Information push method, device and electronic device

Abstract

The disclosed embodiments disclose an information push method, device, and electronic device. A specific implementation of the method includes: importing first push-related information of first adjustment method information and first candidate information into a target model to obtain first scores corresponding to each first adjustment method information, wherein the number of the first adjustment method information is at least two; selecting target first adjustment method information from at least two first adjustment method information according to each first score; adjusting the first to-be-adjusted representation value of the first candidate information using the target first adjustment method information to generate a first adjusted representation value; and pushing the first candidate information based on the first adjusted representation value. Thus, a new information push method is provided.

Description

Technical Field

[0001] The present disclosure relates to the field of Internet technology, and in particular to an information push method, device and electronic device. Background Art

[0002] With the development of the Internet, more and more users use terminal devices to browse various types of information. For example, users can browse information streams that are constantly updated over time through terminal devices. Users may be interested in one or more information in the information stream.

[0003] Generally, the information stream that a user browses may be pushed by a server to a terminal used by the user. The accuracy of the information pushed by the server may affect the amount of valid information obtained by the user within a certain period of time, thereby affecting the length of time the user browses the information stream. Summary of the invention

[0004] This disclosure section is provided to introduce concepts in a brief form, which will be described in detail in the detailed description section below. This disclosure section is not intended to identify key features or essential features of the claimed technical solution, nor is it intended to limit the scope of the claimed technical solution.

[0005] In a first aspect, an embodiment of the present disclosure provides an information push method, the method comprising: importing first push-related information of first adjustment method information and first candidate information into a target model to obtain first scores corresponding to each first adjustment method information, wherein the number of the first adjustment method information is at least two; according to each first score, selecting target first adjustment method information from at least two first adjustment method information; using the target first adjustment method information, adjusting the first to-be-adjusted representation value of the first candidate information to generate a first adjusted representation value; and pushing the first candidate information based on the first adjusted representation value.

[0006] In a second aspect, an embodiment of the present disclosure provides an information pushing device, including: an importing unit, used to import first adjustment method information and first push-related information of first candidate information into a target model to obtain first scores corresponding to each first adjustment method information, wherein the number of the first adjustment method information is at least two; a selecting unit, used to select target first adjustment method information from at least two first adjustment method information according to each first score; a generating unit, used to adjust the first to-be-adjusted representation value of the first candidate information using the target first adjustment method information to generate a first adjusted representation value; and a pushing unit, used to push the first candidate information based on the first adjusted representation value.

[0007] In a third aspect, the disclosed embodiment provides a model training method, the method comprising: importing second adjustment method information and second push-related information of second candidate information into a model to be trained, and obtaining second scores corresponding to each second adjustment method information, wherein the number of second adjustment method information is at least two; selecting target second adjustment method information from at least two second adjustment method information according to each second score; adjusting the second to-be-adjusted representation value of the second candidate information using the target second adjustment method information to generate a second adjusted representation value; and adjusting the parameters of the model to be trained based on the push effect information, wherein the push effect information is used to characterize the push effect of pushing the second candidate information based on the second adjusted representation value.

[0008] In a fourth aspect, an embodiment of the present disclosure provides a model training device, comprising: an import module, used to import second adjustment method information and second push-related information of second candidate information into a model to be trained, and obtain a second score corresponding to each second adjustment method information, wherein the number of the second adjustment method information is at least two; a selection module, used to select target second adjustment method information from at least two second adjustment method information according to each second score; a generation module, used to use the target second adjustment method information to adjust the second to-be-adjusted representation value of the second candidate information, and generate a second adjusted representation value; an adjustment module, used to adjust the parameters of the model to be trained based on push effect information, wherein the push effect information is used to characterize the push effect of pushing the second candidate information based on the second adjusted representation value.

[0009] In a fifth aspect, an embodiment of the present disclosure provides an electronic device, comprising: one or more processors; a storage device for storing one or more programs, when the one or more programs are executed by the one or more processors, the one or more processors implement the information push method as described in the first aspect.

[0010] In a sixth aspect, an embodiment of the present disclosure provides a computer-readable medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the information push method as described in the first aspect.

[0011] The information push method, device and electronic device provided by the embodiments of the present disclosure import the first push related information and the first adjustment method information as input into the target model, and then the target model directly outputs the first score of the first adjustment method information, so that the first adjustment method information matching the first push related information can be determined, and the first to-be-adjusted characterization value can be adjusted for the first push related information, thereby improving the accuracy of the generated first adjusted characterization value, and further improving the accuracy of the push. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the accompanying drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and the originals and elements are not necessarily drawn to scale.

[0013] Figure 1 is a flow chart of an embodiment of the information push method according to the present disclosure;

[0014] Figure 2 is a flow chart of the model training steps according to the present disclosure;

[0015] Figure 3 is a flowchart of an embodiment of a model training method according to the present disclosure;

[0016] Figure 4 is a structural schematic diagram of an embodiment of an information push device according to the present disclosure;

[0017] Figure 5 is a structural schematic diagram of an embodiment of a model training device according to the present disclosure;

[0018] Figure 6 is an exemplary system architecture to which the information push method of an embodiment of the present disclosure can be applied;

[0019] Figure 7 It is a schematic diagram of the basic structure of an electronic device provided according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0020] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein, which are instead provided for a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0021] It should be understood that the various steps described in the method embodiments of the present disclosure may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this respect.

[0022] The term "including" and its variations used herein are open inclusions, i.e., "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0023] It should be noted that the concepts such as "first" and "second" mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0024] It should be noted that the modifications of "one" and "plurality" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0025] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0026] Please refer to Figure 1 , which shows the process of an embodiment of the information push method according to the present disclosure. Figure 1 The information push method shown includes the following steps:

[0027] Step 101: import first adjustment method information and first push related information of first candidate information into a target model to obtain a first score corresponding to the first adjustment method information.

[0028] In this embodiment, the execution subject of the information push method may import the first adjustment method information and the first push related information of the first candidate information into the target model to obtain the first score corresponding to the first adjustment method information.

[0029] In this embodiment, the first adjustment method information may be used to indicate a method for adjusting the first to-be-adjusted representation value of the first candidate information. Here, the number of the first adjustment method information may be at least two.

[0030] In this embodiment, the first candidate information may be information for push. The type of the first candidate information is not limited here.

[0031] In this embodiment, the first push related information may include information related to pushing the first candidate information. The information items specifically included in the first push related information may be set according to actual application scenarios and are not limited here.

[0032] In this embodiment, the target model can be used to characterize the correspondence between the first push related information, the first adjustment method information, and the first score. It can be understood that for each adjustment method, the target model will generate a first score for the first adjustment method information.

[0033] In this embodiment, the target model may be a model constructed based on a neural network. The specific structure of the target model may be set according to the actual application scenario and is not limited here.

[0034] Step 102 : selecting target first adjustment method information from at least two pieces of first adjustment method information according to the first scores.

[0035] In this embodiment, the execution entity may select target first adjustment method information from at least two pieces of first adjustment method information according to each first score.

[0036] In this embodiment, the execution subject may select a first score from at least two first scores generated by the target model as the target first score. Then, the first adjustment method information corresponding to the target first score is used as the target first adjustment method information. In other words, the first adjustment method information may be selected from at least two first adjustment method information based on the first score as the target first adjustment method information.

[0037] Step 103: adjusting the first to-be-adjusted representation value of the first candidate information according to the target first adjustment mode information to generate a first adjusted representation value.

[0038] In this embodiment, the execution subject may adjust the first to-be-adjusted representation value of the first candidate information according to the target first adjustment mode information to generate a first adjusted representation value.

[0039] As an example, the first characterization value to be adjusted may be 10. The first adjustment mode information may indicate an increase of 2, and then the sum of the first characterization value to be adjusted 10 and 2 may be used as the first adjusted characterization value, that is, 12.

[0040] In this embodiment, the characterization value of the first candidate information can be used as a basis for judging whether to push the first candidate information. As an example, when information is pushed, the first candidate information related to the user characteristics or the user search terms can be first recalled based on the content relevance, and then the characterization value of the first candidate information can be referred to to determine whether to display each of the recalled first candidate information, or to determine the ranking of the first candidate information to be displayed.

[0041] Step 104: Push the first candidate information based on the first adjusted representation value.

[0042] In this embodiment, the execution entity may push the first candidate information based on the first adjusted representation value.

[0043] In this embodiment, the first candidate information is pushed based on the first adjusted characterization value evaluation information, which can be implemented in various ways, which are not limited here.

[0044] In some embodiments, at least one of the following may be included but is not limited to: determining whether to display the first candidate information based on the first adjusted representation value, or determining the position of the first candidate information to be displayed in the sequence of information to be pushed.

[0045] As an example, after recalling a number of candidate information according to the search term or user attribute, a predetermined number of candidate information may be selected in descending order of the first characterization value, and the selected candidate information may be the candidate information to be displayed. Then, the candidate information to be displayed may be sorted, and the basis for the sorting may include the first adjusted characterization value.

[0046] It should be noted that the method provided in this embodiment imports the first push-related information and the first adjustment method information as input into the target model, and then the target model directly outputs the first score of the first adjustment method information, so that the first adjustment method information matching the first push-related information can be determined, and the first to-be-adjusted representation value can be adjusted according to the first push-related information, thereby improving the accuracy of the generated first adjusted representation value, and further improving the accuracy of the push.

[0047] In some embodiments, the above step 102 may include: determining the first adjustment method information corresponding to the highest first score as the target first adjustment method information.

[0048] It should be noted that by determining the first adjustment method information corresponding to the highest first score as the target first adjustment method information, the accuracy of the adjustment can be improved, thereby improving the accuracy of the generated first adjusted representation value, thereby improving the accuracy of information push.

[0049] In some embodiments, the first push related information belongs to push related information, that is, the first push related information may inherit related conditions in the push related information.

[0050] In some embodiments, specific items in the push-related information can be set according to actual application scenarios and are not limited here.

[0051] In some embodiments, the push-related information may include but is not limited to at least one of the following: push basis information and push history information.

[0052] It should be noted that by using the push history information as the push related information, the push history information can be used as a basis for judging the response to the adjustment method information. Thus, the push history information of the candidate information can be referred to to determine the adjustment range of the characterization value to be adjusted, thereby improving the accuracy of adjusting the characterization value to be adjusted.

[0053] In some embodiments, the push basis information may include but is not limited to at least one of the following: attribute information of the candidate information, attribute information of the owner of the candidate information.

[0054] In some embodiments, the attribute information of the candidate information may include but is not limited to at least one of the following: the representation value of the candidate information, the form type of the candidate information (such as picture, text or video, etc.), and the field type of the candidate information (such as education, sports or literature, etc.).

[0055] In some embodiments, the attribute information of the candidate information owner may include but is not limited to at least one of the following: the name and type of the candidate information owner. The candidate information owner may include the owner of the candidate information.

[0056] In some embodiments, the push history information may include, but is not limited to, at least one of the following: push process history information, and history operation information performed on the pushed candidate information.

[0057] It should be noted that the push-related information includes real-time determined push history data, and the characterization value to be adjusted can be adjusted according to the real-time push environment, thereby improving the real-time performance of the adjusted characterization value, that is, the characterization value can be adjusted to adapt to the latest push environment, thereby improving the accuracy of the adjustment.

[0058] In some embodiments, the push process history information may include but is not limited to at least one of the following: the start time of push, the duration from the start time of push to the current time, the number of pushes, and the total cost since the push of the candidate information.

[0059] In some embodiments, the historical operation information performed on the pushed candidate information may include but is not limited to at least one of the following: a user portrait of the user who operates the candidate information, the first operation number of the first operation performed on the candidate information, and the second operation number of the second operation performed on the candidate information.

[0060] Here, the operation content of the above-mentioned first operation can be set according to the actual scenario and is not limited here; as an example, the first operation can be an operation of triggering candidate information so as to browse the candidate information.

[0061] Here, the operation content of the second operation can be set according to the actual application scenario and is not limited here; as an example, the second operation can be an operation that triggers a preset control in the candidate information.

[0062] In some application scenarios, the first push-related information may belong to push-related information. The second push-related information may also belong to push-related information.

[0063] In some embodiments, the first push related information includes at least one of the following: first push basis information and first push history information; wherein the first push basis information includes at least one of the following: attribute information of the first candidate information, attribute information of the owner of the first candidate information; the first push history information includes at least one of the following: first push process history information, historical operation information of the first pair of pushed candidate information.

[0064] In some embodiments, the second push related information includes at least one of the following: second push basis information and second push history information; wherein the second push basis information includes at least one of the following: attribute information of the second candidate information, attribute information of the owner of the second candidate information; the second push history information includes at least one of the following: second push process history information, historical operation information of the second pair of pushed candidate information.

[0065] In some embodiments, at least two pieces of adjustment mode information may be set according to actual conditions and are not limited here.

[0066] As an example, the adjustment range and adjustment amplitude are greater than or equal to -10 and less than or equal to 10. The minimum adjustment unit is 0.1. Therefore, 200 adjustment amplitudes can be set. When the adjustment amplitude is a positive number, it means to increase the characterization value to be adjusted; when the adjustment amplitude is a negative number, it means to reduce the characterization value to be adjusted.

[0067] In some embodiments, the target model can be obtained through a model training step. Here, the model training step can be obtained through Figure 2 The process shown is implemented. Figure 2 The process shown may include:

[0068] Step 201: Import the second adjustment method information and the second push related information of the second candidate information into the model to be trained.

[0069] It should be noted that the second adjustment method information in the model training step is Figure 1The first adjustment mode information in the corresponding information push method is the adjustment mode information in different stages. The contents of the first adjustment mode information and the second adjustment mode information may be the same; as an example, there may be two first adjustment mode information and two second adjustment mode information, and the first adjustment mode information may be plus 1 and minus 1 respectively, and the second adjustment mode information may be plus 1 and minus 1 respectively.

[0070] It should be noted that the second candidate information in the model training step is Figure 1 The first candidate information in the corresponding information push method is the candidate information in different stages. The information content of the first candidate information and the second candidate information may be the same. In some application scenarios, candidate information A can be used as the second candidate information in the model training stage, or as the first candidate information after the model training is completed.

[0071] Here, the second push-related information may be used to indicate information related to the push of the second candidate information.

[0072] Here, the model to be trained can be an untrained initial model or a model that has been trained but not completed.

[0073] Here, the model to be trained can be used to characterize the correspondence between the first push related information, the first adjustment method information, and the first score. It can be understood that for each adjustment method, the model to be trained will generate a first score for the first adjustment method information.

[0074] Here, the above-mentioned model to be trained can be a model built based on a neural network. The specific structure of the model to be trained can be set according to the actual application scenario and is not limited here.

[0075] Step 202, selecting target second adjustment method information from at least two second adjustment method information according to each second score,

[0076] Here, the method of selecting the target second adjustment method information according to the second score may be the same as or different from the method of selecting the target adjustment method information according to the first score.

[0077] Step 203: Use the target second adjustment method information to adjust the second to-be-adjusted representation value of the second candidate information to generate a second adjusted representation value.

[0078] Here, the second in the second characterization value to be adjusted is for consistency with the second expression in the second candidate information. It can be understood that the second candidate information does not have the first characterization value to be adjusted.

[0079] Step 204: Adjust the parameters of the model to be trained based on the push effect information.

[0080] Here, the push effect information is used to characterize the push effect of pushing the second candidate information based on the second adjusted characterization value.

[0081] In this embodiment, the specific representation method of the above push effect information can be set according to the actual application scenario and is not limited here.

[0082] In some application scenarios, the push effect information can be represented by the ratio of the first operation times to the display times. As an example, the first operation can be an operation of the user clicking on the second candidate information, or an operation of the user clicking on a preset control in the second candidate information.

[0083] In this embodiment, the parameters in the model to be trained may also include but are not limited to at least one of the following: a weight in the model to be trained and a bias item in the model to be trained.

[0084] In this embodiment, the parameters of the model to be trained are adjusted based on the push effect information, which can be achieved in various ways, which are not limited here.

[0085] As an example, back propagation, gradient descent, etc. can be used to adjust the parameters of the training model.

[0086] In some embodiments, the stopping condition of model training can be set according to the actual application scenario and is not limited here. As an example, the stopping condition of model training may include but is not limited to at least one of the following: the model parameter update amplitude is less than the preset update amplitude threshold, and the model update number is greater than the preset update number threshold.

[0087] In this embodiment, the model to be trained after the parameters are adjusted can be used as the target model, or the model to be trained after the parameters are adjusted can be trained again to obtain the target model.

[0088] It should be noted that the model training method provided in this embodiment imports push-related information of the model to be trained, which can be used to characterize the push environment of the candidate information. The scoring of the adjustment method information for the push environment will be based on the actual push effect information. Adjusting the parameters of the model to be trained can change the scoring in the direction of improving the promotion effect evaluation. Therefore, it is possible to avoid calculating the correlation between the characterization value and the push-related information or the correlation between the characterization value and the push effect from the content. The adjustment method information can be judged directly from the perspective of the push effect information, thereby reducing the complexity of determining more accurate adjustment method information.

[0089] In some embodiments, the above step 202 may include: selecting a target second score from the second scores by a random selection method with a set selection probability; and determining the second adjustment method information corresponding to the selected target second score as the target second adjustment method information.

[0090] In some embodiments, the probability of selection can be set for each score. There are various ways to set the probability of selection.

[0091] In some application scenarios, the specific value of the first preset probability can be set according to the actual application scenario and is not limited here. As an example, the first preset probability can be set to 90%. The first preset probability is less than 1.

[0092] Here, the probability of determining the highest second score as the target second score is the first preset probability. It can be understood that, in general, the number of the highest scores is one.

[0093] It should be noted that by adopting a random calculation method to set the selection probability, the exploratory nature of the information on the selection of various adjustment methods can be increased, thus avoiding the situation where the model falls into a local optimum by selecting the second highest score every time.

[0094] In some application scenarios, the selection probabilities of various non-highest second answers may be equal or unequal.

[0095] As an example, the probability of selection corresponding to the probability corresponding to the highest score can be set as the first preset probability. The ratio of the difference between 1 and the first preset probability and the number of scores of non-highest scores is determined as the second preset probability. The second preset probability can be used as the probability of selection of each non-highest score. In this case, the probability of selection of each non-highest second score is equal.

[0096] In other words, if the product of the number of highest scores and the first preset probability is taken as the first product, the probability of determining the second answer other than the highest score as the target second score is taken as the second preset probability. The product of the number of other answers other than the highest score and the second preset probability is taken as the second product. The sum of the first product and the second product is equal to 1.

[0097] As an example, if there are 10 responses, 10 different selection probabilities may be set; the sum of these 10 selection probabilities is 1. In this case, the selection probabilities of the non-highest second scores are not equal.

[0098] In some embodiments, the model training step includes: increasing the first preset probability based on an increase in the number of model training updates.

[0099] Here, as the model training progresses, the above-mentioned first preset probability can be improved, that is, the probability of selecting the highest second reply can be improved. In this way, the convergence speed of the model can be improved, that is, as the model training progresses, the accuracy of the model is generally improving, that is, the accuracy of the second score is also constantly improving. In this case, the exploration of other adjustment methods can be reduced, and the model parameters can be improved based on the highest second score. When the model is trained using real data and the trained model is continuously applied to real scenes, the push effect of the real scene can be improved.

[0100] In some embodiments, the push effect information may include a push effect representation value. The above step 204 may include: determining a loss function value according to a change in the push effect information of the second candidate information; and adjusting parameters of the model to be trained according to the loss function value.

[0101] In some application scenarios, the push effect information may include a push effect representation value. As an example, the push effect representation value may be a ratio of the number of impressions to the number of clicks. The more clicks there are, the smaller the push effect representation value is, indicating a better push effect.

[0102] As an example, the square of the push effect characterization value may be used as the loss function value. Alternatively, the product of the square of the push effect characterization value and a preset attenuation coefficient may be used as the loss function value.

[0103] As an example, the push effect characterization value of this time may be compared with the push effect characterization value of the previous time. If the push effect characterization value increases, a loss function may be generated according to the highest score.

[0104] It should be noted that if the push effect representation value increases, it means that the push effect is reduced by adopting the target adjustment method, and the score corresponding to the target adjustment method information can be reduced. Therefore, the score corresponding to the target adjustment method can be determined as the loss function value. Based on the parameters of the model to be trained adjusted by this loss function value, the score corresponding to the target adjustment method information can be gradually reduced.

[0105] In some embodiments, the above step 204 includes: adjusting the parameters of the model to be trained according to the push effect information corresponding to the second candidate information set, wherein the second candidate information set includes at least two second candidate information to be pushed.

[0106] It should be noted that, by using relevant data of multiple second candidate information to train the model to be trained, the training model can learn the relevant characteristics of the second candidate information set space, so that when using the model to be trained to process a single second candidate information, the processing method can refer to the overall situation of the second candidate information set space to process the single second candidate information, that is, processing a certain second candidate information with reference to the overall situation of the second candidate information, so that the representation value of the processed second candidate information can match the above-mentioned overall situation, that is, the accuracy of the adjustment is improved.

[0107] Please refer to Figure 3 The present disclosure provides a model training method, which may include Figure 3 The process shown. Figure 3 The process shown may include step 301 , step 302 , step 303 and step 304 .

[0108] Step 301: Import the second adjustment method information and the second push related information of the second candidate information into the model to be trained.

[0109] It should be noted that the second adjustment method information in the model training method is Figure 1 The first adjustment mode information in the corresponding information push method is the adjustment mode information in different stages. The contents of the first adjustment mode information and the second adjustment mode information may be the same; as an example, there may be two first adjustment mode information and two second adjustment mode information, and the first adjustment mode information may be plus 1 and minus 1 respectively, and the second adjustment mode information may be plus 1 and minus 1 respectively.

[0110] It should be noted that the second candidate information in the model training method is Figure 1 The first candidate information in the corresponding information push method is the candidate information in different stages. The information content of the first candidate information and the second candidate information may be the same. In some application scenarios, candidate information A can be used as the second candidate information in the model training stage, or as the first candidate information after the model training is completed.

[0111] Here, the second push-related information may be used to indicate information related to the push of the second candidate information.

[0112] Here, the model to be trained can be an untrained initial model or a model that has been trained but not completed.

[0113] Here, the model to be trained can be used to characterize the correspondence between the first push related information, the first adjustment method information, and the first score. It can be understood that for each adjustment method, the model to be trained will generate a first score for the first adjustment method information.

[0114] Here, the above-mentioned model to be trained can be a model built based on a neural network. The specific structure of the model to be trained can be set according to the actual application scenario and is not limited here.

[0115] Step 302, selecting target second adjustment method information from at least two second adjustment method information according to each second score,

[0116] Here, the method of selecting the target second adjustment method information according to the second score may be the same as or different from the method of selecting the target adjustment method information according to the first score.

[0117] Step 303: Use the target second adjustment method information to adjust the second to-be-adjusted representation value of the second candidate information to generate a second adjusted representation value.

[0118] Here, the second in the second characterization value to be adjusted is for consistency with the second expression in the second candidate information. It can be understood that the second candidate information does not have the first characterization value to be adjusted.

[0119] Step 304: Adjust the parameters of the model to be trained based on the push effect information.

[0120] Here, the push effect information is used to characterize the push effect of pushing the second candidate information based on the second adjusted characterization value.

[0121] In this embodiment, the specific representation method of the above push effect information can be set according to the actual application scenario and is not limited here.

[0122] In some application scenarios, the push effect information can be represented by the ratio of the first operation times to the display times. As an example, the first operation can be an operation of the user clicking on the second candidate information, or an operation of the user clicking on a preset control in the second candidate information.

[0123] In this embodiment, the parameters in the model to be trained may also include but are not limited to at least one of the following: a weight in the model to be trained and a bias item in the model to be trained.

[0124] In this embodiment, the parameters of the model to be trained are adjusted based on the push effect information, which can be achieved in various ways, which are not limited here.

[0125] As an example, back propagation, gradient descent, etc. can be used to adjust the parameters of the training model.

[0126] In some embodiments, the stopping condition of model training can be set according to the actual application scenario and is not limited here. As an example, the stopping condition of model training may include but is not limited to at least one of the following: the model parameter update amplitude is less than the preset update amplitude threshold, and the model update number is greater than the preset update number threshold.

[0127] In this embodiment, the model to be trained after the parameters are adjusted can be used as the target model, or the model to be trained after the parameters are adjusted can be trained again to obtain the target model.

[0128] It should be noted that the model training method provided in this embodiment imports push-related information of the model to be trained, which can be used to characterize the push environment of the candidate information. The scoring of the adjustment method information for the push environment will be based on the actual push effect information. Adjusting the parameters of the model to be trained can change the scoring in the direction of improving the promotion effect evaluation. Therefore, it is possible to avoid calculating the correlation between the characterization value and the push-related information or the correlation between the characterization value and the push effect from the content. The adjustment method information can be judged directly from the perspective of the push effect information, thereby reducing the complexity of determining more accurate adjustment method information.

[0129] In some embodiments, the above step 302 may include: selecting a target second score from the second scores by a random selection method with a set selection probability; and determining the second adjustment method information corresponding to the selected target second score as the target second adjustment method information.

[0130] In some embodiments, the probability of selection can be set for each score. There are various ways to set the probability of selection.

[0131] In some application scenarios, the specific value of the first preset probability can be set according to the actual application scenario and is not limited here. As an example, the first preset probability can be set to 90%. The first preset probability is less than 1.

[0132] Here, the probability of determining the highest second score as the target second score is the first preset probability. It can be understood that, in general, the number of the highest scores is one.

[0133] It should be noted that by adopting a random calculation method to set the selection probability, the exploratory nature of the information on the selection of various adjustment methods can be increased, thus avoiding the situation where the model falls into a local optimum by selecting the second highest score every time.

[0134] In some application scenarios, the selection probabilities of various non-highest second answers may be equal or unequal.

[0135] As an example, the probability of selection corresponding to the probability corresponding to the highest score can be set as the first preset probability. The ratio of the difference between 1 and the first preset probability and the number of scores of non-highest scores is determined as the second preset probability. The second preset probability can be used as the probability of selection of each non-highest score. In this case, the probability of selection of each non-highest second score is equal.

[0136] In other words, if the product of the number of highest scores and the first preset probability is taken as the first product, the probability of determining the second answer other than the highest score as the target second score is taken as the second preset probability. The product of the number of other answers other than the highest score and the second preset probability is taken as the second product. The sum of the first product and the second product is equal to 1.

[0137] As an example, if there are 10 responses, 10 different selection probabilities may be set; the sum of these 10 selection probabilities is 1. In this case, the selection probabilities of the non-highest second scores are not equal.

[0138] In some embodiments, the model training method includes: increasing the first preset probability based on an increase in the number of model training updates.

[0139] Here, as the model training progresses, the above-mentioned first preset probability can be improved, that is, the probability of selecting the highest second reply can be improved. In this way, the convergence speed of the model can be improved, that is, as the model training progresses, the accuracy of the model is generally improving, that is, the accuracy of the second score is also constantly improving. In this case, the exploration of other adjustment methods can be reduced, and the model parameters can be improved based on the highest second score. When the model is trained using real data and the trained model is continuously applied to real scenes, the push effect of the real scene can be improved.

[0140] In some embodiments, the push effect information may include a push effect representation value. The above step 304 may include: determining a loss function value according to a change in the push effect information of the second candidate information; and adjusting parameters of the model to be trained according to the loss function value.

[0141] In some application scenarios, the push effect information may include a push effect representation value. As an example, the push effect representation value may be a ratio of the number of impressions to the number of clicks. The more clicks there are, the smaller the push effect representation value is, indicating a better push effect.

[0142] As an example, the square of the push effect characterization value may be used as the loss function value. Alternatively, the product of the square of the push effect characterization value and a preset attenuation coefficient may be used as the loss function value.

[0143] As an example, the push effect characterization value of this time may be compared with the push effect characterization value of the previous time. If the push effect characterization value increases, a loss function may be generated according to the highest score.

[0144] It should be noted that if the push effect representation value increases, it means that the push effect is reduced by adopting the target adjustment method, and the score corresponding to the target adjustment method information can be reduced. Therefore, the score corresponding to the target adjustment method can be determined as the loss function value. Based on the parameters of the model to be trained adjusted by this loss function value, the score corresponding to the target adjustment method information can be gradually reduced.

[0145] In some embodiments, the above step 304 includes: adjusting the parameters of the model to be trained according to the push effect information corresponding to the second candidate information set, wherein the second candidate information set includes at least two second candidate information to be pushed.

[0146] It should be noted that, by using relevant data of multiple second candidate information to train the model to be trained, the training model can learn the relevant characteristics of the second candidate information set space, so that when using the model to be trained to process a single second candidate information, the processing method can refer to the overall situation of the second candidate information set space to process the single second candidate information, that is, processing a certain second candidate information with reference to the overall situation of the second candidate information, so that the representation value of the processed second candidate information can match the above-mentioned overall situation, that is, the accuracy of the adjustment is improved.

[0147] Further references Figure 4 As an implementation of the methods shown in the above figures, the present disclosure provides an embodiment of an information push device, which is similar to Figure 1 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0148] like Figure 4 As shown, the information push device of this embodiment includes: an import unit 401, a selection unit 402, a generation unit 403 and a push unit 404. The import unit is used to import the first push-related information of the first adjustment method information and the first candidate information into the target model to obtain the first score corresponding to each first adjustment method information, wherein the number of the first adjustment method information is at least two; the selection unit is used to select the target first adjustment method information from at least two first adjustment method information according to each first score; the generation unit is used to adjust the first to-be-adjusted representation value of the first candidate information by using the target first adjustment method information to generate a first adjusted representation value; the push unit is used to push the first candidate information based on the first adjusted representation value.

[0149] In this embodiment, the specific processing of the importing unit 401, the selecting unit 402, the generating unit 403 and the pushing unit 404 of the information pushing device and the technical effects thereof can be referred to in Figure 1 The relevant descriptions of step 101, step 102, step 103 and step 104 in the corresponding embodiment are not repeated here.

[0150] In some embodiments, selecting target first adjustment method information from at least two first adjustment method information according to the first scores includes: determining the first adjustment method information corresponding to the highest first score as the target first adjustment method information.

[0151] In some embodiments, the target model is obtained through the following model training steps, wherein the model training steps include: importing the second adjustment method information and the second push-related information of the second candidate information into the model to be trained to obtain a second score corresponding to each second adjustment method information, wherein the number of the second adjustment method information is at least two; according to each second score, selecting the target second adjustment method information from at least two second adjustment method information; using the target second adjustment method information, adjusting the second to-be-adjusted representation value of the second candidate information to generate a second adjusted representation value; based on the push effect information, adjusting the parameters of the model to be trained, wherein the push effect information is used to characterize the push effect of pushing the second candidate information based on the second adjusted representation value.

[0152] In some embodiments, the second push related information includes at least one of the following: second push basis information and second push history information;

[0153] The second push basis information includes at least one of the following: attribute information of the second candidate information, attribute information of the owner of the second candidate information;

[0154] The second push history information includes at least one of the following: second push process history information, second historical operation information on the candidate information pushed. In some embodiments, the selecting the target second adjustment method information from at least two second adjustment method information according to each second score includes: selecting the target second score from the second scores by a random selection method with a set selection probability, wherein the selection probability of the highest second score is the first preset probability; and determining the second adjustment method information corresponding to the selected target second score as the target second adjustment method information.

[0155] In some embodiments, the model training step includes: increasing the first preset probability based on an increase in the number of model training updates.

[0156] In some embodiments, adjusting the parameters of the model to be trained based on the push effect information includes: determining a loss function value according to a change in the push effect information of the second candidate information; and adjusting the parameters of the model to be trained according to the loss function value.

[0157] In some embodiments, adjusting the parameters of the model to be trained based on the push effect information includes: adjusting the parameters of the model to be trained according to the push effect information corresponding to the second candidate information set, wherein the second candidate information set includes at least two second candidate information to be pushed.

[0158] Further references Figure 5 As an implementation of the methods shown in the above figures, the present disclosure provides an embodiment of an information push device, which is similar to Figure 1 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0159] like Figure 5 As shown, the information push device of this embodiment includes: an import module 501, a selection module 502, a generation module 503 and an adjustment module 504. The import unit is used to import the first adjustment method information and the first push-related information of the first candidate information into the model to be trained to obtain the second score corresponding to each second adjustment method information, wherein the number of the second adjustment method information is at least two; the selection module is used to select the target second adjustment method information from at least two second adjustment method information according to each second score; the generation module is used to adjust the second to-be-adjusted characterization value of the second candidate information by using the target second adjustment method information to generate a second adjusted characterization value; the adjustment module is used to adjust the parameters of the model to be trained based on the push effect information, wherein the push effect information is used to characterize the push effect of pushing the second candidate information based on the second adjusted characterization value.

[0160] In this embodiment, the specific processing of the import module 501, the selection module 502, the generation module 503 and the adjustment module 504 of the information push device and the technical effects thereof can be referred to in Figure 3 The relevant descriptions of step 301, step 302, step 303 and step 304 in the corresponding embodiment are not repeated here.

[0161] In some embodiments, the second push related information includes at least one of the following: second push basis information and second push history information; wherein the second push basis information includes at least one of the following: attribute information of the second candidate information, attribute information of the owner of the second candidate information; the second push history information includes at least one of the following: second push process history information, historical operation information of the second pair of pushed candidate information.

[0162] In some embodiments, the selecting target second adjustment method information from at least two second adjustment method information based on each second score includes: selecting the target second score from the second scores using a random selection method with a set selection probability, wherein the selection probability of the highest second score is a first preset probability; and determining the second adjustment method information corresponding to the selected target second score as the target second adjustment method information.

[0163] In some embodiments, the model training step includes: increasing the first preset probability based on an increase in the number of model training updates.

[0164] Please refer to Figure 6 , Figure 6 An exemplary system architecture is shown in which the information push method according to an embodiment of the present disclosure can be applied.

[0165] like Figure 6 As shown, the system architecture may include terminal devices 601, 602, 603, a network 604, and a server 605. The network 604 is used to provide a medium for communication links between the terminal devices 601, 602, 603 and the server 605. The network 606 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0166] The terminal devices 601, 602, and 603 can interact with the server 605 through the network 604 to receive or send messages, etc. Various client applications, such as web browser applications, search applications, and news information applications, can be installed on the terminal devices 601, 602, and 603. The client applications in the terminal devices 601, 602, and 603 can receive user instructions and perform corresponding functions according to the user instructions, such as adding corresponding information to the information according to the user instructions.

[0167] Terminal devices 601, 602, and 603 can be hardware or software. When terminal devices 601, 602, and 603 are hardware, they can be various electronic devices with display screens and supporting web browsing, including but not limited to smart phones, tablet computers, e-book readers, MP3 players (Moving Picture Experts Group Audio Layer III, Moving Picture Experts Group Audio Layer 3), MP4 (Moving Picture Experts Group Audio Layer IV, Moving Picture Experts Group Audio Layer 4) players, laptop computers, and desktop computers, etc. When terminal devices 601, 602, and 603 are software, they can be installed in the electronic devices listed above. It can be implemented as multiple software or software modules (for example, software or software modules used to provide distributed services), or it can be implemented as a single software or software module. No specific limitation is made here.

[0168] The server 605 may be a server that provides various services, such as receiving information acquisition requests sent by the terminal devices 601, 602, and 603, acquiring display information corresponding to the information acquisition requests in various ways according to the information acquisition requests, and sending the relevant data of the display information to the terminal devices 601, 602, and 603.

[0169] It should be noted that the information push method provided in the embodiment of the present disclosure can be executed by a terminal device, and accordingly, the information push device can be set in the terminal devices 601, 602, and 603. In addition, the information push method provided in the embodiment of the present disclosure can also be executed by a server 605, and accordingly, the information push device can be set in the server 605.

[0170] It should be understood that Figure 6 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to the implementation requirements.

[0171] Reference below Figure 7 , which shows an electronic device (eg, Figure 6 The terminal device in the embodiment of the present disclosure may include but is not limited to mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 7 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0172] like Figure 7 As shown, the electronic device may include a processing device (e.g., a central processing unit, a graphics processor, etc.) 701, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 702 or a program loaded from a storage device 708 into a random access memory (RAM) 703. In the RAM 703, various programs and data required for the operation of the electronic device 700 are also stored. The processing device 701, the ROM 702, and the RAM 703 are connected to each other via a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0173] Typically, the following devices may be connected to the I / O interface 705: input devices 707 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; output devices 707 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 708 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 709. The communication devices 709 may allow the electronic device to communicate with other devices wirelessly or by wire to exchange data. Although Figure 7 An electronic device having various devices is shown, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0174] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program contains program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device 709, or installed from a storage device 708, or installed from a ROM 702. When the computer program is executed by the processing device 701, the above-mentioned functions defined in the method of the embodiment of the present disclosure are executed.

[0175] It should be noted that the computer-readable medium mentioned above in the present disclosure may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0176] In some embodiments, the client and the server may communicate using any currently known or future developed network protocol such as HTTP (HyperText Transfer Protocol), and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), an internet (e.g., the Internet), and a peer-to-peer network (e.g., an ad hoc peer-to-peer network), as well as any currently known or future developed network.

[0177] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0178] The computer-readable medium carries one or more programs. When the one or more programs are executed by the electronic device, the electronic device: imports the first adjustment method information and the first push-related information of the first candidate information into the target model to obtain the first scores corresponding to the respective first adjustment method information, wherein the number of the first adjustment method information is at least two; selects the target first adjustment method information from the at least two first adjustment method information according to the respective first scores; uses the target first adjustment method information to adjust the first to-be-adjusted representation value of the first candidate information to generate a first adjusted representation value; and pushes the first candidate information based on the first adjusted representation value.

[0179] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages ​​or a combination thereof, including, but not limited to, object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0180] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0181] The units involved in the embodiments described in the present disclosure may be implemented by software or hardware. The name of a unit does not limit the unit itself in some cases. For example, an import unit may also be described as an "import target model unit".

[0182] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0183] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or apparatus. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any suitable combination of the foregoing. More specific examples of machine-readable storage media may include an electrical connection based on one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0184] The above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present disclosure is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present disclosure (but not limited to) by each other to form a technical solution.

[0185] In addition, although each operation is described in a specific order, this should not be understood as requiring these operations to be performed in the specific order shown or in a sequential order. Under certain circumstances, multitasking and parallel processing may be advantageous. Similarly, although some specific implementation details are included in the above discussion, these should not be interpreted as limiting the scope of the present disclosure. Some features described in the context of a separate embodiment can also be implemented in a single embodiment in combination. On the contrary, the various features described in the context of a single embodiment can also be implemented in multiple embodiments individually or in any suitable sub-combination mode.

[0186] Although the subject matter has been described in language specific to structural features and / or methodological logical actions, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. On the contrary, the specific features and actions described above are merely example forms of implementing the claims.

Claims

1. An information push method, characterized in that: include: Importing the first adjustment method information into the target model, and importing the first push-related information of the first candidate information into the target model, to obtain a first score corresponding to each piece of first adjustment method information, wherein the number of the first adjustment method information is at least two; Selecting target first adjustment method information from at least two pieces of first adjustment method information according to each first score; Using the target first adjustment method information, adjusting the first to-be-adjusted representation value of the first candidate information to generate a first adjusted representation value, wherein the representation value of the first candidate information is a basis for determining whether to push the first candidate information; Pushing the first candidate information based on the first adjusted representation value; The first push related information includes at least one of the following: first push basis information and first push history information; The first push basis information includes: attribute information of the first candidate information and attribute information of the owner of the first candidate information; The first push history information includes at least one of the following: first push process history information, historical operation information performed on the pushed first candidate information; At least two pieces of first adjustment mode information indicate different adjustment ranges for the same first representation value to be adjusted; Among them, the historical information of the first push process includes at least one of the following: the start time of pushing, the length of time from the start time of pushing to the current time, the number of pushes, and the total cost since the push of the first candidate information; the historical operation information on the pushed first candidate information includes at least one of the following: the user portrait of the user who operates the first candidate information, the first operation number of the first operation on the first candidate information, and the second operation number of the second operation on the second candidate information.

2. The method according to claim 1, characterized in that The selecting target first adjustment method information from at least two pieces of first adjustment method information according to each first score includes: The first adjustment method information corresponding to the highest first score is determined as the target first adjustment method information.

3. The method according to claim 1, characterized in that The pushing of the first candidate information based on the first adjusted representation value includes at least one of the following: Determining whether to display first candidate information based on the first adjusted representation value; Determine the position of the first candidate information to be displayed in the sequence of information to be pushed.

4. The method according to claim 1, characterized in that: The target model is obtained through the following model training steps, wherein the model training steps include: Importing the second adjustment method information into the to-be-trained model, and importing the second push-related information of the second candidate information into the to-be-trained model, to obtain a second score corresponding to each piece of second adjustment method information, wherein the number of the second adjustment method information is at least two; Selecting target second adjustment method information from at least two pieces of second adjustment method information according to each second score; Using the target second adjustment method information, adjusting the second to-be-adjusted representation value of the second candidate information to generate a second adjusted representation value; Based on the push effect information, the parameters of the model to be trained are adjusted, wherein the push effect information is used to characterize the push effect of pushing the second candidate information based on the second adjusted characterization value.

5. The method according to claim 4, characterized in that The second push related information includes at least one of the following: second push basis information and second push history information; The second push basis information includes at least one of the following: attribute information of the second candidate information, attribute information of the owner of the second candidate information; The second push history information includes at least one of the following: second push process history information, and second historical operation information performed on the candidate information to be pushed.

6. The method according to claim 4, characterized in that The selecting target second adjustment method information from at least two pieces of second adjustment method information according to each second score includes: A target second score is selected from the second scores by adopting a random selection method with a set selection probability, wherein the selection probability of the highest second score is the first preset probability; The second adjustment method information corresponding to the selected target second score is determined as the target second adjustment method information.

7. The method according to claim 6, characterized in that The model training step includes: Based on an increase in the number of model training updates, the first preset probability is increased.

8. The method according to claim 4, characterized in that The adjusting the parameters of the model to be trained based on the push effect information includes: Determining a loss function value according to a change in the push effect information of the second candidate information; According to the loss function value, adjust the parameters of the model to be trained.

9. The method according to claim 4, characterized in that The adjusting the parameters of the model to be trained based on the push effect information includes: According to the push effect information corresponding to the second candidate information set, the parameters of the model to be trained are adjusted, wherein the second candidate information set includes at least two second candidate information to be pushed.

10. A model training method, characterized in that: The method comprises: Importing the second adjustment method information into the to-be-trained model, and importing the second push-related information of the second candidate information into the to-be-trained model, to obtain a second score corresponding to each piece of second adjustment method information, wherein the number of the second adjustment method information is at least two; Selecting target second adjustment method information from at least two pieces of second adjustment method information according to each second score; Using the target second adjustment method information, adjusting the second to-be-adjusted representation value of the second candidate information to generate a second adjusted representation value, wherein the representation value of the second candidate information is a basis for determining whether to push the second candidate information; Based on the push effect information, adjusting the parameters of the model to be trained, wherein the push effect information is used to characterize the push effect of pushing the second candidate information based on the second adjusted characterization value; The second push related information includes at least one of the following: second push basis information and second push history information; The second push basis information includes: attribute information of the second candidate information and attribute information of the owner of the second candidate information; The second push history information includes at least one of the following: second push process history information, second historical operation information performed on the candidate information to be pushed; At least two pieces of second adjustment mode information indicate different adjustment ranges for the same second representation value to be adjusted; Among them, the historical information of the second push process includes at least one of the following: the start time of pushing, the length of time from the start time of pushing to the current time, the number of pushes, and the total cost since the push of the second candidate information; the historical operation information on the pushed second candidate information includes at least one of the following: the user portrait of the user who operates the second candidate information, the first number of operations for the first operation on the second candidate information, and the second number of operations for the second operation on the second candidate information.

11. The method according to claim 10, characterized in that The selecting target second adjustment method information from at least two pieces of second adjustment method information according to each second score includes: A target second score is selected from the second scores by adopting a random selection method with a set selection probability, wherein the selection probability of the highest second score is the first preset probability; The second adjustment method information corresponding to the selected target second score is determined as the target second adjustment method information.

12. The method according to claim 11, characterized in that The model training step includes: Based on an increase in the number of model training updates, the first preset probability is increased.

13. An information push device, characterized in that: include: an importing unit, configured to import the first adjustment method information into the target model, and import the first push-related information of the first candidate information into the target model, to obtain a first score corresponding to each piece of first adjustment method information, wherein the number of the first adjustment method information is at least two; A selection unit, configured to select target first adjustment method information from at least two pieces of first adjustment method information according to each first score; a generating unit, configured to adjust the first to-be-adjusted representation value of the first candidate information by using the target first adjustment mode information, and generate a first adjusted representation value, wherein the representation value of the first candidate information is a basis for determining whether to push the first candidate information; A pushing unit, configured to push the first candidate information based on the first adjusted representation value; The first push related information includes at least one of the following: first push basis information and first push history information; The first push basis information includes: attribute information of the first candidate information and attribute information of the owner of the first candidate information; The first push history information includes at least one of the following: first push process history information, historical operation information performed on the pushed first candidate information; At least two pieces of first adjustment mode information indicate different adjustment ranges for the same first representation value to be adjusted; Among them, the historical information of the first push process includes at least one of the following: the start time of pushing, the length of time from the start time of pushing to the current time, the number of pushes, and the total cost since the push of the first candidate information; the historical operation information on the pushed first candidate information includes at least one of the following: the user portrait of the user who operates the first candidate information, the first operation number of the first operation on the first candidate information, and the second operation number of the second operation on the second candidate information.

14. A model training device, characterized in that: The device comprises: an import module, used to import the second adjustment mode information into the to-be-trained model, and import the second push-related information of the second candidate information into the to-be-trained model, to obtain a second score corresponding to each second adjustment mode information, wherein the number of the second adjustment mode information is at least two; A selection module, configured to select target second adjustment method information from at least two pieces of second adjustment method information according to each second score; a generating module, configured to adjust the second to-be-adjusted representation value of the second candidate information by using the target second adjustment mode information, and generate a second adjusted representation value, wherein the representation value of the second candidate information is a basis for determining whether to push the second candidate information; an adjustment module, configured to adjust the parameters of the model to be trained based on the push effect information, wherein the push effect information is used to characterize the push effect of pushing the second candidate information based on the second adjusted characterization value; The second push related information includes at least one of the following: second push basis information and second push history information; The second push basis information includes: attribute information of the second candidate information and attribute information of the owner of the second candidate information; The second push history information includes at least one of the following: second push process history information, historical operation information performed on the second push candidate information; At least two pieces of second adjustment mode information indicate different adjustment ranges for the same second representation value to be adjusted; Among them, the historical information of the second push process includes at least one of the following: the start time of pushing, the length of time from the start time of pushing to the current time, the number of pushes, and the total cost since the push of the second candidate information; the historical operation information on the pushed second candidate information includes at least one of the following: the user portrait of the user who operates the second candidate information, the first number of operations for the first operation on the second candidate information, and the second number of operations for the second operation on the second candidate information.

15. An electronic device, characterized in that: include: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 9 or the method according to any one of claims 10 to 12.

16. A computer readable medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method described in any one of claims 1 to 9 or the method described in any one of claims 10 to 12 is implemented.

Images