Search

CN-111695037-B - Information recommendation method and device based on artificial intelligence and electronic equipment

CN111695037BCN 111695037 BCN111695037 BCN 111695037BCN-111695037-B

Abstract

The application provides an information recommending method, device, electronic equipment and computer readable storage medium based on artificial intelligence, wherein the method comprises the steps of determining relevance characteristics between information to be determined and all information in an information set, wherein the information set comprises at least one of the information to be determined, the information to be determined is information which is already allocated with display positions in a position sequence, the information to be determined is information of display positions in the position sequence to be allocated, corresponding first click rate is determined according to the relevance characteristics of each piece of information to be determined, unallocated display positions with highest priority in the position sequence are allocated to the information to be determined with highest first click rate and marked as new position information, and when the display positions in the position sequence are allocated, recommending operation is executed based on each piece of the information to be determined and the priority of the corresponding allocated display position. The method and the device can improve the accuracy of the recommended information.

Inventors

  • GAO QIAN
  • ZHANG SHENZHENG
  • ZHANG XINYU
  • DU YING

Assignees

  • 腾讯科技（北京）有限公司

Dates

Publication Date
20260505
Application Date
20200611

Claims (20)

  1. 1. An artificial intelligence based information recommendation method, comprising: Determining relevance features between the undetermined position information and all information in the information set; The information set comprises at least one of the following information, the undetermined position information and the information set, wherein the undetermined position information is the information in the information set, to which the display position in the position sequence is allocated, and the undetermined position information is the information in the information set, to which the display position in the position sequence is to be allocated; Based on the full connection parameter of each piece of the undetermined position information, carrying out full connection processing on the relevance characteristic of each piece of undetermined position information to obtain a first click rate when the undetermined position information is displayed at the corresponding undetermined position; distributing unallocated display positions with highest priority in the position sequence to undetermined position information with highest first click rate, and marking the undetermined position information as new set position information; And when the display positions in the position sequence are distributed, executing recommendation operation based on each piece of the information of the preset positions and the priority of the corresponding distributed display positions.
  2. 2. The method of claim 1, wherein prior to determining the correlation characteristics between the pending location information and all information in the information set, the method further comprises: acquiring basic characteristics of each piece of information in an information base; Based on the general full-connection parameters of the information base, carrying out full-connection processing on the basic features to obtain corresponding second click rate; And carrying out descending order sorting processing on the information base based on the second click rate of each piece of information, and selecting a plurality of pieces of information which are sorted in front in the descending order sorting result to form the information set.
  3. 3. The method of claim 1, wherein determining the correlation characteristics between the pending location information and all information in the information set comprises: Acquiring the characteristics of each piece of undetermined position information and the characteristics of each piece of determined position information in the information set; The following processing is performed for each piece of the pending position information: Performing attention coding processing on the characteristics of each piece of undetermined position information to obtain the association degree between the undetermined position information and each piece of information in the information set; And determining the association characteristic of the undetermined position information based on the association degree between the undetermined position information and each piece of information in the information set.
  4. 4. A method according to claim 3, wherein the performing attention encoding processing on the feature of each piece of the pending location information to obtain the degree of association between the pending location information and each piece of information in the information set includes: Performing linear transformation processing on the characteristics of each piece of information in the information set to obtain a query vector, a key vector and a value vector corresponding to each piece of information; And carrying out point multiplication processing on the query vector of the undetermined position information and the key vector of each piece of information in the information set, and carrying out normalization processing based on a maximum likelihood function on a point multiplication processing result to obtain the association degree between the undetermined position information and each piece of information in the information set.
  5. 5. A method according to claim 3, wherein said determining an association characteristic of said pending location information based on a degree of association between said pending location information and each information in said set of information comprises: Determining the degree of association as an attention weight of a value vector corresponding to each piece of information; And weighting the value vector based on the attention weight to obtain the associated characteristic of the undetermined position information based on the attention coding process.
  6. 6. A method according to claim 3, wherein said obtaining a characteristic of each pending location information in said set of information and a characteristic of each determined location information comprises: acquiring basic characteristics of each piece of information in the information set; acquiring the position characteristic of each piece of the preset position information in the information set, wherein the position characteristic is used for representing the display position of the preset position information; Taking the basic characteristics of the undetermined position information as the characteristics of each undetermined position information; and fusing the basic features and the position features of the fixed position information to obtain the features of the fixed position information.
  7. 7. The method of claim 6, wherein the base characteristic comprises at least one of: basic attribute characteristics used for representing basic information of a user to be recommended, interest tag characteristics used for representing interest preference of the user to be recommended, environmental characteristics used for representing a recommendation environment for recommending the information to the user to be recommended, category characteristics used for representing categories of the information, source characteristics used for representing sources of the information and content characteristics used for representing contents of the information.
  8. 8. The method of claim 6, wherein the obtaining the base characteristic of each piece of information in the set of pieces of information comprises: the following is performed for each information in the set of information: inquiring a plurality of feature vectors corresponding to the information from a pre-established feature vector matrix; And carrying out fusion processing on a plurality of feature vectors of the information to obtain basic features corresponding to the information.
  9. 9. The method according to claim 1, wherein the method further comprises: determining new association features between the new determined location information and the new pending location information; determining a corresponding new first click rate according to the relevance characteristic of each piece of new undetermined position information; And allocating the unallocated display positions with the highest priority in the position sequence to new undetermined position information with the highest first click rate until the display positions in the position sequence are allocated.
  10. 10. The method of claim 1, wherein the step of determining the position of the substrate comprises, The first click rate of each piece of information of the undetermined position in the information set is obtained by calling a click rate prediction model; Before determining the correlation characteristic between the determined location information and the pending location information, the method further comprises: acquiring an information sample sequence and a real first click rate of each information sample in the information sample sequence from a recommendation log; Training the kth display position of the position sequence on the click rate prediction model based on the information sample sequence and the corresponding real first click rate to update parameters corresponding to the kth display position in the click rate prediction model, and fixing the parameters corresponding to other display positions in the click rate prediction model in the updating process; Wherein k is an integer greater than or equal to 1, and the other display positions are display positions in the position sequence except the first k-1 display positions; When the parameters corresponding to the kth display position in the click rate prediction model are unchanged, continuing to train the (k+1) th display position of the position sequence on the click rate prediction model so as to update the parameters corresponding to the (k+1) th display position in the click rate prediction model, and fixing the parameters corresponding to other display positions in the click rate prediction model in the updating process; and when the parameters corresponding to each display position in the click rate prediction model are determined, determining that the click rate prediction model training is completed.
  11. 11. The method of claim 10, wherein training the click rate prediction model for a kth presentation position of a sequence of positions to update parameters of the click rate prediction model for the kth presentation position comprises: The following is performed in the training of the click rate prediction model for the kth presentation position of the sequence of positions: Determining a predicted first click rate of each information sample except the first k-1 information samples in the information sample sequence through the click rate prediction model; determining an error between the predicted first click rate and the true first click rate for each of the information samples and back-propagating the error in the click rate prediction model according to a loss function corresponding to the kth presentation position to Determining a parameter change value corresponding to the kth display position in the click rate prediction model when the loss function corresponding to the kth display position obtains the minimum value; and updating the parameters corresponding to the kth display position in the click rate prediction model according to the determined parameter change value.
  12. 12. An artificial intelligence based information recommendation device, comprising: the feature acquisition module is used for determining the relevance features between the undetermined position information and all the information in the information set; The information set comprises at least one of the following information, the undetermined position information and the information set, wherein the undetermined position information is the information in the information set, to which the display position in the position sequence is allocated, and the undetermined position information is the information in the information set, to which the display position in the position sequence is to be allocated; The click rate determining module is used for acquiring the undetermined positions corresponding to the undetermined position information respectively and acquiring full-connection parameters corresponding to the undetermined positions; based on the full connection parameters of each piece of undetermined position information, carrying out full connection processing on the relevance characteristics of each piece of undetermined position information to obtain a first click rate of the undetermined position information when the corresponding undetermined position information is displayed; The position allocation module is used for allocating the unallocated display position with the highest priority in the position sequence to the undetermined position information with the highest first click rate and marking the undetermined position information as new set position information; and the recommending module is used for executing recommending operation based on each piece of the preset position information and the priority of the corresponding allocated display position when the display position in the position sequence is allocated.
  13. 13. The apparatus of claim 12, wherein the device comprises a plurality of sensors, The feature acquisition module is further used for acquiring basic features of each piece of information in the information base before determining the relevance features between the undetermined position information and all pieces of information in the information set, carrying out full connection processing on the basic features based on the general full connection parameters of the information base to obtain corresponding second click rate, carrying out descending order sorting processing on the information base based on the second click rate of each piece of information, and selecting a plurality of pieces of information which are sorted in front in the descending order sorting result to form the information set.
  14. 14. The apparatus of claim 12, wherein the device comprises a plurality of sensors, The feature acquisition module is further used for acquiring the feature of each piece of undetermined position information in the information set and the feature of each piece of determined position information, and executing the following processing on each piece of undetermined position information, namely performing attention coding processing on the feature of each piece of undetermined position information to obtain the association degree between the undetermined position information and each piece of information in the information set, and determining the association feature of the undetermined position information based on the association degree between the undetermined position information and each piece of information in the information set.
  15. 15. The apparatus of claim 14, wherein the feature obtaining module is further configured to perform a linear transformation process on a feature of each piece of information in the information set to obtain a query vector, a key vector, and a value vector corresponding to each piece of information, perform a point multiplication process on the query vector of the pending location information and the key vector of each piece of information in the information set, and perform a normalization process on a point multiplication result based on a maximum likelihood function to obtain a degree of association between the pending location information and each piece of information in the information set.
  16. 16. The apparatus of claim 14, wherein the device comprises a plurality of sensors, The feature acquisition module is further used for determining the association degree as the attention weight of the value vector corresponding to each piece of information, and carrying out weighting processing on the value vector based on the attention weight to obtain the association feature of the undetermined position information based on the attention coding processing.
  17. 17. The apparatus of claim 14, wherein the device comprises a plurality of sensors, The feature acquisition module is further used for acquiring basic features of each piece of information in the information set, acquiring position features of each piece of fixed position information in the information set, wherein the position features are used for representing display positions of the fixed position information, taking the basic features of the to-be-determined position information as the features of each piece of to-be-determined position information, and carrying out fusion processing on the basic features and the position features of the fixed position information to obtain the features of the fixed position information.
  18. 18. The apparatus of claim 17, wherein the base characteristic comprises at least one of: basic attribute characteristics used for representing basic information of a user to be recommended, interest tag characteristics used for representing interest preference of the user to be recommended, environmental characteristics used for representing a recommendation environment for recommending the information to the user to be recommended, category characteristics used for representing categories of the information, source characteristics used for representing sources of the information and content characteristics used for representing contents of the information.
  19. 19. The apparatus of claim 17, wherein the device comprises a plurality of sensors, The feature acquisition module is further used for carrying out the following processing on each piece of information in the information set, wherein the processing comprises the steps of inquiring a plurality of feature vectors corresponding to the information from a pre-established feature vector matrix, and carrying out fusion processing on the plurality of feature vectors of the information to obtain basic features corresponding to the information.
  20. 20. The apparatus of claim 12, wherein the device comprises a plurality of sensors, The feature acquisition module is further used for determining new association features between the new set position information and the new pending position information; the click rate determining module is further configured to determine a corresponding new first click rate according to the relevance feature of each piece of new pending location information; the position allocation module is further configured to allocate an unallocated display position with the highest priority in the position sequence to new pending position information with the highest first click rate until the display position in the position sequence is allocated.

Description

Information recommendation method and device based on artificial intelligence and electronic equipment Technical Field The present application relates to an information recommendation technology of artificial intelligence, and in particular, to an information recommendation method, apparatus, electronic device and computer readable storage medium based on artificial intelligence. Background Cloud computing (clouding) is a computing model that distributes computing tasks across a large pool of computers, enabling various application systems to acquire computing power, storage space, and information services as needed. The network that provides the resources is referred to as the "cloud". Resources in the cloud are infinitely expandable in the sense of users, and can be acquired at any time, used as needed, expanded at any time and paid for use as needed. Artificial intelligence (AI, artificial Intelligence) is the theory, method and technique and application system that uses a digital computer or a machine controlled by a digital computer to simulate, extend and expand human intelligence, sense the environment, acquire knowledge and use the knowledge to obtain optimal results. The information recommendation is an important application of artificial intelligence, a rearrangement module in the related art is the final stage of personalized recommendation of a recommendation system, and the rearrangement module breaks up the information generated by the sequencing module according to preset rules and then presents the information to a user so as to prevent the information with higher repeatability from being continuously presented to the user, and the problem of lack of individuation in recommendation exists. Disclosure of Invention The embodiment of the invention provides an information recommending method, an information recommending device, electronic equipment and a computer readable storage medium based on artificial intelligence, which can improve the accuracy of information recommendation and further improve the user experience. The technical scheme of the embodiment of the invention is realized as follows: the embodiment of the invention provides an information recommendation method based on artificial intelligence, which comprises the following steps: Determining relevance features between the undetermined position information and all information in the information set; The information set comprises at least one of the following information, the undetermined position information and the information set, wherein the undetermined position information is the information in the information set, to which the display position in the position sequence is allocated, and the undetermined position information is the information in the information set, to which the display position in the position sequence is to be allocated; Determining a corresponding first click rate according to the relevance characteristics of each piece of undetermined position information; distributing unallocated display positions with highest priority in the position sequence to undetermined position information with highest first click rate, and marking the undetermined position information as new set position information; And when the display positions in the position sequence are distributed, executing recommendation operation based on each piece of the information of the preset positions and the priority of the corresponding distributed display positions. The embodiment of the invention provides an information recommendation device based on artificial intelligence, which comprises the following components: the feature acquisition module is used for determining the relevance features between the undetermined position information and all the information in the information set; The information set comprises at least one of the following information, namely, the information of the preset position, the information of the undetermined position, wherein the information of the preset position is the information of the display position in the position sequence which is already allocated in the information set, and the information of the undetermined position is the information of the display position in the position sequence which is to be allocated in the information set; The click rate determining module is used for determining a corresponding first click rate according to the relevance characteristics of each piece of undetermined position information; The position allocation module is used for allocating the unallocated display position with the highest priority in the position sequence to the undetermined position information with the highest first click rate and marking the undetermined position information as new set position information; and the recommending module is used for executing recommending operation based on each piece of the preset position information and the priority of the corresponding allocated display position when th