps:
s1: establishing a first mapping between keywords and common questions, establishing a second mapping between the common questions and sales marks, and establishing a third mapping between the common questions subjected to the sales marks and answers to the questions, wherein the sales marks are generated based on training or rules;
s2: a keyword extraction step, wherein a plurality of merchant terminals establish voice channels to a plurality of user terminals belonging to respective merchants through a voice interaction server, users propose natural language questions to corresponding merchant terminals through the voice channels, and after the voice interaction server identifies the questions proposed by the users, the keywords matched in the Chinese semantic knowledge base are obtained after voice-to-word processing and natural language keyword extraction processing;
s3: a common problem matching step, namely matching the keywords extracted in the step S2 to a common problem in a Chinese semantic knowledge base based on a first algorithm, judging whether the common problem has a sales mark attribute, if the common problem has the sales mark attribute, entering the step S4, and if the common problem does not have the sales mark attribute, entering the step S5 and putting the common problem into a sales mark list to be marked;
s4: a first similar question answer generating step, namely generating a plurality of first similar question answers which are ranked based on relevance and similarity by using the matched common questions based on a second algorithm;
s5: a second similar question answer generating step, namely generating a plurality of second similar question answers which are ranked based on the similarity degree by the matched common questions based on a third algorithm;
s6: and the merchant terminal sends the first similar question answer or the second similar question answer which is subjected to voice synthesis processing by the voice interaction server to the client terminal, so that automatic voice interaction between the merchant terminal and the client terminal is realized.
A chinese similarity problem generation system, the generation system comprising:
the Chinese semantic knowledge base building module is used for building a Chinese semantic knowledge base, building a first mapping between key words and common questions, building a second mapping between the common questions and sales marks, building a third mapping between the common questions subjected to the sales marks and answers to the questions, and generating the sales marks based on training or rules;
the keyword extraction module is used for establishing a voice channel between a plurality of merchants and a plurality of user sides belonging to the respective merchants through a voice interaction server, proposing a natural language problem to the corresponding merchant through the voice channel by a user, and obtaining keywords matched in the Chinese semantic knowledge base after the voice interaction server identifies the problem posed by the user, and then the problems are subjected to voice-to-word processing and natural language keyword extraction processing;
the common question matching module is used for matching the keywords extracted from the keyword extraction module to common questions in a Chinese semantic knowledge base based on a first algorithm, judging whether the common questions have sales mark attributes, if the common questions have the sales mark attributes, entering the first similar question answer generation module, and if the common questions do not have the sales mark attributes, entering the second similar question answer generation module and putting the common questions into a to-be-marked sales mark list;
the first similar question answer generating module is used for generating a plurality of first similar question answers which are based on the relevance and similarity ranking for the matched common questions based on a second algorithm;
the second similar question answer generating module is used for generating a plurality of second similar question answers which are ranked based on the similarity degree from the matched common questions based on a third algorithm;
and the interaction module is used for sending the first similar question answer or the second similar question answer which is subjected to voice synthesis processing by the voice interaction server to the client by the merchant so as to realize automatic voice interaction between the merchant and the client.
The invention has the following beneficial effects:
the invention provides a new Chinese similar question generation system and a method, the Chinese similar question generation system and the method can position corresponding answers in a semantic knowledge base according to similarity, and sequence the answers according to the relevance of each sales mark, so as to answer a user, and the positioned answers of the same common question with the sales mark attribute in the semantic knowledge base may be more than one, so that the answer can be more intelligently guided instead of simple question-answer, and benefits can be obtained among a platform, a merchant and a client after the question-answer.
Drawings
FIG. 1 is a flowchart of a method for generating a Chinese similarity problem according to embodiment 1;
FIG. 2 is a flowchart of step S4 in example 1;
fig. 3 is a block diagram showing the structure of the system for generating chinese similarity problems according to embodiment 2.
Detailed Description
The present invention will be described in further detail with reference to the following examples and drawings.
The steps illustrated in the flow charts of the figures may be performed in a computer system such as a set of computer-executable instructions. Although a logical order is shown in the flow diagrams, in some cases, the steps shown or described may be performed in an order different than presented herein.
Example 1
An embodiment 1 of the present invention provides a system and a method for generating a chinese similarity problem, as shown in fig. 1, the method includes:
s1: establishing a first mapping between keywords and common questions, establishing a second mapping between the common questions and sales marks, and establishing a third mapping between the common questions subjected to the sales marks and answers to the questions, wherein the sales marks are generated based on training or rules;
s2: a keyword extraction step, wherein a plurality of merchant terminals establish voice channels to a plurality of user terminals belonging to respective merchants through a voice interaction server, users propose natural language questions to corresponding merchant terminals through the voice channels, and after the voice interaction server identifies the questions proposed by the users, the keywords matched in the Chinese semantic knowledge base are obtained after voice-to-word processing and natural language keyword extraction processing;
s3: a common problem matching step, namely matching the keywords extracted in the step S2 to a common problem in a Chinese semantic knowledge base based on a first algorithm, judging whether the common problem has a sales mark attribute, if the common problem has the sales mark attribute, entering the step S4, and if the common problem does not have the sales mark attribute, entering the step S5 and putting the common problem into a sales mark list to be marked;
s4: a first similar question answer generating step, namely generating a plurality of first similar question answers which are ranked based on relevance and similarity by using the matched common questions based on a second algorithm;
s5: a second similar question answer generating step, namely generating a plurality of second similar question answers which are ranked based on the similarity degree by the matched common questions based on a third algorithm;
s6: and the merchant terminal sends the first similar question answer or the second similar question answer which is subjected to voice synthesis processing by the voice interaction server to the client terminal, so that automatic voice interaction between the merchant terminal and the client terminal is realized.
The invention provides a new Chinese similar question generation method, which can position corresponding answers in a semantic knowledge base according to similarity and sequence the answers according to the relevance of each sales mark, so as to answer a user, and the positioned answers of the same common question with the sales mark attribute in the semantic knowledge base can be more than one, so that the answer can be more intelligently guided instead of simple question-answer, and benefits can be obtained among a platform, a merchant and a customer after the question-answer.
In step S1, a plurality of mapping relationships may be used between the keywords and the common questions in the chinese semantic knowledge base, and between the common questions and the answers to the questions via the sales marks, and preferably, a one-to-one mapping relationship is used between the common questions and the sales marks. The mapping relationship between the common question and the sales mark is established, or the mapping relationship between the answer to the question and the sales mark is established, which is a routine choice for those skilled in the art.
In step S1, sales tags may be represented by letters or numbers, such as 4 categories, which are classified into A, B, C, D. The rule is as follows: according to the sales expert knowledge of the merchant, for example, according to the purchasing stage of the potential customer, different common problem sales marks are respectively endowed for 'purchase at once', 'demand', 'purchase desire' and 'potential desire', the judgment mode includes but is not limited to that sales categories are classified according to crowd interest point sentences, competition sentences, price inquiring sentences and brand sentences, the expert knowledge is not understood to limit the protection scope of the invention, and other classification modes capable of improving the sales conversion Rate (ROI) can be adopted. If the 'crowd interest point statement' is marked as the class A, the 'competitive product statement' is marked as the class B, the 'price inquiry statement' is marked as the class C, and the 'brand statement' is marked as the class D. When the data volume is large, such as exceeding 10000, the data can be generated based on training, the training method includes known machine learning algorithm, such as neural network algorithm and markov algorithm, and it is within the protection of the present invention to select other AI algorithm training sales tags based on the concept of the present invention.
In step S4 of this embodiment, the number of the matched common questions with sales marks is 2 or more than 2, and in an actual question and answer, when the user continuously presents a plurality of questions, or the user presents a question, but at least two questions are matched based on the first algorithm, the first similar question answers ranked based on the relevance and the similarity may be matched according to the steps S2-S5, or the second similar question answers ranked based on the similarity may be matched, so as to better satisfy the needs of the user, and facilitate the merchant to convert the "potentially willing" consumer into a "buying right away" consumer, and meanwhile, the voice interaction between the user and the merchant may bring greater benefit to the platform, so the method may better guarantee the benefit among the platform, the merchant and the user. In the above example, the sales tag is assigned in the following manner when calculating the relevance, where the relevance of the a class is 0.1, the relevance of the B class is 0.2, the relevance of the C class is 0.3, and the relevance of the D class is 0.4, and the process of the system for processing the natural speech problem is shown in table 1 and table 2:
for example:
and when the input question does not have a sales mark, outputting answers through a third algorithm, wherein the output sequence is fresh, the whole market is lowest, and the output answer sequence is well adjusted according to the similarity.
TABLE 2
The relevance of each answer can also be calculated respectively, such as fresh (0.1), facing sun (0.1), with a store in the sea lake (0.1), delivery (0.4), with a broccoli in the Guangxi (0.2), with a promotion (0.4), and the order of the output answers at this time is delivery, with a promotion, with an orchid in the Guangxi, fresh, facing sun, with a store in the sea lake, and the ordering can be realized based on the concept of the invention; when the input question does not have a sales mark, the answer is output through a third algorithm, the output sequence is fresh, the store is in the morning sun, the store is in the sea lake, the goods are delivered, the orchid in the Guangxi is in the promotion, and the output answer sequence is adjusted according to the similarity.
As shown in fig. 2, in the present embodiment, step S4 includes the following steps:
s41: for common questions with sales mark attributes, generating a plurality of first similar question answers in a Chinese semantic knowledge base through a second algorithm based on similarity, wherein the second algorithm comprises but is not limited to calculating the similarity between the common questions and the question answers based on Word2Vec, and the calculation of the similarity by selecting other AI algorithms based on the concept of the invention is protected by the invention;
s42: and calculating the relevance of the common questions with the sales mark attributes, and ranking the generated answers of the first similar questions based on the relevance, wherein the relevance is calculated based on training or rules.
Since a common question may correspond to a plurality of answers to the question, at this time, the similarity between the common question and each answer to the question needs to be calculated through a second algorithm, the answer with the highest similarity value is taken as the alternative output answer to the common question, since the user proposes a plurality of questions with sales marks, the correlation of each sales mark needs to be calculated, the alternative output answers are ranked based on the value of the correlation to output the actual answer, and benefits among the platform, the merchant and the user can be greatly improved.
In step S42, the calculation based on rules includes, but is not limited to, assigning a degree of association to each sales mark in the chinese semantic knowledge base based on the sales expert knowledge of the merchant, the calculation based on training includes assigning a degree of association to each sales mark in the chinese semantic knowledge base, and the calculation model trains the sales marks assigned with the degree of association, where the calculation model is established based on an artificial intelligence deep learning technique; in connection with the above example, the calculation model is trained with the sales tags to which the scores are respectively assigned, and the trained calculation model can directly assign values to the sales tags; when the answer is output according to the relevance, the answer is more targeted, the requirements of the user are met, and the benefits among the platform, the merchant and the user are further ensured.
In this embodiment, the first algorithm in step S3 includes, but is not limited to, a recurrent neural network algorithm, and it is within the protection of the present invention to select other AI algorithms to calculate similarity based on the inventive concept; the third algorithm in step S5 includes, but is not limited to, a natural language processing algorithm.
Natural Language Processing (NLP) algorithms of the present invention include, but are not limited to, statistical-based Machine Learning (Machine Learning) and Deep Learning (Deep Learning) algorithms. The following NLP algorithm selection is within the concept of the present invention, and the classification algorithm may select LR (Logistic Regression, also called Logistic classification), SVM (Support Vector Machine), NB (Naive Bayes), DT (Decision Tree), integrated algorithm (such as 1), Bagging, 2), Random Forest, 3), GB (Gradient Boosting), 4) gbdt (Gradient Boosting Decision Tree), 5, AdaBoost, 6), Xgboost), maximum entropy model; the Regression algorithm can select LR (Linear Regression), SVR (support vector machine Regression), RR (Ridge Regression); the clustering algorithm can select a Knn algorithm, a Kmeans algorithm, hierarchical clustering and density clustering; the dimensionality reduction algorithm may select SGD (random gradient descent); the probability graph model algorithm can select a Bayesian network, an HMM and a CRF (conditional random field); the text mining algorithm may select a model (such as LDA (topic generation model), maximum entropy model), keyword extraction (such as 1, tf-idf, 2), bm25, 3, textrank, 4), pagerank, 5), left-right entropy high as a keyword, 6, mutual information), lexical analysis (such as 1), participle-HMM (due to markov) -CRF (conditional random field), 2), part-of-speech tagging, 3), named entity recognition), syntactic analysis (such as 1, syntactic structure analysis, 2, dependency analysis), text vectorization (such as 1, tf-idf, 2), word2vec, 3), doc2vec, 4), cw2vec, distance calculation (such as 1, euclidean distance, 2), similarity calculation; the optimization algorithm may select regularization (e.g., 1). L1 regularization, 2). L2 regularization); the deep learning algorithm may select BP, CNN, DNN, RNN, LSTM.
Example 2
A chinese similarity problem generation system, as shown in fig. 3, the generation system comprising:
the Chinese semantic knowledge base building module 1 is used for building a Chinese semantic knowledge base, building a first mapping between key words and common questions, building a second mapping between the common questions and sales marks, and building a third mapping between the common questions subjected to the sales marks and answers to the questions, wherein the sales marks are generated based on training or rules;
the keyword extraction module 2 is used for establishing a voice channel between a plurality of merchants and a plurality of user terminals belonging to the respective merchants through a voice interaction server, proposing a natural language question to the corresponding merchant terminal through the voice channel by a user, and obtaining keywords matched in the Chinese semantic knowledge base after the voice interaction server identifies the question posed by the user, and then the speech-to-word processing and the natural language keyword extraction processing are carried out;
the common question matching module 3 is used for matching the keywords extracted in the keyword extraction module 2 to common questions in a Chinese semantic knowledge base based on a first algorithm, judging whether the common questions have sales mark attributes, if the common questions have the sales mark attributes, entering the first similar question answer generation module 4, and if the common questions do not have the sales mark attributes, entering the second similar question answer generation module 5 and putting the common questions into a sales mark list to be marked;
the first similar question answer generating module 4 is used for generating a plurality of first similar question answers which are ranked based on the relevance and the similarity from the matched common questions based on a second algorithm;
the second similar question answer generating module 5 is used for generating a plurality of second similar question answers which are ranked based on the similarity degree from the matched common questions based on a third algorithm;
and the interaction module 6 is used for the merchant end to send the first similar question answer or the second similar question answer which is subjected to the voice synthesis processing by the voice interaction server to the client end, so that the automatic voice interaction between the merchant end and the client end is realized.
In this embodiment, in the chinese semantic knowledge base building module 1, the rules include, but are not limited to, expert knowledge based on sales of merchants, and may be generated based on training when the data size is large, and the training method includes a known machine learning algorithm.
In this embodiment, a plurality of mapping relationships may be used between the keywords and the common questions and between the common questions and the answers to the questions marked by the sales tags in the chinese semantic knowledge base constructing module 1, and the mapping relationships between the common questions and the sales tags are preferably in a one-to-one correspondence.
The invention provides a new Chinese similar question generation system, which can position corresponding answers in a semantic knowledge base according to similarity and sequence the answers according to the relevance of each sales mark, so as to answer a user, and more than one answer can be positioned in the semantic knowledge base for the same common question with the sales mark attribute, so that the answer can be more intelligently guided instead of simple question-answer, and the benefit can be obtained among a platform, a merchant and a customer after the question-answer.
Since the method description of the invention is implemented in a computer system. The computer system may be provided in a processor of a server or a client, for example. For example, the methods described herein may be implemented as software executable with control logic that is executed by a CPU in a server. The functionality described herein may be implemented as a set of program instructions stored in a non-transitory tangible computer readable medium. When implemented in this manner, the computer program comprises a set of instructions which, when executed by a computer, cause the computer to perform a method capable of carrying out the functions described above. Programmable logic may be temporarily or permanently installed in a non-transitory tangible computer-readable medium, such as a read-only memory chip, computer memory, disk, or other storage medium. In addition to being implemented in software, the logic described herein may be embodied using a discrete component, an integrated circuit, programmable logic used in conjunction with a programmable logic device such as a Field Programmable Gate Array (FPGA) or microprocessor, or any other device including any combination thereof. All such implementations are intended to fall within the scope of the present invention.
The above-mentioned embodiments are merely illustrative of the preferred embodiments of the present invention, and do not limit the scope of the present invention, and various modifications and improvements of the technical solution of the present invention by those skilled in the art should fall within the protection scope defined by the claims of the present invention without departing from the spirit of the present invention.
Claims (10)
1. A Chinese similarity problem generation method is characterized by comprising the following steps:
s1: establishing a first mapping between keywords and common questions, establishing a second mapping between the common questions and sales marks, and establishing a third mapping between the common questions subjected to the sales marks and answers to the questions, wherein the sales marks are generated based on training or rules;
s2: a keyword extraction step, wherein a plurality of merchant terminals establish voice channels to a plurality of user terminals belonging to respective merchants through a voice interaction server, users propose natural language questions to corresponding merchant terminals through the voice channels, and after the voice interaction server identifies the questions proposed by the users, the keywords matched in the Chinese semantic knowledge base are obtained after voice-to-word processing and natural language keyword extraction processing;
s3: a common problem matching step, namely matching the keywords extracted in the step S2 to a common problem in a Chinese semantic knowledge base based on a first algorithm, judging whether the common problem has a sales mark attribute, if the common problem has the sales mark attribute, entering the step S4, and if the common problem does not have the sales mark attribute, entering the step S5 and putting the common problem into a sales mark list to be marked;
s4: a first similar question answer generating step, namely generating a plurality of first similar question answers which are ranked based on relevance and similarity by using the matched common questions based on a second algorithm;
s5: a second similar question answer generating step, namely generating a plurality of second similar question answers which are ranked based on the similarity degree by the matched common questions based on a third algorithm;
s6: and the merchant terminal sends the first similar question answer or the second similar question answer which is subjected to voice synthesis processing by the voice interaction server to the client terminal, so that automatic voice interaction between the merchant terminal and the client terminal is realized.
2. The method for generating Chinese similarity questions according to claim 1, wherein the Chinese semantic knowledge base of step S1 includes a plurality of mapping relationships between the keywords and the common questions, the common questions marked for sale and the answers to the questions, and the common questions and the sales marks are in one-to-one correspondence.
3. The chinese similarity problem generation method according to claim 1, wherein in step S1, the rules include but are not limited to sales expert knowledge according to merchants, and are generated based on training when the data volume exceeds 10000 pieces, and the training method includes known machine learning algorithm.
4. The method for generating chinese similarity questions according to claim 1, wherein the number of matched common questions with sales tags in step S4 is at least 2.
5. The method for generating chinese similarity problem according to claim 1, wherein the step S4 includes the steps of:
s41: for common questions with sales mark attributes, generating a plurality of first similar question answers in a Chinese semantic knowledge base through a second algorithm based on similarity, wherein the second algorithm comprises but is not limited to calculating the similarity between the common questions and the question answers based on Word2 Vec;
s42: and calculating the relevance of the common questions with the sales mark attributes, and ranking the generated answers of the first similar questions based on the relevance, wherein the relevance is calculated based on training or rules.
6. The method for generating Chinese similarity problems according to claim 5, wherein the rule-based calculation in step S42 includes but is not limited to assigning a degree of association to each sales tag in the Chinese semantic knowledge base based on the sales expert knowledge of the merchant, the training-based calculation includes assigning a degree of association to each sales tag in the Chinese semantic knowledge base, and the calculation model trains the sales tags assigned with the degree of association, wherein the calculation model is established based on an artificial intelligence deep learning technique.
7. The method for generating Chinese similarity problem according to claim 1, wherein the first algorithm in step S3 includes but is not limited to recurrent neural network algorithm; the third algorithm in step S5 includes, but is not limited to, a natural language processing algorithm.
8. A system for generating chinese similarity problems, the system comprising: the Chinese semantic knowledge base building module (1) is used for building a Chinese semantic knowledge base, building a first mapping between key words and common questions, building a second mapping between the common questions and sales marks, building a third mapping between the common questions subjected to the sales marks and answers of the questions, and generating the sales marks based on training or rules;
the keyword extraction module (2) is used for establishing a voice channel between a plurality of merchant terminals and a plurality of user terminals belonging to respective merchants through a voice interaction server, and the users propose natural language questions to the corresponding merchant terminals through the voice channel;
the common question matching module (3) is used for matching the keywords extracted in the keyword extraction module (2) to common questions in a Chinese semantic knowledge base based on a first algorithm, judging whether the common questions have sales mark attributes, if the common questions have the sales mark attributes, entering the first similar question answer generation module (4), and if the common questions do not have the sales mark attributes, entering the second similar question answer generation module (5) and putting the common questions into a to-be-marked sales mark list;
the first similar question answer generating module (4) is used for generating a plurality of first similar question answers which are ranked based on the relevance and the similarity from the matched common questions based on a second algorithm;
the second similar question answer generating module (5) is used for generating a plurality of second similar question answers which are ranked based on the similarity degree by the matched common questions based on a third algorithm;
and the interaction module (6) is used for sending the first similar question answer or the second similar question answer which is subjected to voice synthesis processing by the voice interaction server to the client side by the merchant side, so that automatic voice interaction between the merchant side and the client side is realized.
9. The chinese similarity problem generation system according to claim 8, wherein in the chinese semantic knowledge base construction module (1), the rules include but are not limited to sales expert knowledge according to merchants, and are generated based on training when the data volume exceeds 10000, and the training method includes known machine learning algorithm.
10. The system for generating chinese similar questions according to claim 8, wherein the chinese semantic knowledge base of the chinese semantic knowledge base constructing module (1) employs a plurality of mapping relationships between the keywords and the common questions, and between the common questions and the answers to the questions via the sales marks, and the common questions and the sales marks employ a one-to-one mapping relationship.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811360413.5A CN109460462B (en) | 2018-11-15 | 2018-11-15 | Chinese similarity problem generation system and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811360413.5A CN109460462B (en) | 2018-11-15 | 2018-11-15 | Chinese similarity problem generation system and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109460462A CN109460462A (en) | 2019-03-12 |
| CN109460462B true CN109460462B (en) | 2021-10-19 |
Family
ID=65610673
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811360413.5A Active CN109460462B (en) | 2018-11-15 | 2018-11-15 | Chinese similarity problem generation system and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109460462B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110134798B (en) * | 2019-05-15 | 2021-07-27 | 八戒科技服务有限公司 | Innovation demand mining method, system, terminal and medium |
| CN110502620B (en) * | 2019-07-02 | 2023-05-30 | 平安科技（深圳）有限公司 | Method, system and computer equipment for generating guide diagnosis similar problem pairs |
| CN113051918B (en) * | 2019-12-26 | 2024-05-14 | 北京中科闻歌科技股份有限公司 | Named entity recognition method, device, equipment and medium based on ensemble learning |
| CN111753055B (en) * | 2020-06-28 | 2024-01-26 | 中国银行股份有限公司 | Automatic prompt method and device for customer questions and answers |
| CN112966071B (en) * | 2021-02-03 | 2023-09-08 | 北京奥鹏远程教育中心有限公司 | User feedback information analysis method, device, equipment and readable storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103150321A (en) * | 2011-12-06 | 2013-06-12 | 纽昂斯通讯公司 | Method and apparatus for operating a frequently asked questions (faq) - based system |
| CN104991887A (en) * | 2015-06-18 | 2015-10-21 | 北京京东尚科信息技术有限公司 | Information providing method and apparatus |
| CN105094315A (en) * | 2015-06-25 | 2015-11-25 | 百度在线网络技术（北京）有限公司 | Method and apparatus for smart man-machine chat based on artificial intelligence |
| CN107423437A (en) * | 2017-08-04 | 2017-12-01 | 逸途（北京）科技有限公司 | A kind of Question-Answering Model optimization method based on confrontation network intensified learning |
| CN107798140A (en) * | 2017-11-23 | 2018-03-13 | 北京神州泰岳软件股份有限公司 | A kind of conversational system construction method, semantic controlled answer method and device |
| CN107908803A (en) * | 2017-12-26 | 2018-04-13 | 上海智臻智能网络科技股份有限公司 | The response method and device, storage medium, terminal of question and answer interaction |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10275454B2 (en) * | 2014-10-13 | 2019-04-30 | International Business Machines Corporation | Identifying salient terms for passage justification in a question answering system |
-
2018
- 2018-11-15 CN CN201811360413.5A patent/CN109460462B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103150321A (en) * | 2011-12-06 | 2013-06-12 | 纽昂斯通讯公司 | Method and apparatus for operating a frequently asked questions (faq) - based system |
| CN104991887A (en) * | 2015-06-18 | 2015-10-21 | 北京京东尚科信息技术有限公司 | Information providing method and apparatus |
| CN105094315A (en) * | 2015-06-25 | 2015-11-25 | 百度在线网络技术（北京）有限公司 | Method and apparatus for smart man-machine chat based on artificial intelligence |
| CN107423437A (en) * | 2017-08-04 | 2017-12-01 | 逸途（北京）科技有限公司 | A kind of Question-Answering Model optimization method based on confrontation network intensified learning |
| CN107798140A (en) * | 2017-11-23 | 2018-03-13 | 北京神州泰岳软件股份有限公司 | A kind of conversational system construction method, semantic controlled answer method and device |
| CN107908803A (en) * | 2017-12-26 | 2018-04-13 | 上海智臻智能网络科技股份有限公司 | The response method and device, storage medium, terminal of question and answer interaction |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109460462A (en) | 2019-03-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109460462B (en) | Chinese similarity problem generation system and method | |
| CN110121706B (en) | Providing responses in a conversation | |
| JP7127150B2 (en) | Question and Answer Interaction Method, Apparatus, Computing Device and Computer Readable Storage Medium | |
| CN109635273B (en) | Text keyword extraction method, device, equipment and storage medium | |
| US11204972B2 (en) | Comprehensive search engine scoring and modeling of user relevance | |
| US9633045B2 (en) | Image ranking based on attribute correlation | |
| CN109271537B (en) | Text-to-image generation method and system based on distillation learning | |
| US8873813B2 (en) | Application of Z-webs and Z-factors to analytics, search engine, learning, recognition, natural language, and other utilities | |
| US20180158078A1 (en) | Computer device and method for predicting market demand of commodities | |
| Shirani et al. | Learning emphasis selection for written text in visual media from crowd-sourced label distributions | |
| CN109871446A (en) | Rejection method for identifying, electronic device and storage medium in intention assessment | |
| CN112100344A (en) | Financial field knowledge question-answering method based on knowledge graph | |
| CN107807968B (en) | Question answering device and method based on Bayesian network and storage medium | |
| US20230316379A1 (en) | Deep learning based visual compatibility prediction for bundle recommendations | |
| US10417328B2 (en) | Text quality evaluation methods and processes | |
| CN112307351A (en) | Model training and recommending method, device and equipment for user behavior | |
| CN110827797B (en) | Voice response event classification processing method and device | |
| US11860932B2 (en) | Scene graph embeddings using relative similarity supervision | |
| CN114491077A (en) | Text generation method, device, equipment and medium | |
| US20230073843A1 (en) | Data compatibility for text-enhanced visual retrieval | |
| KR102344075B1 (en) | Data mining-based online marketing platform | |
| US20210342744A1 (en) | Recommendation method and system and method and system for improving a machine learning system | |
| CN113157876A (en) | Information feedback method, device, terminal and storage medium | |
| CN116976920A (en) | Commodity shopping guide method and device, equipment and medium thereof | |
| CN113971599A (en) | Advertisement putting and selecting method and device, equipment, medium and product thereof |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |