Detailed Description
Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that the relative arrangement of the components and steps, numerical expressions and numerical values set forth in these embodiments do not limit the scope of the present invention unless it is specifically stated otherwise.
The following description of at least one exemplary embodiment is merely exemplary in nature and is in no way intended to limit the invention, its application, or uses.
Techniques, methods, and apparatus known to one of ordinary skill in the relevant art may not be discussed in detail, but are intended to be part of the specification where appropriate.
In all examples shown and discussed herein, any specific values should be construed as merely illustrative, and not a limitation. Thus, other examples of exemplary embodiments may have different values.
It should be noted that like reference numerals and letters refer to like items in the following figures, and thus once an item is defined in one figure, no further discussion thereof is necessary in subsequent figures.
< Hardware configuration >
Fig. 1 is a block diagram illustrating a hardware configuration of an information system 1000 in which an embodiment of the present invention may be implemented.
As shown in fig. 1, information system 1000 includes a server 1100, a client 1200, and a network 1300.
The server 1100 may be, for example, a blade server, a cloud server, a server group composed of a plurality of servers, or the like. In one example, the server 1100 may include a processor 1110, a memory 1120, an interface device 1130, a communication device 1140, a display device 1150, and an input device 1160, as shown in fig. 1. Although the server may also include speakers, microphones, etc., these components are not relevant to the present invention and are omitted here. The processor 1110 may be, for example, a central processing unit CPU, a microprocessor MCU, or the like. The memory 1120 includes, for example, ROM (read only memory), RAM (random access memory), nonvolatile memory such as a hard disk, and the like. The interface device 1130 includes, for example, a USB interface, a serial interface, and the like. The communication device 1140 can perform wired or wireless communication, for example. The display device 1150 is, for example, a liquid crystal display. The input device 1160 may include, for example, a touch screen, a keyboard, and the like.
The client 1200 may be a portable computer (1200-1), a desktop computer (1200-2), a cell phone (1200-3), a tablet computer (1200-4), etc. As shown in fig. 1, client 1200 may include a processor 1210, a memory 1220, an interface device 1230, a communication device 1240, a display device 1250, an input device 1260, a speaker 1270, a microphone 1280, and so forth. The processor 1210 may be a central processing unit CPU, a microprocessor MCU, or the like. The memory 1220 includes, for example, ROM (read only memory), RAM (random access memory), nonvolatile memory such as a hard disk, and the like. The interface device 1230 includes, for example, a USB interface, a headphone interface, and the like. The communication device 1240 can perform wired or wireless communication, for example. The display device 1250 is, for example, a liquid crystal display, a touch display, or the like. The input device 1260 may include, for example, a touch screen, a keyboard, and the like. A user may input/output voice information through the speaker 1270 and the microphone 1280.
The communication network 1300 may be a wireless network or a network, and may be a local area network or a wide area network. In the information system 1000 shown in FIG. 1, clients 1200-1, 1200-2, 1200-3, 1200-4 and a server 1100 may communicate over a communications network 1300.
The information system 1000 shown in fig. 1 is merely illustrative and is in no way intended to limit the invention, its application, or uses. In an embodiment of the present invention, the memory 1120 of the server 1100 is configured to store instructions for controlling the processor 1110 to operate to perform any one of the information acquisition methods provided by the embodiment of the present invention. In addition, the memory 1220 of the client 1200 is configured to store instructions for controlling the processor 1210 to operate to perform any of the information methods provided by the embodiments of the present invention. Those skilled in the art will appreciate that although a plurality of devices are shown in fig. 1 for both server 1100 and client 1200, the present invention may refer to only some of the devices, e.g., server 1100 may refer to only processor 1110 and storage device 1120, or client 1200 may refer to only processor 1210 and storage device 1220, etc. The skilled person can design instructions according to the disclosed solution. How the instructions control the processor to operate is well known in the art and will not be described in detail here.
The general conception of the implementation of the invention is to provide a new scheme for acquiring information, which can respectively acquire target query sentences input by a user and trigger contents which are associated with each recommended query sentence corresponding to the target query sentences in advance through a preset content trigger mode, then determine direct contents which correspond to the target query sentences and can accurately meet the information query requirements of the user according to the target query sentences and the trigger contents of each recommended query sentence, and display the direct contents in association with the target query sentences, so that the user can acquire the information which can accurately meet the information query requirements of the user without waiting for the information searching process and performing information browsing selection after inputting the target query sentences, thereby greatly shortening the information acquisition time of the user and reducing the information acquisition cost of the user.
< First embodiment >
In the present embodiment, an information acquisition method is provided. The information may include any content accessible over a local or wide area network, e.g., the information may include web content, pictures, audio, video, terms, encyclopedias, and the like.
The information obtaining method in this embodiment may be implemented by a server, where the server may include a blade server, a cloud server, or a server cluster. In one example, the information acquisition method in the present embodiment may be implemented by the server 1100 as shown in fig. 1.
As shown in fig. 2, the information acquisition method includes steps S2100 to S2300.
Step S2100, acquiring a recommended query statement set corresponding to the target query statement according to the target query statement input by the user.
The recommended query statement set comprises a plurality of recommended query statements corresponding to the target query statement. In this embodiment, the corresponding recommended query sentence may be obtained by performing semantic analysis, content association, and other processes on the target query sentence, for example, assuming that the sentence of the target query input by the user is "spring", and the corresponding recommended query sentence "spring has come", "spring flower", "poem of spring", and the like may be obtained by performing semantic analysis, content association, and other processes on the "spring". In this embodiment, a plurality of recommended query sentences in which the number of queried times in the latest statistical period is ranked in ascending order according to the number of queried times in the latest statistical period obtained after processing may be selected as recommended query sentences in the recommended query sentence set, so as to reduce the processing amount of subsequent steps and further improve the information acquisition efficiency.
Step S2200, acquiring triggering contents of each recommended query statement in the target query statement and the recommended query statement set respectively through a preset content triggering mode.
The trigger content is content that is pre-associated with the corresponding target query statement or recommended query statement.
The preset content triggering mode is a mode for acquiring triggering content of a target query sentence or a recommended query sentence, different from a general information searching mode, searching and matching in massive information of a wide area network or a local area network according to the query sentence to find a corresponding searching result, and the target query sentence or the recommended query sentence can be used as triggering to quickly and directly acquire triggering content which is pre-established in association with the triggered query sentence.
Through a preset content triggering mode, search matching in massive acquirable information is not needed, and content which is pre-associated with a target query sentence or a recommended query sentence is quickly and directly acquired, so that direct content which corresponds to the target query sentence and can accurately meet the information query requirement of a user is determined from the triggering content in combination with subsequent steps, and the user can immediately meet the information acquisition requirement of the user after inputting the target query sentence without waiting for an information searching process and carrying out information browsing selection.
In one example, the preset content trigger mode at least includes one of a vertical card trigger, a choice question-answer trigger, a official data trigger, and a vertical skill trigger.
The vertical card is an information card constructed by acquiring corresponding field information through means such as machine crawling or manual mining in different vertical subdivision fields. For example, for the "movie" field, the vertical subdivision may include multiple fields such as "comedy movie", "love movie", "ethical movie", "animation movie", "european and american movie", "chinese movie", etc., and for different vertical subdivision fields, corresponding contents may be obtained, and multiple vertical cards may be constructed, for example, "love movie" includes "roman holiday", and for "roman holiday" the vertical cards constructed include entity information such as movie name, movie introduction, movie director, etc. This vertical card, after construction, establishes an association with the movie name "roman holiday", movie director "odili hez book", etc.
Correspondingly, the triggering of the vertical card is to trigger a target query statement or a recommended query statement as a query statement, and acquire the vertical card associated with the query statement as a content triggering mode of corresponding triggering content.
The pick questions and answers may be question and answer content extracted by available knowledge-graph. The purpose of the knowledge graph is to describe information and concepts of entities present in the real world, and relationships present between entities. The currently available knowledge-graph generally covers a large amount of entity information of entities and associations between entities, and the entity information and the associations between entities are generally embodied in the form of question-answer contents, for example, the knowledge-graph includes an "why blue" corresponding accurate answer, and accordingly, the accurate answer is associated with the "why blue of sky. The question and answer contents are extracted from the acquired knowledge graph to serve as carefully chosen question and answer, and accordingly a large number of information contents which are related by taking the questions in the question and answer contents as query statement correspondence can be acquired.
Correspondingly, the choice question-answer triggering is to trigger a target query sentence or a recommended query sentence as a query sentence, and acquire an answer in choice question-answer data associated with the query sentence as a content triggering mode of corresponding triggering content.
The official network data is data posted by an official website corresponding to an entity. Entity information or entity concepts of their corresponding entities and associations between other entities are typically included in the data published by the corporate network. For example, the most accurate and detailed star data of a star is published in the official website of the star, and the association between the star data and the query statement related to the name of the star is correspondingly established. These network data for which a corresponding association has been established can be obtained from a plurality of official websites accessible and authenticated by machine crawling or manual mining.
Correspondingly, the triggering of the official data is to trigger a target query statement or a recommended query statement as a query statement, and acquire the official data associated with the query statement as a content triggering mode of corresponding triggering content.
Vertical skills are application skills developed for application requirements of different vertical subdivision domains. For example, aiming at express inquiry requirements, the developed express inquiry skills call the vertical skills, and the corresponding express logistics information can be inquired through the input express list numbers. Similarly, there are weather queries, sight spot queries, and the like. By invoking vertical skills, associated content can be obtained directly, quickly, accurately, etc.
Correspondingly, the vertical skill triggering is to trigger a target query statement or a recommended query statement as a query statement, call the corresponding vertical skill, and acquire the content associated with the query statement as a content triggering mode of the corresponding triggering content.
After step S2200, enter:
Step S2300, determining direct content corresponding to the target query statement according to the target query statement and the trigger content of each recommended query statement, so as to provide information expected to be acquired for the user by performing associated display on the direct content and the target query statement.
The trigger content of the target query statement and each recommended query statement is content which is associated with the target query statement or the recommended query statement in advance, and the direct content corresponding to the target query statement is determined from the trigger content, so that the information query requirement of a user can be accurately met, the user can immediately meet the information acquisition requirement after inputting the target query statement without waiting for an information searching process and carrying out information browsing selection, and the information acquisition cost is greatly reduced.
In one example, the step of determining the direct content corresponding to the target query statement based on the target query statement and the trigger content of each recommended query statement includes steps S2310-S2330.
In step S2310, the target query sentence and each recommended query sentence are respectively used as a query sentence, the confidence judgment is performed on the query sentence according to the trigger content of the query sentence, and the confidence query sentence determined by the confidence judgment is determined.
The trigger content of the target query statement and each recommended query statement is content which is pre-associated with the corresponding target query statement or recommended query statement. In some cases, although the triggering content can embody the information query intention of the corresponding query statement to a certain extent, the information query main requirement (actual and most important information query requirement) embodied by the corresponding query statement can be met to a certain extent, so in step S2310, the confidence judgment can be performed on the query statement according to the triggering content of the query statement, and the confidence query statement judged by the confidence judgment is the query statement of which the corresponding triggering content can meet the information query main requirement embodied by the query statement, so that the following steps are combined, the direct content which can more accurately meet the actual information query requirement of the user can be achieved, and the information acquisition efficiency of the user is improved.
In a more specific example, the step S2310 of performing confidence judgment using the target query term or the recommended query term as one query term may include steps S2311-S2313.
In step S2311, it is determined whether the query sentence is a high-frequency query sentence according to the historical query times of the query sentence.
The historical query times of the query statement are times of the query statement used by the user for querying information in a preset historical statistics period, and the historical query times can be obtained through analysis and statistics according to a log of an application providing information query service in the historical statistics period. The historical statistics period can be set according to specific application scenarios or application requirements, and is not limited by specific numerical values.
In this example, the high-frequency query threshold may be set according to engineering experience or experimental simulation, and the query statement whose historical query number is greater than the high-frequency query threshold may be determined as the high-frequency query statement.
According to the historical query times of the query sentences, whether the query sentences are high-frequency query sentences or not is distinguished, different confidence judgment strategies can be implemented for the high-frequency query sentences and the non-high-frequency sentences in combination with the subsequent steps, so that more accurate confidence judgment is realized, more accurate confidence query sentences are correspondingly obtained, and further the direct content of the actual information query requirement of a user can be more accurately met.
In step S2312, when the query sentence is determined to be a high-frequency query sentence, a click score of the triggering content of the query sentence is obtained, and when the click score meets the score confidence condition, the query sentence is determined to pass through the confidence judgment, so as to obtain a corresponding confidence query sentence.
The click score of the trigger content of the query statement is a score obtained according to the user click data meter obtained by the corresponding trigger content.
The historical query times of the high-frequency query sentences are higher, the probability that the corresponding trigger content is clicked by the user is higher, the high-frequency query sentences are subjected to confidence judgment through the corresponding click scores, and accurate confidence judgment results can be obtained.
In a more specific example, when the query term is determined to be a high frequency query term, the step of obtaining the click score of the trigger content of the query term includes:
s23121, obtaining user click data of the trigger content in the latest statistical period, and fitting the user click data according to a pre-constructed click scoring model to obtain click scores of the trigger content.
The statistical period is a period for counting user click data of the trigger content, and can be set according to a specific application scenario or application requirement, for example, the statistical period can be 1 week, 1 month, N days, or the like.
The user click data of the trigger content in the most recent statistical period is the overall data related to the user click that occurs on the piece of trigger content in the most recent statistical period. The user click data at least comprises the corresponding display times, click times, navigation click times, last click times and skip click times of the trigger content. The number of times the trigger content is presented is the total number of times the trigger content is presented after being queried. The number of clicks of the trigger content is the total number of clicks of the user after the trigger content is displayed by the query. The number of navigation clicks of the trigger content is the total number of times that the trigger content is triggered, and that happens when the user clicks only on the piece of trigger content and no longer clicks on other trigger content. The last click times of the trigger content are the total times of the trigger content after the trigger content is inquired and displayed and the user clicks other trigger content and then clicks the trigger content. The number of skip clicks of the trigger content is the total number of times that the user skips the piece of trigger content to click the trigger content after the trigger content is queried and displayed.
The click scoring model may adopt a training method related to an SVM (Support Vector Machine ), for example, a training method such as support vector regression, support vector clustering, etc., and a machine learning model obtained by training with a large number of collected samples is used for performing linear fitting with the display times, the click sequence, the navigation click times, the last click times, the skip click times, etc., included in the input user click data as the feature values input by the model, and outputting the corresponding click scoring.
And fitting the click data of the user through a click scoring model to obtain the click score of the corresponding trigger content, so that the user information acquisition requirement reflected by the trigger content can be accurately reflected.
The scoring confidence condition is a condition for judging whether the triggering content meets the information query main requirement embodied by the corresponding query statement according to the click score.
In this example, the scoring confidence condition is that the click score is above the click threshold and the difference between the click score and the historical highest click score is less than the click difference. The click threshold and the click difference value can be set according to engineering experience or experimental simulation aiming at a specific confidence judgment scene. The historical highest click score is the highest score of the click scores available in the historical scoring period. The historical scoring period may be set according to a specific application scenario or application requirement.
The trigger content with the click score higher than the click threshold value shows that the trigger content can meet the information query requirement represented by the corresponding query statement to a certain extent, on the basis, the difference value between the click score of the trigger content and the historical highest score is smaller than the click difference value, and the trigger content can accurately meet the information query requirement represented by the corresponding query statement with higher probability.
Step S2313, when the query sentence is determined not to be a high-frequency query sentence, determining a target requirement category and a target query mode corresponding to the query sentence according to the triggering content of the query sentence, and when the mode state of the target query mode under the target requirement category is determined to be a confidence according to the acquired mode state matching list, determining that the query sentence passes through confidence judgment to obtain a corresponding confidence query sentence.
The pattern state matching list comprises pattern states corresponding to a plurality of query patterns under each requirement category in a plurality of requirement categories. The mode state includes confidence and non-confidence.
In the present embodiment, the step of acquiring the pattern matching state list may include steps S2301 to S2304.
Step S2301, a history query statement set is acquired.
All query statements triggering a query in a historical statistics period are included in the set of historical query statements. The historical statistics period may be set according to statistics demand, for example, may be set to the past 1 year or the like.
Step S2302, classifying all query sentences included in the historical query sentence set according to the main requirement category corresponding to each query sentence in the historical query sentence set, to obtain a plurality of query sentences included under different requirement categories.
The main requirement category corresponding to each query statement is the requirement category which is most matched with the information query intention corresponding to the query statement, and is the requirement category which effectively reflects the real information query requirement of the user. The content type of the direct content corresponding to the query statement, the content type of the trigger content with the highest click score corresponding to the query statement and the like can be determined.
Each query sentence in the historical query sentence set has a corresponding main requirement category, the main requirement category of each query sentence can be used as a requirement category, a plurality of requirement categories can be obtained, the query sentences included in the historical query sentence set are classified according to the requirement categories, the query sentences with the same main requirement category are classified into one category, and a plurality of query sentences included in each requirement category can be obtained correspondingly.
Step S2303, randomly extracting a predetermined number of query sentences under each requirement category, extracting a corresponding query pattern for each query sentence in the predetermined number of query sentences, and obtaining the frequency of occurrence of a plurality of query patterns under each requirement category.
The predetermined number may be set according to a specific application scenario or application requirement. Assuming that there may be tens of thousands of query sentences under a demand category for a "movie" of the demand category, a predetermined number is 100, and 100 sentences can be randomly extracted from the tens of thousands of query sentences.
For each requirement category, for each of the randomly extracted query statements, a corresponding query pattern may be extracted according to the method of extracting a target query pattern from the target query statement hereinafter. After obtaining the query pattern of each query sentence, the number of query sentences with the same query pattern can be counted and used as the occurrence frequency of the query pattern, so as to obtain the occurrence frequency of a plurality of query patterns under the requirement category, and the like, so as to obtain the occurrence frequency of a plurality of query patterns under each requirement category.
Step S2304, setting mode states corresponding to the multiple query modes under each requirement category according to the frequency of occurrence of the multiple query modes under each requirement category, and correspondingly generating a mode matching state list.
In this example, after the frequencies of the multiple query patterns under each requirement category are obtained, the multiple query patterns may be sorted in descending order according to the frequencies of the multiple query patterns, and the query patterns with different sorting orders are correspondingly set to different pattern states.
For example, for the "movie" requirement category, the frequency of occurrence of the query pattern of the "downloading of the movie" is 5, the frequency of occurrence of the query pattern of the "director of the movie" is 20, the frequency of occurrence of the query pattern of the "downloading of the movie" after descending order is highest, the mode status of the query pattern of the "director of the movie" may be set as confidence, the frequency of occurrence of the query pattern of the "downloading of the movie" may be lowest, and the mode status of the query pattern of the "downloading of the movie" may be set as non-confidence.
The specific mode state setting process can be automatically set by a computer after setting a setting rule which can be automatically executed by the computer, or can be set by combining engineering experience through manually sorting results according to the occurrence frequency of a plurality of query modes under each requirement category, for example, the mode state of the query mode of 'movie playing' in the previous example can be set to be non-confidence by manually and directly setting the mode state of the query mode of 'movie playing'.
In a more specific example, when the query sentence is not a high-frequency query sentence, the step of determining the target requirement category and the target query pattern corresponding to the query sentence according to the trigger content of the query sentence includes S23131-S23133.
Step S23131, determining the target requirement category of the query statement according to the content type of the trigger content.
For example, the trigger content is a vertical card, the content type of the trigger content can be used as a target requirement category of the query statement, or the trigger content is an answer in a choice question and answer, the content type of the trigger content can be determined according to the classification in a knowledge graph of the source of the choice question and answer, the trigger content can be used as the target requirement category of the query statement, or the trigger content is a content acquired by calling a vertical skill, the content type of the trigger content can be determined according to the application scene classification described by the vertical skill, the target requirement category of the query statement, or the trigger content is official network data, the content type of the trigger content can be determined according to the data classification of the official network data, the target requirement category of the query statement, and the like.
Step S23132, when the trigger content belongs to the vertical card, acquiring an entity name according to entity information included in the trigger content, and taking the card type of the trigger content as an entity tag, when the trigger content does not belong to the vertical card, performing natural language processing on a query statement corresponding to the trigger content, acquiring the entity name according to a pre-acquired knowledge graph, and taking a category attribute with highest information heat corresponding to the entity name as the entity tag.
The vertical card is constructed for the content included in the corresponding content obtained in different vertical subdivision fields, and often has entity information of corresponding entities. For example, the card type is a vertical card for showing related information of the movie "no double", the displayed content is a brief introduction of the movie "no double", a staff table, a poster and the like, and the vertical card includes entity information of one entity of the movie "no double", and when the trigger content is the vertical card, the entity name "no double" can be obtained from the entity information of the vertical card.
The vertical card has a corresponding card type, which can be used as an entity tag. Taking the information card with the card type being a movie as an example, the corresponding obtained entity tag is a movie.
When the trigger content does not belong to the vertical card, the entity corresponding to the target query sentence can be mined according to a pre-acquired knowledge graph for describing entity information and the relation between the entities by technical means such as automatic word segmentation, syntactic analysis, natural language classification, information extraction and the like included in natural language processing (Natural Language Processing, NLP), the entity name is correspondingly obtained, and the category attribute of the entity with the highest information heat degree in the plurality of entities with the same entity name is used as an entity label, for example, for the entity name 'no double', the knowledge graph can have the entities with different types of attributes such as 'no double' of novels, no double 'of movies, no double' of television shows and the like, wherein the information heat degree of the movie 'no double' is the highest, and the corresponding entity label is a 'movie'.
Step S23133, replacing the entity label with the entity name in the query statement corresponding to the trigger content to obtain a corresponding target query mode.
Taking the example that the target query sentence is "no double director", assuming that the extracted entity tag is "movie", the entity name is "no double", after executing step S23133, it may be obtained that the target query pattern is "the director of movie".
When the query statement is not a high-frequency query statement, more user click data cannot be acquired, the target requirement category and the target query mode corresponding to the query statement are determined, the mode state of the target query mode under the target requirement category is determined according to the corresponding mode state matching list to carry out confidence judgment, and accurate confidence judgment results can be correspondingly acquired.
After determining the confidence query statement of the confidence judgment, enter:
step S2320, the relevance scores of the target query statement and each confidence query statement are obtained respectively.
The confidence query statement is a query statement whose corresponding trigger content can satisfy the information query main requirement embodied by the corresponding trigger content. The relevance score of the target query statement and each of the confidence query statements is an indicator reflecting the relevance between the target query statement and the corresponding confidence query statement. By acquiring the relevance scores of the target query statement and each confidence query statement, the method can combine the follow-up steps, can select the triggering content of the information main requirement of the actual full target query statement as the direct content, and effectively meets the information query requirement of the user.
In a more specific example, the step of separately obtaining the relevance scores of the target query statement and each of the confidence query statements includes:
Step S2321, the query characteristics of the target query statement, the recommended query characteristics of the confidence query statement and the correlation characteristics between the target query statement and the confidence query statement are used as model characteristics to be input into a correlation model, and the correlation scores of the target query statement and the confidence query statement output by the correlation model are obtained.
In this example, the query features of the target query statement are features related to the information query requirement represented by the target query statement, and at least include the number of queries of the corresponding query statement in the recommended scene of the latest statistics period, the number of clicks of the recommended query statement, the number of continuous inputs, the number of queries of the query scene, and the statement length. The specific time length of the recent statistics period may be set according to a specific application scenario or application requirement, which is not limited herein.
The recommended scene is a scene in which when a query statement is input, the corresponding recommended query statement is displayed for a user to click to select to replace the query statement for information query.
The number of times a query sentence is queried in a recommended scenario is the number of times the user inputs the query sentence in the recommended scenario.
The number of clicks of a query statement on a recommended query statement of a recommended scene is the total number of clicks of the recommended query statement of the query statement by a user in the recommended scene.
The number of times that the query sentence is continuously input in the recommended scene refers to the number of times that the user performs information query after inputting the query sentence continuously. For example, if the query term input by the user is "spring" and the continuation input is not ended for "day" thereafter, the number of times of the continuation input is considered to be increased by 1.
The query scene refers to a scene in which a user finishes information query after inputting a query sentence.
The number of times of inquiry sentences in an inquiry scene refers to the number of times that a user directly performs information inquiry without inputting the inquiry sentences after inputting the inquiry sentences.
The sentence length of a query sentence is the number of words or the number of characters of the query sentence.
The recommended query features of the recommended query statement are features related to information query requirements reflected by the recommended query statement, and at least comprise total click times of a recommended scene, query times of the query scene, main requirement satisfaction and statement length when the corresponding query statement is used as the recommended query statement in the latest statistical period.
The total clicking times of the recommended scenes of the recommended query sentences refer to the times of information query by clicking and selecting the query sentences instead of the input query sentences by a user when the query sentences are taken as the recommended query sentences.
The number of times of query in a query scenario in which a query term is recommended refers to the number of times of information query by the query term as an input query term.
The main requirement satisfaction degree of the recommended query statement refers to the degree that the triggering content of the recommended query statement corresponds to the information query requirement represented by the recommended query statement. According to the step of performing confidence judgment on the query statement according to the corresponding trigger content, the main requirement satisfaction degree of the recommended query statement judged by the confidence can be set to 1, otherwise, the main requirement satisfaction degree is set to 0.
The relevant features are features showing the correlation between the target query statement and the recommended query statement, and at least comprise the showing times, clicking times and content types of triggering content of the confidence query statement in the latest statistical period when the confidence query statement is used as the recommended query statement under the target query statement.
The relevance model is a machine learning model trained according to a gradient lifting decision tree-based algorithm. The gradient lifting decision tree algorithm, also called GBDT (Gradient Boosting Decision Tree), is an iterative decision tree algorithm, which consists of a plurality of decision trees, and the conclusions of all the trees are accumulated to be the final answer, so that the algorithm has stronger generalization capability (general ization) and can be used for model training in an information query scene to obtain better performance.
In this embodiment, the collected samples including the query feature, the recommended query feature and the related feature may be trained based on a gradient lifting decision tree algorithm to construct a correlation model, and by using the correlation model, the correlation strength between the target query statement and the confidence query statement may be accurately evaluated with respect to the query feature of the target query statement, the recommended query feature of the confidence query statement and the related feature therebetween. In this example, the higher the phototropism score output by the relevance model, the stronger the relevance between the target query statement and the confidence query statement is represented, and the more the confidence query statement can embody the information query intention actually reflected by the target query statement.
Through the above step S2321, the relevance score between the target query sentence and each confidence query sentence may be obtained respectively.
In another example, the step of separately obtaining the relevance scores of the target query statement and each of the confidence query statements includes steps S23201-S23202.
The method comprises 23201, respectively inputting query characteristics of a target query sentence, recommended query characteristics of a confidence query sentence and correlation characteristics between the target query sentence and the confidence query sentence into a first correlation model and a second correlation model as model characteristics, and obtaining a first correlation score of the target query sentence and the confidence query sentence output by the first correlation model and a second correlation score of the target query sentence and the confidence query sentence output by the second correlation model.
The query features of the target query statement, the recommended query features of the confidence query statement, and the relevant features between the target query statement and the confidence query statement are described in the above examples and are not described here again.
The first correlation model is a machine learning model trained based on a gradient boosting decision tree algorithm. The correlation model used in the above example is not described here again.
The second correlation model is a machine learning model trained based on a Wide & Deep learning algorithm. The Wide & Deep learning algorithm is a Deep learning algorithm based on which a Wide & Deep model is constructed. The Wide & Deep model is a model for classification and regression, and the core idea is to combine the memory capacity (memorization, i.e. the correlation between items or features is found from the historical data) of the linear model and the generalization capacity (general ization, the correlation between features is found from the historical data) of the DNN model, and optimize the parameters of 2 models simultaneously in the training process, so as to achieve the optimal prediction capacity of the whole model.
In this embodiment, the collected samples including the query feature, the recommended query feature, and the related feature may be trained based on a Wide & Deep learning algorithm, so as to construct and form a second correlation model.
Step S23202, the first relevance score and the second relevance score are multiplied by corresponding weights and summed to obtain a relevance score.
The weight corresponding to the first relevance score and the weight corresponding to the second relevance score can be set according to engineering experience values or experimental simulation data.
By multiplying the first relevance score output by the first relevance model and the second relevance score output by the second relevance model by corresponding weights and summing the obtained relevance scores, the model performances of the two relevance models can be integrated, and the relevance between the target query statement and the recommended query statement can be evaluated more accurately.
After obtaining the relevance score, enter:
Step S2330, determining the triggering content of the confidence query statement meeting the content direct condition as the direct content corresponding to the target query statement according to the relevance scores of the target query statement and each confidence query statement.
The content direct condition is that the relevance score between the confidence query statement and the target query statement is above a relevance threshold and the relevance score is highest. The correlation threshold is a threshold for judging whether the confidence query sentence is actually correlated with the target query sentence according to the correlation score, and can be set according to engineering experience values or experimental simulation data.
The confidence query statement is a query statement whose corresponding trigger content can satisfy the information query main requirement embodied by the corresponding trigger content. A confidence query statement is one that has a relevance score to the target query statement above a relevance threshold, meaning that the confidence query statement is actually relevant to the target query statement. The confidence query statement is the highest in correlation score with the target query statement, which means that the confidence query statement is most relevant to the target query statement and the real information query requirement represented by the target query statement can be represented most.
Correspondingly, the trigger content which can actually meet the real information query requirement reflected by the target query statement can be rapidly determined as the direct content according to the set content direct condition, so that the direct content can actually meet the corresponding information acquisition requirement.
< Information acquisition device >
In this embodiment, an information acquisition apparatus 3000 is further provided, as shown in fig. 3, including a recommendation acquisition unit 3100, a content trigger unit 3200, and a direct acquisition unit 3300, which are used to implement the information acquisition method provided in this embodiment, and are not described herein.
The recommendation obtaining unit 3100 is configured to obtain, according to a target query sentence input by a user, a recommendation query sentence set corresponding to the target query sentence.
The content triggering unit 3200 is configured to obtain, by using a preset content triggering manner, triggering content of each of the target query sentence and each of the recommended query sentence included in the recommended query sentence set, where the triggering content is content that is associated with the corresponding target query sentence or the recommended query sentence in advance.
Optionally, the preset content triggering modes at least include one of a vertical card triggering mode, a carefully chosen question-answering triggering mode, a official network data triggering mode and a vertical skill triggering mode.
The direct acquiring unit 3300 is configured to determine, according to the target query statement and the trigger content of each recommended query statement, direct content corresponding to the target query statement, so as to provide information that the user desires to acquire by displaying the direct content in association with the target query statement.
Optionally, the direct acquisition unit 3300 is further configured to implement the following steps:
Taking the target query statement and each recommended query statement as a query statement respectively, performing confidence judgment on the query statement according to the triggering content of the query statement, and determining a confidence query statement passing the confidence judgment;
Respectively obtaining the relevance scores of the target query statement and each confidence query statement;
Determining triggering content of the confidence query statement meeting the content direct condition as direct content corresponding to the target query statement according to the relevance scores of the target query statement and each confidence query statement;
The content direct condition is that the relevance score between the confidence query statement and the target query statement is above a relevance threshold and the relevance score is highest.
Optionally, the step of performing, by the direct obtaining unit 3300, confidence judgment on the query statement according to the trigger content of the query statement, to obtain a confidence query statement through confidence judgment includes:
determining whether the query sentence is a high-frequency query sentence or not according to the historical query times of the query sentence;
when the query statement is determined to be a high-frequency query statement, acquiring a click score of the triggering content of the query statement, and when the click score meets a score confidence condition, determining that the query statement passes confidence judgment to obtain the corresponding confidence query statement, wherein the score confidence condition is that the click score is higher than a click threshold value and the difference between the click score and a historical highest click score is smaller than a click difference value;
When the query statement is not a high-frequency query statement, determining a target requirement category and a target query mode corresponding to the query statement according to the triggering content of the query statement, and when the mode state of the target query mode under the target requirement category is determined to be confidence according to an acquired mode state matching list, determining that the query statement passes confidence judgment to obtain the corresponding confidence query statement, wherein the mode state matching list comprises mode states corresponding to a plurality of query modes under each requirement category in the plurality of requirement categories, and the mode states comprise confidence and non-confidence.
Optionally, when the direct obtaining unit 3300 is configured to determine that the query sentence is a high-frequency query sentence, the step of obtaining the click score of the trigger content of the query sentence includes:
Acquiring user click data of the trigger content in the latest statistical period, and fitting the user click data to obtain click scores of the trigger content according to a pre-constructed click score model, wherein the user click data at least comprises the display times, the click times, the navigation click times, the last click times and the skip click times of the trigger content;
And
When the query statement is not a high-frequency query statement, determining a target requirement category and a target query mode corresponding to the query statement according to the trigger content of the query statement comprises the following steps:
Determining a target demand category of the query statement according to the content type of the trigger content;
When the triggering content does not belong to the vertical card, carrying out natural language processing on the query statement corresponding to the triggering content, acquiring the entity name according to a pre-acquired knowledge graph, and taking the category attribute with highest information heat corresponding to the entity name as an entity label;
And replacing the entity name with the entity label in the query statement corresponding to the trigger content to obtain the corresponding target query mode.
Optionally, the step of directly obtaining, by the obtaining unit 3300, the relevance score of each confidence query term and the target query term respectively includes:
inputting the query characteristics of the target query statement, the recommended query characteristics of the confidence query statement and the correlation characteristics between the target query statement and the confidence query statement into a correlation model as model characteristics, and obtaining a correlation score of the target query statement and the confidence query statement output by the correlation model;
The method comprises the steps of selecting a target query statement, wherein the query feature at least comprises the query times of a recommended scene of the corresponding query statement in the latest statistics period, the click times of the recommended query statement, the continuous input times, the query times of the query scene and the statement length, the recommended query feature at least comprises the recommended scene total click times, the query times of the query scene, the main requirement satisfaction degree and the statement length when the corresponding query statement is used as the recommended query statement in the latest statistics period, and the related feature at least comprises the display times, the click times and the content types of the trigger contents of the confidence query statement in the latest statistics period when the confidence query statement is used as the recommended query statement;
Or alternatively
The step of obtaining the relevance scores of the target query statement and each confidence query statement respectively comprises the following steps:
Respectively inputting the query characteristics of the target query statement, the recommended query characteristics of the confidence query statement and the correlation characteristics between the target query statement and the confidence query statement into a first correlation model and a second correlation model as model characteristics, and acquiring a first correlation score of the target query statement and the confidence query statement output by the first correlation model and a second correlation score of the target query statement and the confidence query statement output by the second correlation model;
The method comprises the steps of selecting a target query sentence, wherein the query characteristics at least comprise the query times of a recommended scene of the corresponding query sentence in the latest statistics period, the click times of the recommended query sentence, the continuous input times, the query times of the query scene and the sentence length, the recommended query characteristics at least comprise the recommended scene total click times, the query times of the query scene, the main requirement satisfaction degree and the sentence length when the corresponding query sentence is used as the recommended query sentence in the latest statistics period, the related characteristics at least comprise the display times, the click times and the content types of the trigger contents of the confidence query sentence in the latest statistics period when the confidence query sentence is used as the recommended query sentence in the target query sentence, and the first relevance model is a machine learning model obtained by training based on a gradient lifting decision tree algorithm;
And multiplying the first correlation score and the second correlation score by corresponding weights respectively and summing the weights to obtain the correlation score.
It should be apparent to those skilled in the art that the information acquisition apparatus 3000 may be implemented in various ways. For example, the information acquisition apparatus 3000 may be implemented by an instruction configuration processor. For example, instructions may be stored in a ROM, and when the apparatus is started, the instructions are read from the ROM into a programmable device to realize the information acquisition apparatus 3000. For example, the information acquisition apparatus 3000 may be solidified into a dedicated device (for example, ASIC). The information acquisition apparatus 3000 may be divided into units independent of each other, or they may be realized by combining them together. The information acquisition device 3000 may be implemented by one of the above-described various implementations, or may be implemented by a combination of two or more of the above-described various implementations.
In the present embodiment, the information acquisition apparatus 3000 may be a server program that provides an information acquisition service, or may be a software development kit (for example, SDK) or the like packaged so as to implement a data acquisition method after being called.
< Information acquisition device >
In the present embodiment, there is also provided an information acquisition apparatus 4000, as shown in fig. 4, including:
a memory 4100 for storing executable instructions;
A processor 4200 for executing the information acquisition device 4000 to perform any one of the information acquisition methods as provided in the present embodiment according to the executable instructions.
In this embodiment, the information obtaining device 4000 may be any device having data management and processing functions, for example, may be a blade server, a cloud server, or a server group. In one example, the information acquisition apparatus 4000 may be a server 1100 as shown in fig. 1, further including a communication device, and the like.
< Readable storage Medium >
In the present embodiment, there is also provided a readable storage medium storing a computer program readable and executable by a computer for executing the information acquisition method described in the present embodiment when the computer program is read and executed by the computer.
The readable storage medium may be a tangible device that can hold and store instructions for use by an instruction execution device. The readable storage medium may be, for example, but not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples of a readable storage medium (a non-exhaustive list) include a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a Static Random Access Memory (SRAM), a portable compact disc read-only memory (CD-ROM), a Digital Versatile Disc (DVD), a memory stick, a floppy disk, a mechanical encoding device, punch cards or in-groove protrusion structures such as those having instructions stored thereon, and any suitable combination of the foregoing. A readable storage medium as used herein is not to be construed as a transitory signal itself, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (e.g., a pulse of light through a fiber optic cable), or an electrical signal transmitted through an electrical wire.
The first embodiment of the present invention has been described above with reference to the accompanying drawings, and according to this embodiment, an information acquisition method, apparatus, device, and readable storage medium are provided, in which trigger content that is pre-associated with a target query sentence input by a user and each recommended query sentence corresponding to the target query sentence is acquired respectively in a preset content trigger manner, then direct content that corresponds to the target query sentence and that can accurately satisfy information query requirements of the user is determined according to the target query sentence and the trigger content of each recommended query sentence, and the direct content is displayed in association with the target query sentence, so that the user can acquire information that accurately satisfies information query requirements of the user immediately without waiting for the process of information search and performing information browsing selection after inputting the target query sentence, thereby greatly shortening the time for the user to acquire information, and reducing the information acquisition cost of the user.
< Second embodiment >
In the present embodiment, an information acquisition method is provided. The information may include any content accessible over a local or wide area network, e.g., the information may include web content, pictures, audio, video, terms, encyclopedias, and the like.
The information acquisition method in this embodiment may be implemented by a client, where the client may include a mobile phone, a notebook computer, a desktop computer, a tablet computer, and the like. In one example, the information acquisition method in the present embodiment may be implemented by the client 1200 shown in fig. 1.
As shown in fig. 5, the information acquisition method includes steps S4100-S4300.
In step S4100, an input operation of the user is received, and a target query term corresponding to the input operation is determined.
In the present embodiment, the input operation by the user may include a text input operation, a voice input operation, and the like. For example, as shown in fig. 6, the input operation of the user may be received by setting an information query field in the information query interface, and after receiving the user input, the corresponding target query sentence "japan" is determined.
Step S4200, according to the target query sentence, triggering to obtain the direct content corresponding to the target query sentence.
The direct content corresponding to the target query statement is acquired according to the information acquisition method provided in the first embodiment, and will not be described herein.
In this embodiment, the client implementing the present embodiment may receive an input operation of a user, determine that a corresponding target query sentence is sent to the server implementing the first embodiment, and obtain, by the server, direct content corresponding to the target query sentence according to the information obtaining method of the first embodiment, and return the direct content to the client.
Step S4300, the direct content and the target query statement are displayed in a correlated mode, and information expected to be acquired by the user is provided for the user.
And the direct content corresponding to the target query statement is associated with the target query statement for display, so that a user can acquire the direct content which accurately meets the information query requirement of the user in real time without waiting for the information searching process and carrying out information browsing selection after inputting the target query statement, the information acquisition time of the user is greatly shortened, and the information acquisition cost of the user is reduced.
In a specific example, the target query sentence is displayed after receiving an input operation through an information query field provided in the information query interface, for example, as shown in fig. 6, a user inputs "japan" in the information query field in the information query interface, and correspondingly, displays "japan" in the information query field:
And displaying the content item corresponding to the direct content through a content display area arranged above the information inquiry bar.
The content item is used for displaying the direct content after receiving the click operation of the user.
For example, assuming that the user inputs "japan" and the corresponding acquired direct content is a vertical card introducing japan, as shown in fig. 6, a content item corresponding to the vertical card, which is displayed with a picture related to the date in the vertical card as an icon and information in a part of the vertical card, may be displayed above the information query field, and the user may be allowed to click directly and then display the specific content of the vertical card.
In this embodiment, in step S2200, the direct content corresponding to the target query term may be acquired, and the recommended query term set corresponding to the target query term may be acquired at the same time, and in step S2300, the direct content and the target query term are displayed in association with each other, and the recommended query term and the target query term are displayed in association with each other, for example, as shown in fig. 6, in an area above the content display area above the information input frame, and the recommended query term "weather in japan" corresponding to "japan", "movie in japan", and the like are displayed. The user can obtain the information by clicking the recommended query statement besides the direct content, and the information obtaining range is enlarged.
In this embodiment, the provided information acquisition method further includes:
and when the input operation is detected to change, re-determining a target query statement corresponding to the changed input operation, and executing the steps of triggering and acquiring the direct content according to the target query statement, and displaying the direct content and the target query statement in an associated mode.
For example, after the user inputs "japan", the user continues to input "wandering", re-determines that the target query sentence corresponding to the changed input operation is "japan wandering", re-performs the step of triggering acquisition of the direct content according to "japan wandering" and displaying the direct content in association with the target query sentence, and obtains the information query interface of the refresh display, or the user deletes "book" after inputting "japan", re-determines that the target query sentence corresponding to the changed input operation is "day", re-performs the step of triggering acquisition of the direct content according to "day", and displaying the direct content in association with the target query sentence, and obtains the information query interface of the refresh display.
Through dynamically acquiring corresponding direct content for associated display according to the change of user input operation, the change of the information acquisition requirement of the user can be adaptively adapted, and the direct content meeting the changed information acquisition requirement is acquired for display so as to meet the information acquisition requirement of the user in real-time dynamic change.
< Information acquisition device >
In this embodiment, an information obtaining apparatus 5000 is further provided, as shown in fig. 7, and includes a query determining unit 5100, a direct obtaining unit 5200, and an associated display unit 5300, which are used to implement the information obtaining method provided in this embodiment, and are not described herein.
The query determining unit 5100 is configured to receive an input operation of a user and determine a target query sentence corresponding to the input operation.
A direct acquiring unit 5200, configured to trigger, according to the target query statement, acquiring direct content corresponding to the target query statement, where the direct content is acquired according to any one of the information acquiring methods in the first embodiment.
And the association display unit 5300 is configured to perform association display on the direct content and the target query sentence, so as to provide the user with information that the user desires to obtain.
Optionally, the target query sentence is displayed after receiving the input operation through an information query column provided in the information query interface, and the association display unit 5300 is further configured to:
And displaying a content item corresponding to the direct content through a content display area arranged above the information query field, wherein the content item is used for displaying the direct content after receiving clicking operation of a user.
Optionally, the information acquisition device 5000 is further configured to:
And when the input operation is detected to change, re-determining the target query statement corresponding to the changed input operation, executing the trigger according to the target query statement to obtain the direct content, and displaying the direct content and the target query statement in an associated mode.
It will be appreciated by those skilled in the art that the information acquisition device 5000 may be implemented in various ways. The information acquisition device 5000 may be implemented by an instruction configuration processor, for example. For example, instructions may be stored in a ROM, and when the apparatus is started, the instructions are read from the ROM into a programmable device to implement the information acquisition apparatus 5000. For example, the information acquiring apparatus 5000 may be solidified into a dedicated device (for example, ASIC). The information acquiring apparatus 5000 may be divided into units independent of each other, or they may be implemented by being combined together. The information acquisition apparatus 5000 may be implemented by one of the above-described various implementations, or may be implemented by a combination of two or more of the above-described various implementations.
In the present embodiment, the information acquisition apparatus 5000 may be client application software that provides an information acquisition service, and may be, for example, a browser client, an information streaming application client, or the like.
< Information acquisition apparatus >
In the present embodiment, there is also provided an information acquisition apparatus 6000, as shown in fig. 8, including:
A display 6100;
A memory 6200 for storing executable instructions;
And a processor for executing the information acquisition device 6000 to perform the information acquisition method as provided in the present embodiment according to the executable instructions.
In this embodiment, the information obtaining device 6000 may be an electronic device such as a mobile phone, a desktop computer, a notebook computer, or a tablet computer. For example, the information acquisition device 6000 may be a cellular phone in which client application software providing an information acquisition service is installed.
In one example, the information acquisition apparatus 6000 may also include a communication device or the like as in the client 1200 shown in fig. 1.
< Readable storage Medium >
In the present embodiment, there is also provided a readable storage medium storing a computer program readable and executable by a computer for executing the information acquisition method described in the present embodiment when the computer program is read and executed by the computer.
The readable storage medium may be a tangible device that can hold and store instructions for use by an instruction execution device. The readable storage medium may be, for example, but not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples of a readable storage medium (a non-exhaustive list) include a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a Static Random Access Memory (SRAM), a portable compact disc read-only memory (CD-ROM), a Digital Versatile Disc (DVD), a memory stick, a floppy disk, a mechanical encoding device, punch cards or in-groove protrusion structures such as those having instructions stored thereon, and any suitable combination of the foregoing. A readable storage medium as used herein is not to be construed as a transitory signal itself, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (e.g., a pulse of light through a fiber optic cable), or an electrical signal transmitted through an electrical wire.
The second embodiment of the present invention has been described above with reference to the accompanying drawings, and according to this embodiment, an information acquisition method, apparatus, device, and readable storage medium are provided, where, according to a target query statement input by a user, direct content that meets a user information query requirement and the target query statement are triggered to be displayed in association, so that after the user inputs the target query statement, the user can acquire information that accurately meets the user information query requirement without waiting for a process of information search and performing information browsing selection, thereby greatly shortening a time for the user to acquire information, and reducing an information acquisition cost of the user.
< Third embodiment >
In this embodiment, there is also provided an information acquisition system 7000, including:
The information acquisition apparatus 3000 provided in the first embodiment and the information acquisition apparatus 5000 provided in the second embodiment;
Or alternatively
The information acquisition apparatus 4000 provided in the first embodiment and the information acquisition apparatus 6000 provided in the second embodiment.
In one example, the information acquisition system 7000 may include the server 1100 as the information acquisition device 4000 and the client 1200 as the information acquisition device 6000 as the information acquisition system 1000 shown in fig. 1.
< Example >
The information acquisition method implemented by the information acquisition system 7000 in this embodiment will be further exemplified below in conjunction with fig. 9, in which the information acquisition system 7000 includes a server as the information acquisition device 4000 and a client as the information acquisition device 6000.
As shown in fig. 9, the information acquisition method includes S301 to S312.
S301, the client receives input operation of a user and determines a corresponding target query statement.
S302, the client sends the target query statement to the server.
S303, the server determines a recommended query statement set corresponding to the target query statement according to the received target query statement.
S304, the server acquires the target query statement and the triggering content of each recommended query statement in the recommended query statement set through a preset content triggering mode.
In this example, the content trigger modes include a vertical card trigger, a carefully selected document trigger, a official data trigger, and a vertical skill trigger.
S305, the server judges whether the corresponding target query statement or the recommended query statement is a high-frequency query statement according to the target query statement and the triggering content of each recommended query statement, if so, the server goes to S306, and if not, the server goes to S307.
S306, the server calculates click scores according to the trigger content, and confidence judgment is carried out according to the score confidence conditions.
S307, the server acquires the target requirement category and the target query mode of the corresponding query statement, acquires the corresponding mode state according to the pre-acquired mode state matching list, and carries out confidence judgment.
S308, the server obtains a confidence query statement judged by the confidence.
S309, the server calculates the relevance scores of the target query statement and each recommended query statement according to the two relevance models, namely GDBT model and the Wide & Deep model.
S310, the server determines that the trigger content of the confidence query statement with the relevance score meeting the content direct condition is direct content corresponding to the target query statement.
S311, the server returns the direct content corresponding to the target query statement and the recommended query statement set to the client.
S312, the client displays the direct content and the recommended query statement set corresponding to the target query statement in association with the target query statement.
In this example, assuming that the target query sentence is a "non-name lifetime", the corresponding obtained direct content is a vertical card of the movie of "non-name lifetime", the recommended query sentence set includes a "non-name lifetime movie" and the like, the information query interface for performing the associated display may be as shown in fig. 10, the user may trigger to display the vertical card by clicking on the content item of the vertical card displayed in the content display area above the information query frame, and similarly, when the target query sentence is the name of a certain star, the direct content may be the vertical card of the star, and the information query interface corresponding to the associated display is also similar to that shown in fig. 10.
Or assuming that the target query statement is 'five mountain weather', the corresponding acquired direct content is content acquired by calling the vertical skill of the weather query, the recommended query statement set comprises 'five mountain weather forecast' and the like, an information query interface for carrying out associated display can be shown in fig. 11, a user can directly check the corresponding five mountain weather through a content item displayed in a content display area above an information query frame and can click the content item to enter a more detailed weather forecast page, and similarly, the target query statement is an express bill number, a national exchange rate, an express content can be express logistics, a real-time exchange rate, a test time place and the like which call the corresponding vertical skill acquisition when a certain test is carried out, and the information query interface corresponding to the associated display is similar to that shown in fig. 11.
Or assuming that the target query statement is "survival" and the corresponding acquired direct content is "survival" game official network data, the recommended query statement set includes "survival update notice" and the like, and the information query interface for performing associated display can be as shown in fig. 12, and the user can trigger to jump into the "survival" official network page by clicking the content item of the official network data displayed in the content display area above the information query frame.
Or assuming that the target query statement is "why sky is", the corresponding acquired direct content is related content in the choice question and answer extracted from the available knowledge graph, the recommended query statement set includes "why sky is blue" and the like, and the information query interface for performing associated display may be as shown in fig. 13, and the user may trigger to display all the content of the corresponding direct content by clicking on the content item displayed in the content display area above the information query frame.
The specific implementation method of each of the steps S301 to S312 may be as described in the first and second embodiments, and will not be described herein.
Through the information acquisition system in the example, after the client receives the target query statement input by the user, the target query statement input by the user can be sent to the server, the trigger server respectively acquires the target query statement input by the user and the trigger content which is associated with each recommended query statement corresponding to the target query statement in advance in a preset content trigger mode, then determines the direct content which corresponds to the target query statement and can accurately meet the information query requirement of the user according to the target query statement and the trigger content of each recommended query statement, the server returns the direct content to the client, and the client carries out association display on the direct content and the target query statement, so that the user can acquire the information which accurately meets the information query requirement of the user without waiting for the process of information searching and carrying out information browsing selection after inputting the target query statement, the time for acquiring the information of the user is greatly shortened, and the information acquisition cost of the user is reduced.
The present invention may be a system, method, and/or computer program product. The computer program product may include a computer readable storage medium having computer readable program instructions embodied thereon for causing a processor to implement aspects of the present invention.
The computer readable storage medium may be a tangible device that can hold and store instructions for use by an instruction execution device. The computer readable storage medium may be, for example, but not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of the computer-readable storage medium include a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a Static Random Access Memory (SRAM), a portable compact disc read-only memory (CD-ROM), a Digital Versatile Disc (DVD), a memory stick, a floppy disk, a mechanical encoding device, punch cards or intra-groove protrusion structures such as those having instructions stored thereon, and any suitable combination of the foregoing. Computer-readable storage media, as used herein, are not to be construed as transitory signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., optical pulses through fiber optic cables), or electrical signals transmitted through wires.
The computer readable program instructions described herein may be downloaded from a computer readable storage medium to a respective computing/processing device or to an external computer or external storage device over a network, such as the internet, a local area network, a wide area network, and/or a wireless network. The network may include copper transmission cables, fiber optic transmissions, wireless transmissions, routers, firewalls, switches, gateway computers and/or edge servers. The network interface card or network interface in each computing/processing device receives computer readable program instructions from the network and forwards the computer readable program instructions for storage in a computer readable storage medium in the respective computing/processing device.
Computer program instructions for carrying out operations of the present invention may be assembly instructions, instruction Set Architecture (ISA) instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source or object code written in any combination of one or more programming languages, including an object oriented programming language such as SMALLTALK, C ++ or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The computer readable program instructions may be executed entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computer (for example, through the Internet using an Internet service provider). In some embodiments, aspects of the present invention are implemented by personalizing electronic circuitry, such as programmable logic circuitry, field Programmable Gate Arrays (FPGAs), or Programmable Logic Arrays (PLAs), with state information for computer readable program instructions, which can execute the computer readable program instructions.
Various aspects of the present invention are described herein with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer-readable program instructions.
These computer readable program instructions may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions/acts specified in the flowchart and/or block diagram block or blocks. These computer readable program instructions may also be stored in a computer readable storage medium that can direct a computer, programmable data processing apparatus, and/or other devices to function in a particular manner, such that the computer readable medium having the instructions stored therein includes an article of manufacture including instructions which implement the function/act specified in the flowchart and/or block diagram block or blocks.
The computer readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other devices to cause a series of operational steps to be performed on the computer, other programmable apparatus or other devices to produce a computer implemented process such that the instructions which execute on the computer, other programmable apparatus or other devices implement the functions/acts specified in the flowchart and/or block diagram block or blocks.
The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of instructions, which comprises one or more executable instructions for implementing the specified logical function(s). In some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions. It is well known to those skilled in the art that implementation by hardware, implementation by software, and implementation by a combination of software and hardware are all equivalent.
The foregoing description of embodiments of the invention has been presented for purposes of illustration and description, and is not intended to be exhaustive or limited to the embodiments disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the various embodiments described. The terminology used herein was chosen in order to best explain the principles of the embodiments, the practical application, or the technical improvements in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein. The scope of the invention is defined by the appended claims.