CN111104591A

CN111104591A - Recommendation information generation method and device

Info

Publication number: CN111104591A
Application number: CN201911199008.4A
Authority: CN
Inventors: 张琳; 梁忠平
Original assignee: Alipay Hangzhou Information Technology Co Ltd
Current assignee: Alipay Hangzhou Information Technology Co Ltd
Priority date: 2019-11-29
Filing date: 2019-11-29
Publication date: 2020-05-05
Anticipated expiration: 2039-11-29
Also published as: CN111104591B

Abstract

The present specification provides a recommendation information generation method and apparatus, the method including: when information is recommended, a plurality of pieces of recommendation information are generated once in a single round, the recommendation information is converted into a plurality of pieces of recommendation information generated in multiple times in a single round, one piece of recommendation information is generated each time, and the recommendation information generated last time and the information recommendation correlation factor are used as input information for next generation of recommendation information. By converting the mode that a plurality of pieces of recommendation information are generated once in a single round into the mode that the plurality of pieces of recommendation information are generated in multiple times in a single round, the historical recommendation information can be added in the generation process of each piece of recommendation information.

Description

Recommendation information generation method and device

Technical Field

The specification belongs to the technical field of computers, and particularly relates to a recommendation information generation method and device.

Background

With the development of science and technology, more and more intelligent recommendation software or equipment appears in people's lives, such as: the conversation robot can recommend questions that the user wants to ask for the user according to the input of the user, and some shopping websites can recommend commodities that the user may need when the user opens the website. Some software may recommend a corresponding answer to the user according to the questioning information input by the user, such as: recommending scenic spot information for the user according to the questioning or browsing records of the user, and the like. Therefore, how to generate accurate recommendation information is increasingly important.

Disclosure of Invention

An object of the embodiments of the present specification is to provide a recommendation information generation method, an apparatus, and a conversation robot, which enhance the output effect of an information recommendation model and improve the accuracy of recommendation information.

In one aspect, an embodiment of the present specification provides a recommendation information generation method, including:

obtaining historical recommendation information generated by an information recommendation model, wherein the historical recommendation information comprises at least one of the following: historical output recommendation information generated and output by the information recommendation model and historical generation recommendation information which is not output by the information recommendation model;

taking the obtained information recommendation association factor and the historical recommendation information as input information of the information recommendation model to generate recommendation information;

adding the generated recommendation information into a set of history generation recommendation information to serve as input information of the information recommendation model in the next information recommendation, and generating recommendation information until the generated recommendation information meets the output rule of the information recommendation model;

and outputting the generated recommendation information according to the output rule.

In another aspect, the present specification provides a recommendation information generating apparatus including:

the historical recommendation information acquisition module is used for acquiring historical recommendation information generated by an information recommendation model, and the historical recommendation information comprises at least one of the following: historical output recommendation information generated and output by the information recommendation model and historical generation recommendation information which is not output by the information recommendation model;

the recommendation information generation module is used for taking the acquired information recommendation association factor and the historical recommendation information as input information of the information recommendation model to generate recommendation information;

the recommendation information cycle generation module is used for adding the generated recommendation information into a set of history generation recommendation information to be used as input information of the information recommendation model in the next information recommendation, and generating recommendation information until the generated recommendation information meets the output rule of the information recommendation model;

and the recommendation information output module is used for outputting the generated recommendation information according to the output rule.

In another aspect, an embodiment of the present specification provides a conversation robot, including: at least one processor and a memory for storing processor-executable instructions, the memory having an information recommendation model stored therein;

the processor is configured to obtain an information recommendation association factor, historical recommendation information generated by the information recommendation model, information entropy of candidate recommendation information, and location information of currently generated recommendation information, where the historical recommendation information includes at least one of: historical output recommendation information generated and output by the information recommendation model and historical generation recommendation information which is not output by the information recommendation model;

the processor is further configured to input the information recommendation association factor, the historical recommendation information, the information entropy of the candidate recommendation information, and the location information into the information recommendation model, generate recommendation information using the information recommendation model, and output the recommendation information based on an output rule by the information recommendation model.

In yet another aspect, the present specification provides a data processing apparatus for recommendation information generation, including: the recommendation information generation device comprises at least one processor and a memory for storing processor-executable instructions, wherein the processor executes the instructions to realize the recommendation information generation method.

In the recommendation information generation method, the recommendation information generation device, the processing device and the conversation robot provided by the specification, when information is recommended, a single round of once generating a plurality of pieces of recommendation information is converted into a single round of repeatedly generating a plurality of pieces of recommendation information, one piece of recommendation information is generated each time, and the recommendation information generated last time is used as input information for next generation of recommendation information. By converting the mode that the recommendation information is generated once in a single round into the mode that the recommendation information is generated repeatedly in the single round, the historical recommendation information can be added in the generation process of the recommendation information each time, learning of the association relation between the recommendation information is achieved, learning prediction of the recommendation information is performed for multiple times through one-time output, the effect that the information recommendation model recommends multiple recommendation information once is strengthened, and the accuracy of the recommendation information is improved.

Drawings

In order to more clearly illustrate the embodiments of the present specification or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below, it is obvious that the drawings in the following description are only some embodiments described in the present specification, and for those skilled in the art, other drawings can be obtained according to the drawings without any creative effort.

FIG. 1 is a flow chart illustrating a method for generating recommendation information in an embodiment of the present disclosure;

FIG. 2 is a schematic block diagram of a schematic framework for recommendation information generation in one embodiment of the present description;

FIG. 3 is a schematic diagram of a recommendation generation method in yet another embodiment of the present disclosure;

fig. 4 is a schematic block diagram of an embodiment of a recommendation information generation apparatus provided in the present specification;

fig. 5 is a block diagram of a hardware configuration of the recommendation information generation server in one embodiment of the present specification.

Detailed Description

In order to make those skilled in the art better understand the technical solutions in the present specification, the technical solutions in the embodiments of the present specification will be clearly and completely described below with reference to the drawings in the embodiments of the present specification, and it is obvious that the described embodiments are only a part of the embodiments of the present specification, and not all of the embodiments. All other embodiments obtained by a person skilled in the art based on the embodiments in the present specification without any inventive step should fall within the scope of protection of the present specification.

Along with the development of computer technology, intelligent recommendation technology is more and more appeared in the life of people, and the work and the life of people are facilitated. Intelligent recommendation can be generally implemented based on a machine learning model constructed from historical data, and the machine learning model can recommend n pieces of recommendation information or 1 piece of recommendation information at a time. Such as: the dialogue robot may recommend 1 question to the user, which is most likely to be a question of the user, or recommend the top 5 questions as questions the user may ask, selected by the user. When recommending n pieces of recommendation information, the model generates a plurality of pieces of recommendation information at a time, scores the generated recommendation information, and outputs the top n pieces of recommendation information with the highest scores, but the n pieces of recommendation information may have similar recommendation information, so that the recommended information is not accurate enough.

The embodiment of the specification provides a recommendation information generation method, which can divide a mode of generating n pieces of recommendation information once into n times to complete, 1 piece of recommendation information is generated each time, the previous piece of recommendation information is used as input of the next piece of recommendation information, historical recommendation information is added in the generation process of the recommendation information, the relation among the recommendation information is effectively controlled, the effect of recommending a plurality of pieces of recommendation information at a time is strengthened, and the accuracy of the recommendation information is improved.

The recommendation information generation method in the specification can be applied to a client or a server, and the client can be an electronic device such as a smart phone, a tablet computer, a smart wearable device (a smart watch, virtual reality glasses, a virtual reality helmet, and the like), and a smart vehicle-mounted device.

Fig. 1 is a schematic flow diagram of a recommendation information generation method in an embodiment of this specification, and as shown in fig. 1, the recommendation information generation method provided in an embodiment of this specification may include:

102, obtaining historical recommendation information generated by an information recommendation model, wherein the historical recommendation information comprises at least one of the following: and historical output recommendation information generated and output by the information recommendation model and historical generation recommendation information which is not output by the information recommendation model are generated by the information recommendation model.

In a specific implementation process, in some embodiments of the present description, an information recommendation model may be built in advance according to historical data, and the information recommendation model may generate new recommendation information based on an information recommendation association factor, historical recommendation information, and the like. The information recommendation model may be a machine learning model, and specifically may be a reinforcement learning-based machine learning model or a supervised learning-based machine learning model, and the embodiments of this specification are not particularly limited with respect to the specific form of the model. The information recommendation model in the embodiment of the present specification may generate recommendation information multiple times in each round of information recommendation, but the generated recommendation information is not necessarily output and displayed to the user immediately. When information recommendation is performed, the embodiment of the specification can acquire historical recommendation information recommended by current information together with other information as input information of an information recommendation model, and the historical recommendation information and the other information are used for prediction generation of next recommendation information. The history recommendation information may include history output recommendation information generated and output by the information recommendation model, and history generation recommendation information not output by the information recommendation model. The history output recommendation information may be recommendation information output in a history specified turn, and the history generation recommendation information may be recommendation information generated in the current information recommendation process but not output yet. The historical output recommendation information can be the same recommendation process such as: when recommendation information is continuously generated for the same topic or for one user, the recommendation information generated and output in the front is taken as history output recommendation information. When the topic is changed or the user is changed, the history output recommendation information should be acquired again.

For example: in one embodiment, each round of information recommendation can output 3 pieces of recommendation information, but the information recommendation model generates the recommendation information by predicting 3 times, and generates one piece of recommendation information each time. And generating recommendation information, and then not outputting the recommendation information, wherein the recommendation information generated last time is history generation recommendation information and is used as input information for next generation of recommendation information. Meanwhile, when each information recommendation is generated, the recommendation information which is already output by the previous information recommendation can be used as historical output recommendation information and also as input information of an information recommendation model, and 3 pieces of recommendation information are output together until 3 pieces of recommendation information are generated. When the user switches topics, the history recommendation information is 0 when the information recommendation of the topic is switched for the first time, and when the information recommendation is performed for a new user, the history recommendation information should be 0 when the recommendation information is generated for the first time.

And step 104, taking the acquired information recommendation association factor and the historical recommendation information as input information of the information recommendation model to generate recommendation information.

In a specific implementation process, the recommendation association factor may represent a feature related to the recommendation information, and may be set according to an actual application scenario and information to be recommended, which is not specifically limited in this specification. Such as: when the dialogue robot recommends questions asked by the user, the information recommendation association factor may be questions asked by the user in history, text information currently input by the user, and the like, and for other application scenario information recommendation association factor may include characteristic information of the user (such as age, academic history, occupation, gender, income, and the like), track of using services, user account information, behavior information such as browsing records, device information, historical shopping records, and the like. It should be noted that the information recommendation association factor may be obtained only when the first recommendation information is generated, and the obtained information recommendation association factor may be directly used when the recommendation information is generated subsequently. When information recommendation is generated every time, the acquired historical recommendation information and the information recommendation association factor are used as input information of an information recommendation model, and the information recommendation model can generate recommendation information based on the input information.

For example: in one embodiment, the conversation robot is provided with the information recommendation model in the embodiments of the present specification, a user makes multiple rounds of questions in the conversation robot, and each round of information recommendation can output 3 questions for the user to select. And when a second round of question answering is carried out, the second round of first question forecasting can obtain 3 questions output in the first round and information recommendation correlation factors as input of the information recommendation model to generate recommendation information, and then second round of second question forecasting is carried out. The second round of second-time problem prediction can obtain 3 problems output in the first round, and the recommendation information generated for the first time but not output in the second round and the information recommendation association factor as the input of the information recommendation model, so as to generate new recommendation information, and so on. The information recommendation association factor may include text information input by the user, question information of historical questions of the user, related information of the user account, and the like.

And 106, adding the generated recommendation information into a set of history generation recommendation information to serve as input information of the information recommendation model in the next information recommendation, and generating recommendation information until the generated recommendation information meets the output rule of the information recommendation model.

In a specific implementation process, in the embodiment of the present specification, a single round of information recommendation is converted into a single round of multiple information recommendation. Referring to the description of the above embodiment, in the same round of information recommendation process, the generated recommendation information may be used as input information for generating information recommendation next time in the current round, that is, the generated recommendation information may be added to the set of history generated recommendation information and input into the information recommendation model together with other input information to generate new recommendation information. By analogy, newly generated recommendation information is added into a set of history generation recommendation information, and is input into the information recommendation model together with other input information to generate recommendation information recommended by information for the next time until the generated recommendation information meets the output rule of the information recommendation model. The output rule can be preset, and whether the generated recommendation information can be output is judged based on the output rule. The specific content of the output rule can be set according to actual needs, such as: the generated recommendation information may be scored each time the recommendation information is generated, and when the score of the generated recommendation information is greater than a preset threshold, it may be considered that the output rule is satisfied. Or the output number of the recommendation information may be set, and when the number of the generated recommendation information is reached, the output rule is considered to be satisfied. Or a time threshold value for outputting the recommendation information can be set, and when the time is up, the output rule is considered to be satisfied. Specifically, the output rule of the information recommendation model may be set according to an actual usage scenario, and the embodiment of this specification is not particularly limited.

For example: in the above-described embodiment regarding the interactive robot, the recommendation information generated by the second round of second-time problem prediction may be input to the information recommendation model together with other input information as the history generation recommendation information in the second round of third-time problem prediction, and the information recommendation for the second round and the third time may be performed to generate new recommendation information. And judging whether the newly generated recommendation information meets the output rule, if not, adding the newly generated recommendation information into a set of history generation recommendation information, inputting the set of history generation recommendation information and other input information into an information recommendation model, and generating recommendation information recommended by the next information until the generated recommendation information meets the output rule of the information recommendation model.

And step 108, outputting the generated recommendation information according to the output rule of the information recommendation model.

In a specific implementation process, all the generated recommendation information (excluding the already output recommendation information) may be output based on the preset output rule of the information recommendation model. Such as: the generated recommendation information can be scored when the recommendation information is generated every time, and when the score of the generated recommendation information is larger than a preset threshold value, the recommendation information is considered to meet the output rule, and the recommendation information which is not output can be directly output and generated. Or the output number of the recommendation information can be set, and when the number of the generated recommendation information is reached, the recommendation information is considered to meet the output rule, and the generated recommendation information which is not output yet is output. Or a time threshold value for outputting the recommendation information can be set, and when the time is up, the recommendation information is considered to meet the output rule, and the recommendation information which is not output is output.

In some embodiments of the present specification, the outputting the recommendation information according to the output rule of the information recommendation model may include:

counting the number of the recommendation information which is generated by the information recommendation model and is not output, and if the number of the recommendation information reaches the information output number specified by the output rule, outputting the recommendation information which is generated by the information recommendation model and is not output;

and if the quantity of the recommendation information does not reach the information output quantity specified by the output rule, inputting the recommendation information serving as history generation recommendation information into the information recommendation model, generating the next recommendation information, and outputting the recommendation information which is generated by the information recommendation model and is not output until the generated recommendation information reaches the information output quantity specified by the output rule.

In a specific implementation process, the information output number of the information recommendation model may be set as an output rule, and when the information recommendation model generates recommendation information each time, the number of recommendation information that is not output is generated by statistics. And if the number of the generated recommendation information reaches the information output number, outputting all the generated recommendation information which is not output. And if the number of the recommendation information does not reach the information output number in the statistics mode, adding the generated recommendation information serving as history generation recommendation information into a set of history generation recommendation information, inputting the history generation recommendation information and other input information into an information recommendation model, and generating the recommendation information for the next time until the generated recommendation information reaches the information output number. The information output quantity in the output rule of the information recommendation model is set, so that the information output of the information recommendation model can be controlled, the output information of the information recommendation model is adjusted, the flexible output of the recommendation information is realized, and the output recommendation information of the book is more in line with the requirements of users.

For example: in the above-described embodiment of the interactive robot, the interactive robot is provided with the information recommendation model in the embodiment of the present specification, the user makes multiple rounds of questions in the interactive robot, 3 questions can be output for the user to select for each round of information recommendation, and the number of information outputs set in the output rule of the information recommendation model can be considered to be 3. When the question and answer of the second round is carried out, the first question prediction of the second round can obtain 3 questions output by the first round and the information recommendation association factor as the input of the information recommendation model, and generate recommendation information. The statistically generated and non-output recommendation information is 1 and less than 3, 3 questions output in the first round can be acquired as historical output recommendation information, the recommendation information generated for the first time but not output in the second round can be acquired as historical generation recommendation information, the historical generation recommendation information and the information recommendation association factor are used as input of an information recommendation model, second-round problem prediction is carried out, and new recommendation information is generated. And counting that 2 is less than 3 for the recommendation information which is generated by the current information recommendation model and is not output, acquiring 3 questions output in the first round as historical output recommendation information, acquiring the recommendation information which is generated for the first time and the second time but is not output in the second round as historical generation recommendation information, and taking the recommendation information and the information recommendation correlation factor as the input of the information recommendation model to perform third-time question prediction in the second round to generate new recommendation information. And counting the number of the recommendation information which is generated by the current information recommendation model and is not output to be 3, outputting the generated 3 recommendation information when the requirement of the information output number is met, and ending the problem prediction of the second round.

In the recommendation information generation method provided in the embodiment of the present specification, when information is recommended, a single round of generating a plurality of pieces of recommendation information at a time is converted into a single round of generating a plurality of pieces of recommendation information multiple times, one piece of recommendation information is generated each time, and the recommendation information generated last time is used as input information for next generation of recommendation information. By converting the mode that the recommendation information is generated once in a single round into the mode that the recommendation information is generated repeatedly in the single round, the historical recommendation information can be added in the generation process of the recommendation information each time, learning of the association relation between the recommendation information is achieved, learning prediction of the recommendation information is performed for multiple times through one-time output, the effect that the information recommendation model recommends multiple recommendation information once is strengthened, and the accuracy of the recommendation information is improved.

On the basis of the foregoing embodiments, in some embodiments of the present specification, the method may further include:

and calculating the information entropy of the candidate recommendation information, and generating recommendation information by taking the information entropy of each candidate recommendation information, the information recommendation association factor and the historical recommendation information as input information of the information recommendation model.

In a specific implementation process, when information recommendation is performed, the information entropy of the candidate recommendation information can be calculated and input into the information recommendation model together with other input information to generate recommendation information each time information recommendation is generated. The candidate recommendation information may be understood as recommendation information that may be recommended, and may be constructed and obtained in advance according to data such as a historical recommendation record, a user query record, and an actual application scenario, and the candidate recommendation information may include the historical recommendation information. The information entropy may be understood as the occurrence probability of a certain specific information, and the information entropy of the candidate recommendation information in the embodiment of the present specification may be understood as the probability that the candidate recommendation information may occur, or the magnitude of the association relationship between the candidate recommendation information and the newly generated recommendation information. The calculation mode of the information entropy of each candidate recommendation information may be set according to an actual usage scenario, and the embodiment of the present specification is not particularly limited. Such as: 1/(the number of associated recommendation information of the candidate recommendation information) or 1/(the number of associated recommendation information of the candidate recommendation information excluding the historical recommendation information) may be used as the information entropy of the candidate recommendation information, and of course, other calculation methods may be adopted according to actual use needs.

For example: in an application scenario of tag and question recommendation, tag information can be understood as being composed of words or phrases and used for describing services or appeals, question information can be understood as a natural sentence and used for representing a question of a user, generally, a plurality of similar user questions correspond to one piece of question information, and one piece of question information can correspond to a plurality of pieces of tag information. The information entropy of the candidate recommended label can be calculated by calculating the number of questions additionally included in the candidate recommended label based on the questions included in the historical recommended label. Such as: if the information recommendation model has already been used as the recommendation label in the last time, "flower over money" is no longer used as any additional (i.e. beyond the label included in the "flower over money") label number for the candidate recommendation information during the information recommendation, because the coverage of "flower over money" is certainly larger. If the last recommended label of the information recommendation model is "borrow", then "repayment with flower" will cover the additional label because the coverage of the two labels is different. The number of questions additionally contained in 1/historical recommendation label can be used as the information entropy of the candidate recommendation label. Therefore, the information entropy of the candidate recommendation information is determined based on the historical recommendation information, the information entropy of the candidate recommendation information is different due to different historical recommendation information, the information entropy is used as the input information of the information recommendation model, the relationship between the recommended information can be controlled, and the recommended information is more accurate. In the embodiment of the specification, in each information recommendation process, the information entropy of the candidate recommendation information is added as the input information of the model to control the generation of the subsequent recommendation information, and the relationship between the recommendation information can be controlled, so that the recommended information better meets the requirements of users.

On the basis of the above embodiments, in some embodiments of the present specification, the method may further include:

determining the position information of the recommendation information, wherein the position information is the generation sequence of the recommendation information generated by the information recommendation model;

and generating recommendation information by taking the position information of the recommendation information as input information of an information recommendation model.

In a specific implementation process, the position information of the recommendation information can be added as the input of the information recommendation model in each information recommendation, and the position information can be understood as the sequence of the recommendation information generated by the information recommendation model. For example: in one information recommendation process, the information recommendation model has generated 2 pieces of recommendation information, and the position information of the recommendation information to be currently generated may be 3. It should be noted that the location information may refer to the total ranking of the recommendation information currently prepared to be generated in the recommendation information already output and the recommendation information not output in the generation of the information recommendation model. The generation of already output recommendation information can be understood as belonging to the same conversation process, and when a new topic or a new user exists, the calculation is restarted. Such as: in the above embodiment, the dialog robot outputs 3 pieces of recommendation information per round, and when performing the second round of information recommendation, the position information of the recommendation information generated for the second round should be 3+1 — 4.

The goal of the information recommendation model optimization in the embodiments of the present specification may include spitting appropriate recommendation information at an appropriate position, such as: there is recommended information that is more in line with the user's expectations that it should appear at an earlier location. In each round of information recommendation process, the position information of the recommendation information can be added as input, and the generated recommendation information and the feedback of the user recommendation information can be used as training data to optimize the information recommendation model. In the training process of the information recommendation model, the position information of each piece of recommendation information can be used as input information for model training, so that the model finds the relation with the position information from the training data, the optimization direction of the model is controlled, the better recommendation information is output earlier, and the generation efficiency of the recommendation information is improved.

On the basis of the foregoing embodiments, in some embodiments of the present specification, the outputting the recommendation information according to the output rule of the information recommendation model includes:

setting an output label, a non-output label and an output list;

storing the recommendation information into the output list;

marking the output label on the output recommended information and marking the non-output label on the non-output recommended information;

and outputting the recommendation information with the non-output label in the output list based on the output label and the non-output label on the recommendation information.

In a specific implementation process, an output list, an output tag and an un-output tag of the information recommendation model may be set, the output list may include all information that has been output and is not output, and the output tag and the un-output tag may indicate whether data in the output list is output, such as: 0, 1 may be used as the output tag and the non-output tag, respectively. Storing the generated recommendation information into an output list, marking output labels and non-output labels for each recommendation information according to whether the recommendation information is output, and when the recommendation information is required to be output, for example: and when the number of the recommendation information which is not output reaches the output number of the information recommendation model, outputting all the recommendation information with the label which is not output in the output list. The output of the model is controlled through the output label, the non-output label and the output list, so that the flexible output of the model is realized.

acquiring recommendation information clicked by a user according to the click data of the user on the output recommendation information;

and adding the recommendation information clicked by the user into the input information of the information recommendation model during the next round of information recommendation.

In a specific implementation process, as can be seen from the description of the above embodiment, information recommendation in the embodiment of the present specification may be divided into multiple rounds of information recommendation, where each round of information recommendation includes generation of multiple pieces of recommendation information, but one round of information recommendation is output. In some embodiments of the present specification, when there are multiple information recommendations, click data of recommendation information output in a previous information recommendation or previous information recommendations, that is, recommendation information output in a previous information recommendation or previous information recommendations clicked by a user, may be used as input information of an information recommendation model together with other input information to generate new recommendation information when a new information recommendation is generated.

For example: in the above embodiment, the conversation robot is provided with the information recommendation model in the embodiment of the present specification, a user makes multiple rounds of questions in the conversation robot, and each round of information recommendation can output 3 questions for the user to select. When information recommendation to the second round for the first time is performed, click data of 3 questions output by the user for the first round may be added to the input information. Such as: the model outputs a question A, B, C, and the user clicks on a, which can be added to the input information, along with other input information, as input information for the second round of information recommendation by the information recommendation model.

In the embodiment of the specification, click data of the front-wheel user is added as input information in the multi-round information recommendation process, so that the accuracy of generation of the next round of recommendation information is improved.

Fig. 2 is a schematic diagram of a principle framework of generation of recommendation information according to an embodiment of the present disclosure, as shown in fig. 2, in some embodiments of the present disclosure, the recommendation information may be tag information, and the method may further include:

acquiring the label information clicked by the user according to the click data of the user on the label information;

taking the label information clicked by the user as input information of the information recommendation model when the next round of label information is generated, and generating label information recommended by the next round of information;

judging whether the label information recommended by the next round of information meets a preset requirement or not, if not, acquiring the label information clicked by the user as input information of the information recommendation model when the label information is generated by the next round of label information according to click data of the output label information recommended by the next round of information by the user until the output label information meets the preset condition;

and according to the click data of the last round of output label information by the user, acquiring the last round of output label information clicked by the user as input information of the information recommendation model, generating question information and outputting the question information.

In a specific implementation process, the tag information may be understood as being composed of words or phrases for describing a service or a appeal, the question information may be understood as a natural sentence for representing a question of a user, generally, a plurality of similar user questions correspond to one question information, and one question information may correspond to a plurality of tag information. Generally, in a question and answer scene such as a conversation robot, label information can be recommended for a user, and based on click data of the user on the recommended label information, question information can be recommended for the user. As shown in fig. 2, in an example scenario of this specification, it may be set that 3 rounds of information are recommended to a user first, the first two rounds are recommended to tag information, each round is recommended to 3 tag information, the last round is recommended to question information, and finally 3 question information is recommended to the user. The following describes generation of the challenge information in the embodiment of the present specification with reference to fig. 2:

as shown in fig. 2, in the first round of generation of tag information, the information recommendation association factor, i.e. S in the graph, may be obtained first₀Recommending the information to a correlation factor S₀Inputting the information into an information recommendation model, and generating first label information t of a first round₁. Then will beS₀And t₁Generating a first round of second label information t as input information for the first round of second label information generation₂. Then the S is mixed₀、t₁、t₂Generating the first round of third label information t as the input information generated by the first round of third label information generation₃. Outputting the 3 label information t generated in the first round₁、t₂、t₃According to the click data of the user on the label information output in the first round, the click label information of the user, such as t, is obtained₃. Tag information t clicked by user₃Other input information (such as information recommendation association factor, label information t output in the first round) can also be used as the input information of the information recommendation model in the next round of label information generation₁、t₂Whether the user clicks or not can be marked on the label information generated in the first round) and the label information is input into the information recommendation model, and the label information generated in the next round is generated. The generation process of the second round of label information is the same as that of the first round of label information, and is not repeated here.

And judging whether the label information generated in the second round meets preset requirements (such as whether the requirement of the label information generation round is met or not, or whether the score of the generated label information reaches a threshold value or not), if not, acquiring the label information clicked by the user according to click data of the label information output by the user in the second round as input information of an information recommendation model in the next round, namely the third round, until the output label information meets the preset requirements. As shown in fig. 2, if the label information generated in the second round meets the preset requirement, the label information t generated in the second round may be obtained₄、t₅、t₆And (6) outputting. The specific content of the preset requirement may be set according to actual needs, may be set as the number of label output rounds, or other requirements, and the embodiments of the present specification are not specifically limited.

According to the click data of the last round of output label information by the user, obtaining the last round of output label information clicked by the user as the input information of the information recommendation model, generating question information, and generating question informationAnd outputting the information. As shown in FIG. 2, the label information t output for the second round of user click may be output₆The information is input into the information recommendation model, and is input into the information recommendation model together with other input information (for example, information recommendation association factor, label information output by the first round and the second round, whether a user clicks the label information generated by the first round and the second round) to generate question information. The generation of the mark information is the same as the generation process of the first round of label information, and 3 mark information q are generated respectively for 3 times₁、q₂、q₃For a specific process, refer to the process of generating the first round of tag information, which is not described herein again. C in FIG. 2_TCan represent input information at the time of generation of tag information, c_QInput information at the time of generation of the question information may be represented.

In the generation process of the first round of label information, the second round of label information and the third round of question information, each time one label information or question information is generated, whether the generated label information or question information meets the output rule (for example, whether the output quantity of the information reaches each round) or not can be judged, and if the generated label information or question information does not meet the output rule, the generated label information or question information is used as the input information of the information recommendation model to generate the next label information or question information. And continuously judging whether the generated label information or the generated question information conforms to the output rule, if not, taking the generated label information or the generated question information as input information of the information recommendation model until the generated label information or the generated question information conforms to the output rule, and outputting all the label information or all the question information generated in each round.

It should be noted that, in the above embodiments, the number of the label information or the question information generated in each round and in the round may be adjusted according to actual needs, and the embodiments of the present specification are not particularly limited.

In the embodiment of the specification, a mode of converting single-round generation and multiple generation of the labels and the questions into single-theory multiple generation is adopted, for a user, the labels seen by the user are three labels in each round, and when the model is actually predicted, the three labels are predicted three times to obtain three labels respectively. Through the form of single-round multi-time prediction, the characteristics of historical label information, information entropy, position information, output rules and the like of one round can be added in each round of prediction to control the recommendation of subsequent labels, so that the relation among the recommended labels can be controlled to meet the requirements of users, the flexible control of model output is realized, and the accuracy of label and question recommendation is improved.

Fig. 3 is a schematic diagram illustrating a method for generating recommendation information according to another embodiment of the present disclosure, and as shown in fig. 3, the information recommendation process according to the embodiment of the present disclosure is as follows:

besides the information recommendation association factors used normally, the input of the information recommendation model also adds contents such as history recommendation information, information entropy, location information, and forcing rules (i.e. the output rules in the above embodiments), where the input information is introduced as follows:

the history recommendation information may be input in the form of a list, such as the beijiao value already recommended, and the beijiao value is input as a list containing the beijiao value and the beijiao value. The specific meaning of the historical recommendation information can refer to the description of the above embodiments, and is not described herein again.

Information entropy: here, the information entropy of each candidate recommendation information (which may be constructed and obtained in advance according to data such as a historical recommendation record) is calculated (if the recommendation information is tag information, the information entropy may be calculated by calculating how many additional tags can be included). Position information: currently, the first information recommendation.

And (3) forcing rules: is a string of 0, 1 generated by a rule, and can be used to suppress the output of specific recommendation information

And other contents are as follows: the embodiment of the present specification is not particularly limited, and the addition of other content may enable the output of the model to have a control advantage for each recommended information output.

It should be noted that, in the embodiment of the present specification, a process of generating recommendation information each time when information is recommended multiple times in a single theory is given. In practical application, when there are multiple rounds of information recommendation processes, the click data of the recommendation information recommended by the user in the previous round can also be used as input information of an information recommendation model in the next round of information recommendation, so that the accuracy of the recommended information is improved, and the recommended information is more in line with the requirements of the user.

In the embodiment of the specification, the single-round multi-label recommendation process is split, and the training process of a single round once is converted into the training process of multiple times of a single round, so that the characteristic of learning the label relation can be added during training, and the effect of recommending multiple labels by the recommendation model at a time is enhanced. By the step-by-step label recommendation training method, the labels which are pushed out before can be added as the input of the model, and on the basis, the contents such as information entropy, position information, strong rules and the like are allowed to be added to control the recommendation of the subsequent labels, so that the relation among the recommended labels can be controlled to meet the requirements of users.

In the present specification, each embodiment of the method is described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. The relevant points can be obtained by referring to the partial description of the method embodiment.

Based on the recommendation information generation method, one or more embodiments of the present specification further provide a recommendation information generation apparatus. The apparatus may include systems (including distributed systems), software (applications), modules, components, servers, clients, etc. that use the methods described in the embodiments of the present specification in conjunction with any necessary apparatus to implement the hardware. Based on the same innovative conception, embodiments of the present specification provide an apparatus as described in the following embodiments. Since the implementation scheme of the apparatus for solving the problem is similar to that of the method, the specific apparatus implementation in the embodiment of the present specification may refer to the implementation of the foregoing method, and repeated details are not repeated. As used hereinafter, the term "unit" or "module" may be a combination of software and/or hardware that implements a predetermined function. Although the means described in the embodiments below are preferably implemented in software, an implementation in hardware, or a combination of software and hardware is also possible and contemplated.

Specifically, fig. 4 is a schematic block structure diagram of an embodiment of a recommendation information generation apparatus provided in this specification, and as shown in fig. 4, the recommendation information generation apparatus provided in this specification may include: a history recommendation information acquisition module 41, a recommendation information generation module 42, a recommendation information cycle generation module 43, and a recommendation information output module 44, wherein:

the historical recommendation information obtaining module 41 may be configured to obtain historical recommendation information generated by an information recommendation model, where the historical recommendation information includes at least one of: historical output recommendation information generated and output by the information recommendation model and historical generation recommendation information which is not output by the information recommendation model;

the recommendation information generation module 42 may be configured to use the obtained information recommendation association factor and the historical recommendation information as input information of the information recommendation model to generate recommendation information;

the recommendation information cycle generation module 43 may be configured to add the generated recommendation information to a set of history generated recommendation information, as input information of the information recommendation model in the next information recommendation, and generate recommendation information until the generated recommendation information meets an output rule of the information recommendation model;

and a recommendation information output module 44, configured to output the generated recommendation information according to the output rule.

In the recommendation information generation apparatus provided in the embodiment of the present specification, when information is recommended, a single round of generating a plurality of pieces of recommendation information at a time is converted into a single round of generating a plurality of pieces of recommendation information at a plurality of times, one piece of recommendation information is generated at a time, and the recommendation information generated last time is used as input information for next generation of recommendation information. By converting the mode that the recommendation information is generated once in a single round into the mode that the recommendation information is generated multiple times in the single round, the historical recommendation information can be added in the generation process of the recommendation information every time, learning of the association relation between the recommendation information is achieved, 3 times of learning prediction of the recommendation information is performed through one-time output, the effect that the information recommendation model recommends multiple recommendation information once is strengthened, and the accuracy of the recommendation information is improved.

On the basis of the foregoing embodiments, in some embodiments of the present specification, the recommendation information output module is specifically configured to:

In the embodiment of the specification, by setting the information output quantity in the output rule of the information recommendation model, the information output of the information recommendation model can be controlled, the output information of the information recommendation model is adjusted, the flexible output of the recommendation information is realized, and the output recommendation information of the book is more in line with the user requirements.

On the basis of the foregoing embodiments, in some embodiments of the present specification, the apparatus further includes an information entropy calculation module configured to: calculating the information entropy of the candidate recommendation information according to the historical recommendation information, and sending the calculated information entropy of each candidate recommendation information to the recommendation information generation module;

the recommendation information generation module is used for taking the information entropy of each candidate recommendation information, the information recommendation association factor and the historical recommendation information as input information of the information recommendation model to generate recommendation information.

In the embodiment of the specification, in each information recommendation process, the information entropy of the historical recommendation information is added as the input information of the model to control the generation of the subsequent recommendation information, and the relationship between the recommendation information can be controlled, so that the recommended information better meets the requirements of users.

On the basis of the foregoing embodiments, in some embodiments of the present specification, the apparatus further includes a location information determining module configured to: determining the position information of the recommendation information, and sending the position information to the recommendation information generation module, wherein the position information is the generation sequence of the recommendation information generated by the information recommendation model;

the recommendation information generation module is used for generating recommendation information by taking the position information of the recommendation information as input information of an information recommendation model.

In the embodiments of the present specification, the goal of optimizing the information recommendation model may include issuing appropriate recommendation information at a suitable location, for example, recommendation information that is more consistent with the user's expectation should appear at an earlier location. In each round of information recommendation process, the position information of the recommendation information can be added as input, and the generated recommendation information and the feedback of the user recommendation information can be used as training data to optimize the information recommendation model. In the training process of the information recommendation model, the position information of each piece of recommendation information can be used as input information for model training, so that the model finds the relation with the position information from the training data, the optimization direction of the model is controlled, the better recommendation information is output earlier, and the generation efficiency of the recommendation information is improved.

setting an output label, a non-output label and an output list;

storing the recommendation information into the output list;

In the embodiment of the specification, the flexible output of the model is realized by controlling the output of the model through the output label, the non-output label and the output list.

On the basis of the foregoing embodiments, in some embodiments of the present specification, the apparatus further includes a click data obtaining module configured to:

On the basis of the foregoing embodiments, in some embodiments of this specification, the recommendation information is tag information, and the apparatus further includes a tag information output module configured to:

In the embodiment of the specification, a mode of converting single-round generation of multiple labels and questions into single-theory multiple generation is adopted, for a user, the label seen by the user is three labels in each round, and when the model is actually predicted, the three labels are predicted three times to obtain three labels respectively. Through the form of single-round multi-time prediction, the characteristics of historical label information, information entropy, position information, output rules and the like of one round can be added in each round of prediction to control the recommendation of subsequent labels, so that the relation among the recommended labels can be controlled to meet the requirements of users, the flexible control of model output is realized, and the accuracy of label and question recommendation is improved.

It should be noted that the above-described apparatus may also include other embodiments according to the description of the method embodiment. The specific implementation manner may refer to the description of the above corresponding method embodiment, and is not described in detail herein.

An embodiment of the present specification further provides a data processing apparatus for generating recommendation information, including: at least one processor and a memory for storing processor-executable instructions, where the processor executes the instructions to implement the recommendation information generation method in the above embodiments, such as:

Embodiments of the present specification further provide a conversation robot, including: at least one processor and a memory for storing processor-executable instructions, the memory having an information recommendation model stored therein;

It should be noted that the above-mentioned processing device and the dialogue robot may also include other embodiments according to the description of the method embodiment. The specific implementation manner may refer to the description of the above corresponding method embodiment, and is not described in detail herein.

The recommendation information generating apparatus, processing device, or dialogue robot provided in the present specification can also be applied to various data analysis processing systems. The system or the apparatus or the processing device may include any one of the recommendation information generating apparatuses in the above embodiments. The system or apparatus or processing device may be a single server, or may include a server cluster, a system (including a distributed system), software (applications), an actual operation device, a logic gate device, a quantum computer, etc. using one or more of the methods or one or more of the embodiments of the present disclosure, and a terminal device incorporating necessary hardware for implementation. The system for checking for discrepancies may comprise at least one processor and a memory storing computer-executable instructions that, when executed by the processor, implement the steps of the method of any one or more of the embodiments described above.

The method embodiments provided by the embodiments of the present specification can be executed in a mobile terminal, a computer terminal, a server or a similar computing device. Taking an example of the present invention running on a server, fig. 5 is a block diagram of a hardware configuration of a recommendation information generation server in an embodiment of the present invention, and the server may be a recommendation information generation device or a session robot in the above-described embodiment. As shown in fig. 5, the server 10 may include one or more (only one shown) processors 100 (the processors 100 may include, but are not limited to, a processing device such as a microprocessor MCU or a programmable logic device FPGA, etc.), a memory 200 for storing data, and a transmission module 300 for communication functions. It will be understood by those skilled in the art that the structure shown in fig. 5 is merely illustrative and is not intended to limit the structure of the electronic device. For example, the server 10 may also include more or fewer components than shown in FIG. 5, and may also include other processing hardware, such as a database or multi-level cache, a GPU, or have a different configuration than shown in FIG. 5, for example.

The memory 200 may be used to store software programs and modules of application software, such as program instructions/modules corresponding to the recommendation information generation method in the embodiments of the present specification, and the processor 100 executes various functional applications and resource data updates by running the software programs and modules stored in the memory 200. Memory 200 may include high speed random access memory and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, memory 200 may further include memory located remotely from processor 100, which may be connected to a computer terminal through a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.

The transmission module 300 is used for receiving or transmitting data via a network. Specific examples of the network described above may include a wireless network provided by a communication provider of the computer terminal. In one example, the transmission module 300 includes a Network adapter (NIC) that can be connected to other Network devices through a base station so as to communicate with the internet. In one example, the transmission module 300 may be a Radio Frequency (RF) module, which is used for communicating with the internet in a wireless manner.

The foregoing description has been directed to specific embodiments of this disclosure. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.

The method or apparatus provided by the present specification and described in the foregoing embodiments may implement service logic through a computer program and record the service logic on a storage medium, where the storage medium may be read and executed by a computer, so as to implement the effect of the solution described in the embodiments of the present specification.

The storage medium may include a physical device for storing information, and typically, the information is digitized and then stored using an electrical, magnetic, or optical media. The storage medium may include: devices that store information using electrical energy, such as various types of memory, e.g., RAM, ROM, etc.; devices that store information using magnetic energy, such as hard disks, floppy disks, tapes, core memories, bubble memories, and usb disks; devices that store information optically, such as CDs or DVDs. Of course, there are other ways of storing media that can be read, such as quantum memory, graphene memory, and so forth.

The method or apparatus for generating recommendation information provided in the embodiments of the present specification may be implemented in a computer by a processor executing corresponding program instructions, for example, implemented in a PC end using a c + + language of a windows operating system, implemented in a linux system, or implemented in an intelligent terminal using android, iOS system programming languages, implemented in processing logic based on a quantum computer, or the like.

It should be noted that descriptions of the apparatus, the computer storage medium, and the system described above according to the related method embodiments may also include other embodiments, and specific implementations may refer to descriptions of corresponding method embodiments, which are not described in detail herein.

The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments may be referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the hardware + program class embodiment, since it is substantially similar to the method embodiment, the description is simple, and the relevant points can be referred to only the partial description of the method embodiment.

The embodiments of the present description are not limited to what must be consistent with industry communications standards, standard computer resource data updating and data storage rules, or what is described in one or more embodiments of the present description. Certain industry standards, or implementations modified slightly from those described using custom modes or examples, may also achieve the same, equivalent, or similar, or other, contemplated implementations of the above-described examples. The embodiments using the modified or transformed data acquisition, storage, judgment, processing and the like can still fall within the scope of the alternative embodiments of the embodiments in this specification.

In the 90 s of the 20 th century, improvements in a technology could clearly distinguish between improvements in hardware (e.g., improvements in circuit structures such as diodes, transistors, switches, etc.) and improvements in software (improvements in process flow). However, as technology advances, many of today's process flow improvements have been seen as direct improvements in hardware circuit architecture. Designers almost always obtain the corresponding hardware circuit structure by programming an improved method flow into the hardware circuit. Thus, it cannot be said that an improvement in the process flow cannot be realized by hardware physical modules. For example, a Programmable Logic Device (PLD), such as a Field Programmable Gate Array (FPGA), is an integrated circuit whose Logic functions are determined by programming the Device by a user. A digital system is "integrated" on a PLD by the designer's own programming without requiring the chip manufacturer to design and fabricate application-specific integrated circuit chips. Furthermore, nowadays, instead of manually making an integrated Circuit chip, such Programming is often implemented by "logic compiler" software, which is similar to a software compiler used in program development and writing, but the original code before compiling is also written by a specific Programming Language, which is called Hardware Description Language (HDL), and HDL is not only one but many, such as abel (advanced Boolean Expression Language), ahdl (alternate Language Description Language), traffic, pl (core unified Programming Language), HDCal, JHDL (Java Hardware Description Language), langue, Lola, HDL, laspam, hardsradware (Hardware Description Language), vhjhd (Hardware Description Language), and vhigh-Language, which are currently used in most common. It will also be apparent to those skilled in the art that hardware circuitry that implements the logical method flows can be readily obtained by merely slightly programming the method flows into an integrated circuit using the hardware description languages described above.

The controller may be implemented in any suitable manner, for example, the controller may take the form of, for example, a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro) processor, logic gates, switches, an Application Specific Integrated Circuit (ASIC), a programmable logic controller, and an embedded microcontroller, examples of which include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320, the memory controller may also be implemented as part of the control logic for the memory. Those skilled in the art will also appreciate that, in addition to implementing the controller as pure computer readable program code, the same functionality can be implemented by logically programming method steps such that the controller is in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, embedded microcontrollers and the like. Such a controller may thus be considered a hardware component, and the means included therein for performing the various functions may also be considered as a structure within the hardware component. Or even means for performing the functions may be regarded as being both a software module for performing the method and a structure within a hardware component.

The systems, devices, modules or units illustrated in the above embodiments may be implemented by a computer chip or an entity, or by a product with certain functions. One typical implementation device is a computer. In particular, the computer may be, for example, a personal computer, a laptop computer, a vehicle-mounted human-computer interaction device, a cellular telephone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

Although one or more embodiments of the present description provide method operational steps as described in the embodiments or flowcharts, more or fewer operational steps may be included based on conventional or non-inventive approaches. The order of steps recited in the embodiments is merely one manner of performing the steps in a multitude of orders and does not represent the only order of execution. When the device or the end product in practice executes, it can execute sequentially or in parallel according to the method shown in the embodiment or the figures (for example, in the environment of parallel processors or multi-thread processing, even in the environment of distributed resource data update). The terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, the presence of additional identical or equivalent elements in a process, method, article, or apparatus that comprises the recited elements is not excluded. The terms first, second, etc. are used to denote names, but not any particular order.

For convenience of description, the above devices are described as being divided into various modules by functions, and are described separately. Of course, when implementing one or more of the present description, the functions of each module may be implemented in one or more software and/or hardware, or a module implementing the same function may be implemented by a combination of multiple sub-modules or sub-units, etc. The above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units is only one logical division, and other divisions may be realized in practice, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.

The present invention is described 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 flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable resource data updating apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable resource data updating apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable resource data update apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable resource data update apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

In a typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.

The memory may include forms of volatile memory in a computer readable medium, Random Access Memory (RAM) and/or non-volatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). Memory is an example of a computer-readable medium.

Computer-readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), Digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage, graphene storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, a computer readable medium does not include a transitory computer readable medium such as a modulated data signal and a carrier wave.

As will be appreciated by one skilled in the art, one or more embodiments of the present description may be provided as a method, system, or computer program product. Accordingly, one or more embodiments of the present description may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, one or more embodiments of the present description may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

One or more embodiments of the present description may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. One or more embodiments of the present specification can also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.

The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments may be referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for system embodiments, since they are substantially similar to method embodiments, the description is relatively simple, and the relevant points can be referred to only part of the description of the method embodiments. In the description of the specification, reference to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the specification. In this specification, the schematic representations of the terms used above are not necessarily intended to refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.

The above description is merely exemplary of one or more embodiments of the present disclosure and is not intended to limit the scope of one or more embodiments of the present disclosure. Various modifications and alterations to one or more embodiments described herein will be apparent to those skilled in the art. Any modification, equivalent replacement, improvement or the like made within the spirit and principle of the present specification should be included in the scope of the claims.

Claims

1. A recommendation information generation method includes:

2. The method of claim 1, the outputting the recommendation information according to the output rule comprising:

3. The method of claim 1, further comprising:

and calculating the information entropy of the candidate recommendation information according to the historical recommendation information, and generating recommendation information by taking the information entropy of each candidate recommendation information, the information recommendation association factor and the historical recommendation information as input information of the information recommendation model.

4. The method of claim 1, further comprising:

5. The method of claim 1, the outputting the recommendation information according to the output rules of the information recommendation model, comprising:

setting an output label, a non-output label and an output list;

storing the recommendation information into the output list;

6. The method of claim 1, further comprising:

7. The method of claim 1, the recommendation information being tag information, the method further comprising:

8. A recommendation information generating apparatus comprising:

9. The apparatus of claim 8, wherein the recommendation information output module is specifically configured to:

10. The apparatus of claim 8, the apparatus further comprising an information entropy calculation module to: calculating the information entropy of the candidate recommendation information according to the historical recommendation information, and sending the calculated information entropy of each candidate recommendation information to the recommendation information generation module;

11. The apparatus of claim 8, the apparatus further comprising a location information determination module to: determining the position information of the recommendation information, and sending the position information to the recommendation information generation module, wherein the position information is the generation sequence of the recommendation information generated by the information recommendation model;

12. The apparatus of claim 8, wherein the recommendation information output module is specifically configured to:

setting an output label, a non-output label and an output list;

storing the recommendation information into the output list;

13. The apparatus of claim 8, further comprising a click data acquisition module to:

14. The apparatus of claim 12, wherein the recommendation information is tag information, the apparatus further comprising a tag information output module for:

15. A conversation robot, comprising: at least one processor and a memory for storing processor-executable instructions, the memory having an information recommendation model stored therein;

16. A data processing apparatus for recommendation information generation, comprising: at least one processor and a memory for storing processor-executable instructions, the processor implementing the method of any one of claims 1-7 when executing the instructions.