Information search method and device, electronic equipment and storage medium

Abstract

The invention discloses an information search method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining current search information and account information, carrying out the intention recognition of the current search information based on an intention recognition model, and obtaining the intention information corresponding to the current search information, the intention recognition model is obtained by training a preset recognition model in combination with a to-be-searched data set in at least one second system, and the correlation degree of the search attribute of the first system and the search attribute of each second system meets a preset correlation condition; and performing matching processing on the intention information corresponding to the current search information and feedback contents in a feedback content set contained in the first system to obtain a matching result, and determining a candidate feedback content set from the feedback content set according to the matching result and the account information. According to the embodiment of the invention, the annotation data with good quality can be quickly and effectively obtained by means of the search results of other systems related to the attributes for training the intention recognition model, so that the application of the search function of the first system can be quickly realized.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and particularly to an information search method, apparatus, electronic device, and storage medium. Background Art

[0002] With the development of artificial intelligence technology, various artificial intelligence data processing models have developed rapidly. Model training is an important part of the model construction process. In some model training methods, model training depends on training data with labels.

[0003] Traditional data labeling methods include manual labeling and constructing training data based on historical data. However, due to the lack of historical exposure click data in the cold start phase, and the manual labeling processing method faces the problem of low labeling efficiency, so how to obtain the training data required for model training has always remained to be solved. Summary of the Invention

[0004] To solve the problems of the prior art, embodiments of the present invention provide an information search method, apparatus, electronic device, and storage medium. The technical solutions are as follows:

[0005] On the one hand, an information search method is provided, which is applied to the first system. The method includes:

[0006] Obtain the current search information and account information;

[0007] Perform intent recognition on the current search information based on an intent recognition model to obtain intent information corresponding to the current search information; the intent recognition model is obtained by training an untrained preset recognition model in combination with at least one search dataset in the second system; the search dataset includes search data to be searched and search feedback data; the correlation degree between the search attributes of the first system and the search attributes of each second system satisfies a preset correlation condition;

[0008] Match the intent information corresponding to the current search information with the feedback content in the feedback content set included in the first system to obtain a matching result;

[0009] Determine a candidate feedback content set from the feedback content set according to the matching result and the account information.

[0010] On the other hand, an information search apparatus is provided, which is applied to the first system. The apparatus includes:

[0011] An information acquisition module, configured to obtain the current search information and account information;

[0012] An intent recognition module, configured to perform intent recognition on the current search information based on an intent recognition model to obtain intent information corresponding to the current search information; the intent recognition model is obtained by training a preset recognition model that has not been trained in combination with a to-be-searched dataset in at least one second system; the to-be-searched dataset includes to-be-searched data and search feedback data; the correlation degree between the search attributes of the first system and the search attributes of each second system satisfies a preset correlation condition;

[0013] A content matching module, configured to perform matching processing on the intent information corresponding to the current search information and the feedback content in the feedback content set included in the first system to obtain a matching result;

[0014] A content determination module, configured to determine a candidate feedback content set from the feedback content set according to the matching result and the account information.

[0015] In some possible embodiments, the content matching module is configured to:

[0016] Determine the intent label of each feedback content in the feedback content set included in the first system;

[0017] Perform matching processing on the intent information corresponding to the current search information and the intent label of each feedback content to obtain a matching result; the matching result indicates the matching degree between the current search information and each feedback content.

[0018] In some possible embodiments, the apparatus further includes an intent label modification module, configured to:

[0019] Display the candidate feedback content in the candidate feedback content set;

[0020] In response to an account consumption instruction for the candidate feedback content, update the intent label of the candidate feedback content based on the intent information.

[0021] In some possible embodiments, the intent label modification module is configured to:

[0022] In response to an account consumption instruction for the candidate feedback content, if the account consumption instruction indicates that the candidate feedback content is consumed and the intent label of the candidate feedback content does not include the intent information, add the intent information to the intent label of the candidate feedback content;

[0023] Or;

[0024] In response to an account consumption instruction for the candidate feedback content, if the account consumption instruction indicates that the candidate feedback content is consumed and the intent label of the candidate feedback content includes the intent information, increase the weight data of the intent information in the intent label of the candidate feedback content.

[0025] In some possible embodiments, the intent label modification module is configured to:

[0026] In response to an account consumption instruction for candidate feedback content, if the account consumption instruction indicates that the candidate feedback content has not been consumed and the intention label of the candidate feedback content contains intention information, update the consumption record of the intention information in the intention label of the candidate feedback content;

[0027] The consumption record is used to delete the intention information from the intention label of the candidate feedback content when the consumption record meets the deletion condition.

[0028] In some possible embodiments, the apparatus further includes an intention recognition model update module for:

[0029] When it is detected that the candidate feedback content set has not been consumed and the updated feedback content in the updated feedback content set associated with the updated search information corresponding to the current search information has been consumed, determine the intention information corresponding to the updated search information;

[0030] Update and train the intention recognition model based on the current search information and the intention information corresponding to the updated search information to obtain an updated intention recognition model.

[0031] In some possible embodiments, the intention recognition model update module is used for:

[0032] When the updated search information is a subordinate information or an equivalent information of the current search information, the difference between the time when the current search information is obtained and the time when the updated search information is obtained is less than or equal to the time threshold, and the updated search information and the current search information correspond to the same account information, determine that the updated search information is the updated search information corresponding to the current search information.

[0033] In some possible embodiments, the apparatus further includes an intention recognition model training module for:

[0034] Determine the search attributes of the first system and the search attributes of the alternative systems;

[0035] Determine a second system from the alternative systems based on the search attributes of the first system and the search attributes of the alternative systems;

[0036] Obtain the data to be searched corresponding to the first system;

[0037] Obtain the search feedback data corresponding to the data to be searched from the second system;

[0038] Obtain the intention key data of the search feedback data;

[0039] Use the intention key data in the search feedback data as the annotation data of the data to be searched corresponding to the search feedback data to train a preset recognition model to obtain an intention recognition model.

[0040] In some possible embodiments, an intention recognition model training module is configured to:

[0041] Use key information to determine the key information of each search feedback data by the model, and obtain the key information corresponding to each search feedback data;

[0042] Perform a universality statistical process on the key information corresponding to each search feedback data to obtain a key information statistical result;

[0043] Determine the intention key data of multiple search feedback data based on the key information statistical result.

[0044] On the other hand, an electronic device is provided, including a processor and a memory. At least one instruction or at least one program segment is stored in the memory, and the at least one instruction or at least one program segment is loaded and executed by the processor to implement the above information search method.

[0045] On the other hand, a computer-readable storage medium is provided. At least one instruction or at least one program segment is stored in the computer-readable storage medium, and the at least one instruction or at least one program segment is loaded and executed by the processor to implement the information search method as described above.

[0046] On the other hand, a computer program product or a computer program is provided. The computer program product or the computer program includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the above information search method.

[0047] In the embodiments of the present invention, by obtaining the current search information and account information, and performing intention recognition on the current search information based on the intention recognition model, the intention information corresponding to the current search information is obtained. The intention recognition model is trained by combining a preset recognition model that has not been trained with a to-be-searched data set in at least one second system. The to-be-searched data set includes to-be-searched data and search feedback data. The correlation degree between the search attributes of the first system and the search attributes of each second system satisfies a preset correlation condition. The intention information corresponding to the current search information is matched with the feedback content in the feedback content set included in the first system to obtain a matching result, and the candidate feedback content set is determined from the feedback content set according to the matching result and the account information. The embodiments of the present application use the search results of other systems related to attributes to implement a semi-supervised annotation method for intention training data. In the case of lacking historical exposure click data and manually annotated training data, this method can quickly and effectively obtain good-quality annotated data for the training of the intention recognition model, so as to quickly implement the application of the search function of the first system. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] To more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the accompanying drawings required for the description of the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0049] Figure 1 It is a schematic diagram of an implementation environment provided by an embodiment of the present invention;

[0050] Figure 2 It is a schematic flowchart of an information search method provided by an embodiment of the present invention;

[0051] Figure 3 It is a schematic flowchart of a method for training an intent recognition model provided by an embodiment of the present invention;

[0052] Figure 4 It is a schematic flowchart of a method for determining intent key data provided by an embodiment of the present invention;

[0053] Figure 5 It is a schematic diagram of determining intent key data provided by an embodiment of the present invention;

[0054] Figure 6 It is a schematic flowchart of an information search method provided by an embodiment of the present invention;

[0055] Figure 7 It is a schematic flowchart of an information search method provided by an embodiment of the present invention;

[0056] Figure 8 It is a structural block diagram of an information search device provided by an embodiment of the present invention;

[0057] Figure 9 It is a hardware structural block diagram of an electronic device provided by an embodiment of the present invention. Detailed implementation manners

[0058] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0059] It should be noted that in the description of the present invention, the claims and the above-mentioned drawings, the terms "first", "second", etc. are used to distinguish similar objects, and do not necessarily describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or server that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0060] It can be understood that in the specific implementation of this application, when it comes to relevant data such as user information, when the above embodiments of this application are applied to specific products or technologies, user permission or consent needs to be obtained, and the collection, use, and processing of relevant data need to comply with relevant laws, regulations, and standards of relevant countries and regions.

[0061] To facilitate the understanding of the above technical solutions of the embodiments of the present disclosure and the technical effects they produce, a brief introduction is given to the nouns involved in the embodiments of the present disclosure:

[0062] Artificial Intelligence (AI) is a theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to obtain the best results. In other words, artificial intelligence is a comprehensive technology in computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can react in a way similar to human intelligence. Artificial intelligence also studies the design principles and implementation methods of various intelligent machines, enabling the machines to have the functions of perception, reasoning, and decision-making.

[0063] Artificial intelligence technology is an interdisciplinary subject that involves a wide range of fields, including both hardware-level technologies and software-level technologies. The basic technologies of artificial intelligence generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, and mechatronics. The software technologies of artificial intelligence mainly include several major directions such as computer vision technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0064] Computer Vision Technology (CV) Computer vision is a science that studies how to enable machines to "see". More specifically, it refers to machine vision that uses cameras and computers to replace human eyes for object recognition, measurement, etc., and further performs image processing to make the computer-processed images more suitable for human eye observation or transmission to instrument detection. As a scientific discipline, computer vision studies related theories and technologies, and attempts to establish artificial intelligence systems that can obtain information from images or multi-dimensional data. Computer vision technology usually includes image processing, image recognition, image semantic understanding, image retrieval, OCR, video processing, video semantic understanding, video content / behavior recognition, 3D object reconstruction, 3D technology, virtual reality, augmented reality, simultaneous localization and mapping, etc. technologies, and also includes common biometric recognition technologies such as face recognition and fingerprint recognition.

[0065] Natural Language Processing (NLP) is an important direction in the fields of computer science and artificial intelligence. It studies various theories and methods that can achieve effective communication between humans and computers in natural language. Natural language processing is a science that integrates linguistics, computer science, and mathematics. Therefore, the research in this field will involve natural language, that is, the language people use in daily life, so it has a close connection with the research of linguistics. Natural language processing technology usually includes text processing, semantic understanding, machine translation, robot question answering, knowledge graph, etc. technologies.

[0066] Machine Learning (ML) is an interdisciplinary subject that involves multiple disciplines such as probability theory, statistics, approximation theory, convex analysis, and algorithm complexity theory. It specifically studies how computers simulate or implement human learning behaviors to acquire new knowledge or skills and reorganize the existing knowledge structure to continuously improve their own performance. Machine learning is the core of artificial intelligence and the fundamental way to make computers intelligent, and its applications cover all fields of artificial intelligence. Machine learning and deep learning usually include technologies such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and rote learning. Machine learning can be divided into supervised machine learning, unsupervised machine learning, and semi-supervised machine learning.

[0067] Autopilot technology usually includes technologies such as high-precision maps, environmental perception, behavior decision-making, path planning, and motion control. Autopilot technology has a wide range of application prospects.

[0068] With the research and progress of artificial intelligence technology, artificial intelligence technology has been studied and applied in multiple fields, such as common smart homes, smart wearable devices, virtual assistants, smart speakers, smart marketing, driverless, autonomous driving, drones, robots, smart healthcare, smart customer service, etc. It is believed that with the development of technology, artificial intelligence technology will be applied in more fields and play an increasingly important role.

[0069] Embodiments of the present invention can be applied to various scenarios, including but not limited to cloud technology, artificial intelligence, intelligent transportation, assisted driving, etc.

[0070] Among them, cloud technology refers to a hosting technology that unifies a series of resources such as hardware, software, and networks within a wide area network or local area network to achieve data computing, storage, processing, and sharing. Cloud technology is the general term for network technology, information technology, integration technology, management platform technology, application technology, etc. based on the cloud computing business model, which can form a resource pool, be used on demand, and be flexible and convenient. Cloud computing technology will become an important support. The back-end services of the technical network system require a large amount of computing and storage resources, such as video websites, picture-based websites, and more portal websites. With the high development and application of the Internet industry, in the future, each item may have its own identification mark and needs to be transmitted to the back-end system for logical processing. Data at different levels will be processed separately, and various industry data requires a powerful system back-end support, which can only be achieved through cloud computing.

[0071] Cloud storage is a new concept extended and developed from the concept of cloud computing. A distributed cloud storage system (hereinafter referred to as the storage system) refers to a storage system that combines a large number of different types of storage devices (storage devices are also called storage nodes) in the network through cluster applications, grid technology, and distributed storage file systems, etc., and works together through application software or application interfaces to provide data storage and business access functions externally. Currently, the storage method of the storage system is as follows: Create a logical volume. When creating a logical volume, physical storage space is allocated for each logical volume. This physical storage space may be composed of the disks of a certain storage device or several storage devices. The client stores data on a certain logical volume, that is, stores the data on the file system. The file system divides the data into many parts, and each part is an object. The object not only contains data but also additional information such as data identification (ID, ID entity), etc. The file system writes each object into the physical storage space of the logical volume respectively, and the file system will record the storage location information of each object. Thus, when the client requests to access the data, the file system can enable the client to access the data according to the storage location information of each object. The process of the storage system allocating physical storage space for the logical volume is specifically as follows: According to the capacity estimation of the objects stored in the logical volume (this estimation often has a large margin relative to the actual capacity of the objects to be stored) and the group of the redundant array of independent disks (RAID), the physical storage space is pre-divided into stripes. A logical volume can be understood as a stripe, thereby allocating physical storage space for the logical volume.

[0072] A database, in short, can be regarded as an electronic filing cabinet - a place to store electronic files. Users can perform operations such as adding, querying, updating, and deleting data in the files. The so-called "database" is a data set stored together in a certain way, can be shared by multiple users, has the smallest possible redundancy, and is independent of application programs.

[0073] Blockchain is a new application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, and encryption algorithms. Blockchain, in essence, is a decentralized database, a series of data blocks generated by using cryptographic methods. Each data block contains information about a batch of network transactions, which is used to verify the validity of the information (anti-counterfeiting) and generate the next block. Blockchain can include the blockchain underlying platform, the platform product service layer, and the application service layer.

[0074] The underlying blockchain platform may include processing modules such as user management, basic services, smart contracts, and operation detection. Among them, the user management module is responsible for the identity information management of all blockchain participants, including maintaining the generation of public and private keys (account management), key management, and the maintenance of the correspondence between the real identity of users and blockchain addresses (permission management). And under authorization, it supervises and audits the transaction situations of certain real identities, and provides the rule configuration for risk control (risk control and audit); the basic service module is deployed on all blockchain node devices, used to verify the validity of business requests, and record them on the storage after consensus for valid requests. For a new business request, the basic service first performs interface adaptation parsing and authentication processing (interface adaptation), then encrypts the business information through a consensus algorithm (consensus management), transmits it to the shared ledger completely and consistently after encryption (network communication), and records and stores it; the smart contract module is responsible for the registration and issuance of contracts, as well as contract triggering and contract execution. Developers can define contract logic through a certain programming language, publish it to the blockchain (contract registration), trigger the execution by calling keys or other events according to the logic of the contract terms, complete the contract logic, and at the same time provide functions for contract upgrade and cancellation; the operation detection module is mainly responsible for the deployment, configuration modification, contract setting, cloud adaptation during the product release process, and the visual output of the real-time state during product operation, such as: alarming, detecting network conditions, detecting the health status of node devices, etc.

[0075] The platform product service layer provides the basic capabilities and implementation frameworks of typical applications. Based on these basic capabilities, developers can overlay the characteristics of the business to complete the blockchain implementation of the business logic. The application service layer provides application services based on the blockchain solution for business participants to use.

[0076] Please refer to Figure 1 , which shows a schematic diagram of an implementation environment provided by an embodiment of the present invention. The implementation environment may include a client 110, a server 120, and a database 130.

[0077] Among them, the client 110 and the server 120, and the server 120 and the database 130 can be connected and communicated through a network.

[0078] Among them, the client 110 includes, but is not limited to, mobile phones, computers, intelligent voice interaction devices, smart home appliances, in-vehicle clients, aircraft, etc. An application program with a human-computer interaction function can run on the client 110, and this application program can launch virtual item distribution activities for different business scenarios, such as flash sales, lucky draws, activities for receiving rewards by completing tasks, and so on. The client 110 can obtain the current search information and account information, perform intent recognition on the current search information based on the intent recognition model, and obtain the intent information corresponding to the current search information. The intent recognition model is trained by combining an untrained preset recognition model with the to-be-searched data set in at least one second system. The to-be-searched data set includes to-be-searched data and search feedback data. The correlation degree between the search attributes of the first system and the search attributes of each second system satisfies the preset correlation condition. The intent information corresponding to the current search information is matched with the feedback content in the feedback content set included in the first system to obtain a matching result, and the candidate feedback content set is determined from the feedback content set according to the matching result and the account information. The embodiment of the present application uses the search results of other systems related to attributes to implement the semi-supervised annotation method of intent training data. In the case of lacking historical exposure click data and manually annotated training data, this method can quickly and effectively obtain good-quality annotation data for the training of the intent recognition model, so as to quickly realize the application of the search function of the first system.

[0079] The server 120 can obtain the current search information and account information, perform intent recognition on the current search information based on the intent recognition model, and obtain the intent information corresponding to the current search information. The intent recognition model is trained by combining an untrained preset recognition model with the to-be-searched data set in at least one second system. The to-be-searched data set includes to-be-searched data and search feedback data. The correlation degree between the search attributes of the first system and the search attributes of each second system satisfies the preset correlation condition. The intent information corresponding to the current search information is matched with the feedback content in the feedback content set included in the first system to obtain a matching result, and the candidate feedback content set is determined from the feedback content set according to the matching result and the account information. The embodiment of the present application uses the search results of other systems related to attributes to implement the semi-supervised annotation method of intent training data. In the case of lacking historical exposure click data and manually annotated training data, this method can quickly and effectively obtain good-quality annotation data for the training of the intent recognition model, so as to quickly realize the application of the search function of the first system.

[0080] The database 130 may include an in-memory database and a relational database. It should be noted that the server, database, node, etc. in the embodiments of the present invention may be independent physical servers, or a server cluster or distributed system composed of multiple physical servers. It may also be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN (Content Delivery Network), as well as big data and artificial intelligence platforms.

[0081] In an exemplary embodiment, the client 110, the server 120, and the database 130 may all be node devices in a blockchain system, capable of sharing the acquired and generated information with other node devices in the blockchain system to achieve information sharing among multiple node devices. Multiple node devices in the blockchain system may be configured with the same blockchain, which is composed of multiple blocks, and adjacent blocks have an association relationship, so that when the data in any block is tampered with, it can be detected by the next block, thereby avoiding the tampering of the data in the blockchain and ensuring the security and reliability of the data in the blockchain.

[0082] Please refer to Figure 2 , which shows a schematic flowchart of an information search method provided by an embodiment of the present invention. This method can be applied to Figure 1 the implementation environment shown in Figure 1 . The execution subject of this method may be the server that determines the clustering result in Figure 2 , or it may also be a client or other server node that determines the clustering result. It should be noted that this specification provides method operation steps as described in the embodiments or flowcharts, but based on routine or non-creative labor, there may be more or fewer operation steps. The step order listed in the embodiments is only one of the execution orders of numerous steps and does not represent the only execution order. In actual system or product execution, it may be executed in the order of the embodiments or the method shown in the drawings, or executed in parallel (for example, in an environment of parallel processors or multi-threaded processing). Specifically, as

[0083] S201, obtain the current search information and account information.

[0084] In the embodiments of the present application, the current search information may be a search text, search image, etc. input by an account object. The search text may be obtained in forms such as manual input, voice input, or previous historical search records. The account information may be the account information of the account object, such as one or more of account attribute information, historical operation information, historical interest information, etc.

[0085] S203. Perform intent recognition on the current search information based on the intent recognition model to obtain the intent information corresponding to the current search information. The intent recognition model is obtained by training a preset recognition model that has not been trained in combination with the to-be-searched data set in at least one second system. The to-be-searched data set includes to-be-searched data and search feedback data. The correlation degree between the search attributes of the first system and the search attributes of each second system meets the preset correlation condition.

[0086] In the embodiments of the present application, the first system may represent a first search platform, a first server providing search services, or other devices. Similarly, the second system may represent a second search platform, a second server providing search services, or other devices.

[0087] In the embodiments of the present application, the intent recognition model may be constructed based on the machine learning technology described above. Among them, the intent recognition model may include, but is not limited to, deep learning models such as convolutional neural networks, recurrent neural networks, or recursive neural networks.

[0088] In some possible embodiments, the training of the intent recognition model may be implemented based on data annotation. Specifically, a data set to be annotated can be obtained, and each data in the data set to be annotated can be manually annotated to obtain the annotation data of each data. Subsequently, the preset recognition model can be trained using each data and its corresponding annotation data to obtain a trained intent recognition model.

[0089] However, the above training method of all data annotation has the disadvantages of long time consumption, low annotation efficiency, and high annotation cost, which results in huge resource consumption in the cold start stage of the preset recognition model that has not been trained.

[0090] Based on this, in some other possible embodiments, in the cold start stage of the preset recognition model in the first system, the recognition ability of the recognition model in the second system can be utilized to achieve the goal of learning from and reducing resource consumption.

[0091] In some alternative implementation manners, the intent recognition model may be obtained by training a preset recognition model that has not been trained in combination with the to-be-searched data set in at least one second system. Among them, the to-be-searched data set includes to-be-searched data and search feedback data obtained after the to-be-searched data is searched by the second system. In order to enable the data in the second system to be used by the first system and have strong relevance, the correlation degree between the search attributes of the first system and the search attributes of each second system may meet the preset correlation condition.

[0092] Please refer to Figure 3 , which shows a schematic flowchart of a process for training an intent recognition model provided by an embodiment of the present invention, including at least:

[0093] S301. Determine the search attributes of the first system and the search attributes of the alternative systems.

[0094] In the embodiments of the present application, the alternative system is the search system currently in use, and the first system is a search system not yet in use.

[0095] In some possible embodiments, the search attributes of the alternative system can be determined according to the search application scenarios of its corresponding search system. Optionally, the search application scenario can be determined based on what web page or what application the search system is applied to, or rather, determined by what kind of web page or application.

[0096] For example, assume that a certain search system is applied in the application scenario of a game application. Then the search attribute of this search system is a game attribute, because the search applications set in the game application are mostly likely to search for game-related information. Similarly, assume that a certain search system is applied in the application scenario of a music application. Then the search attribute of this search system is a music attribute, because the search applications set in the music application are mostly likely to search for music-related information. Assume that a certain search system is called by various applications. Then the search attribute of this search system is a comprehensive attribute, that is, this search system can be searched for any problem.

[0097] In some other possible embodiments, the search attributes of the alternative system can be determined according to the search problems of its corresponding search system. Specifically, when the search system receives search information, it can determine multiple attributes corresponding to this group of search information based on each search information and its corresponding search result, such as games, music, travel, etc. After collecting to a certain extent, such as when the time for analyzing the search information reaches a preset duration, or the quantity of the analyzed search information meets the preset quantity, the main attribute can be determined from multiple attributes and used as the search attribute of the search system.

[0098] In the embodiments of the present application, since the search function of the first system has not been applied yet, the search attributes of the first system can be determined according to the defined search setting parameters.

[0099] S303. Determine the second system from the alternative systems based on the search attributes of the first system and the search attributes of the alternative systems.

[0100] Optionally, assume that the search attribute of the first system is a music attribute. Then the second system is a system with a music attribute among the alternative systems.

[0101] Optionally, assume that the search attribute of the first system is a comprehensive attribute. Then the second system is a system with a comprehensive attribute among the alternative systems.

[0102] Optionally, assuming that the search attribute of the first system is a comprehensive attribute, multiple second systems can be determined from the alternative systems. Some of the second systems are systems with a music attribute as the search attribute among the alternative systems, some are systems with a game attribute as the search attribute, some are systems with a travel attribute as the search attribute, and some are systems with a teaching attribute as the search attribute... By splicing each second system with a professional tendency and different attributes, the finally obtained first system can have a more comprehensive search ability.

[0103] S305. Obtain the data to be searched corresponding to the first system.

[0104] In some possible embodiments, when the number of second systems is one, for example, the second system is a system with a music attribute or a system with a comprehensive attribute, the data to be searched corresponding to the first system can be obtained. Among them, for the consideration of the training of the subsequent corresponding preset recognition model, the number of data to be searched can be multiple.

[0105] In some other possible embodiments, when the number of second systems is multiple, for example, the search attribute of the first system is a comprehensive attribute, and the search attribute of each second system among the multiple second systems is one of the attributes in the comprehensive attribute. After obtaining the data to be searched of the first system (for the consideration of the training of the subsequent corresponding preset recognition model, the number of data to be searched can be multiple), classification can be performed according to the attribution attribute of each data to be searched of the first system. The attribution attribute indicates what search attribute system the data to be searched is related to, and then all the data to be searched can be divided into data packets to be searched corresponding to each of the multiple second systems one by one. For example, the data packet to be searched corresponding to the second system with a music attribute contains all the data to be searched related to the music attribute, such as "Blue and White Porcelain", etc.

[0106] S307. Obtain the search feedback data corresponding to the data to be searched from the second system.

[0107] In the embodiments of the present application, the data to be searched can be input into the second system to obtain the search feedback data corresponding to the data to be searched.

[0108] Optionally, the search feedback data can be one or more media resources, and the resource types of the media resources can include, but are not limited to, one or more of text, pictures, videos, information, etc. Among them, each search feedback data can carry information such as a title.

[0109] S309. Obtain the intent key data of the search feedback data;

[0110] In some possible embodiments, the intent key data of each search feedback data can be obtained from the configuration information of each search feedback data. Among them, the intent key data is used to summarize the main content of the search feedback data.

[0111] In some other possible embodiments, a mature key information determination model can be used to extract key information from the title of the search feedback data or the text, charts, voices, pictures, etc. in the search feedback data.

[0112] Please refer to Figure 4 , which shows a schematic flowchart of a method for determining intent key data provided by an embodiment of the present invention, including at least:

[0113] S401, use the key information determination model to determine the key information of each search feedback data, and obtain the key information corresponding to each search feedback data.

[0114] In the embodiments of the present application, the key information determination model can be constructed based on technologies such as Bert, Bert-CRF, CNN-CRF, LSTM-CRF, etc.

[0115] Please refer to Figure 5 , which shows a schematic diagram of a method for determining intent key data provided by an embodiment of the present invention. As Figure 5 shown, the to-be-searched data "A" is input into the second system, and the search feedback data corresponding to the to-be-searched data "A" with the highest sorting position is obtained. Among them, "A" can be the name of a song, and the search feedback data corresponding to "A" is 5 videos, and the title of each video is on the right side of the video.

[0116] Then, the title of each search feedback data in the search feedback data corresponding to "A" can be obtained, and the title of each search feedback data is input into the key information determination model to obtain the key information corresponding to each search feedback data on the right side of the key information determination model.

[0117] S403, perform a universality statistical process on the key information corresponding to each search feedback data to obtain a key information statistical result.

[0118] Continuing with Figure 5 as an example for elaboration, perform a universality statistical process on the key information corresponding to each search feedback data, and the obtained statistical result is as follows: the number of "A" is 3, the number of "A, XX" is 1, and the number of "A, light music" is 1.

[0119] S405, determine the intent key data of multiple search feedback data based on the key information statistical result.

[0120] Optionally, the "A" with the largest quantity can be used as the intent key data for the search feedback data.

[0121] S311. Use the intent key data in the search feedback data as the annotation data for the data to be searched corresponding to the search feedback data, and train a preset recognition model to obtain an intent recognition model.

[0122] Based on this, the intent key data in the search feedback data can be used as the annotation data for the data to be searched corresponding to the search feedback data to train a preset recognition model to obtain an intent recognition model. For example, use the intent key data "A" as the annotation data for the data to be searched "A" to train a preset recognition model to obtain an intent recognition model.

[0123] The embodiment of the present application also includes an implementation method for training a preset recognition model to obtain an intent recognition model.

[0124] Optionally, after obtaining the annotation data for each data to be searched in multiple data to be searched, based on the preset recognition model, perform intent recognition on the data to be searched to determine the predicted intent information for each data to be searched. Subsequently, based on the predicted intent information and annotation data for each data to be searched, determine a loss value. When the loss value is greater than a preset threshold, perform backpropagation based on the loss value to update the preset recognition model to obtain an updated preset recognition model. Repeat the steps: based on the updated preset recognition model, perform intent recognition on each data to be searched. When the loss value is less than or equal to the preset threshold, or when the number of training rounds meets the preset number of rounds, determine the preset recognition model as the intent recognition model.

[0125] In this way, we obtain an intent recognition model applied to the first system. And since the intent recognition model of the first system obtains annotation data by leveraging the functions of the second system during the cold start phase, a large amount of annotation work can be saved.

[0126] S205. Match the intent information corresponding to the current search information with the feedback content in the feedback content set included in the first system to obtain a matching result.

[0127] In an alternative embodiment, the feedback content included in the feedback content set of the first system includes, but is not limited to, one or more of text, pictures, videos, news, etc. Among them, each feedback data can carry information such as a title.

[0128] Based on this, the intention information corresponding to the current search information can be matched with the title of each feedback content in the feedback content set included in the first system, and whether the title of the feedback content contains or is equivalent to containing the intention information corresponding to the current search information is used as the matching result. In this embodiment, the matching result includes identification information of "yes" or "no".

[0129] In some other possible embodiments, the intention label of each feedback content included in the first system can be determined, and the intention information corresponding to the current search information is matched with the intention label of each feedback content to obtain a matching result. Optionally, the matching result indicates the matching degree between the current search information and each feedback content, such as "90%".

[0130] S207. Determine a candidate feedback content set from the feedback content set according to the matching result and the account information.

[0131] Based on the above, in the embodiment of the present application, a candidate feedback content set can be determined from the feedback content set according to the matching result. For example, the feedback content with the matching result indicating "yes" is used as the candidate feedback content in the candidate feedback content set, or the candidate feedback content in the candidate feedback content set is determined according to the matching degree.

[0132] Alternatively, when the account information is one or more of account attribute information, historical operation information, historical interest information, etc., the feedback content in the feedback content set can be double-screened according to the matching result and the account information to obtain a personalized candidate feedback content set that is more adapted to the account information.

[0133] For example, assuming that a number of candidate feedback contents related to the intention information "travel" are determined according to the matching result, when the historical interest information "food" is obtained, the candidate feedback contents strongly related to "food" can be selected from the number of candidate feedback contents.

[0134] In summary, the embodiment of the present application provides a semi-supervised annotation method for a search system to obtain intention training data by means of the sorting results of other systems related to attributes in the cold start stage. In the case of lacking historical exposure click data and manually annotated training data, this method can quickly and effectively obtain good-quality annotation data for the training of the intention recognition model.

[0135] Please refer to Figure 6 , which shows a schematic flowchart of an information search method provided by an embodiment of the present invention. After determining the candidate feedback content set, it further includes:

[0136] S209. Display the candidate feedback content in the candidate feedback content set.

[0137] Optionally, a part of the candidate feedback content can be selected from the candidate feedback content set and displayed on the search page.

[0138] S211, in response to an account consumption instruction for the candidate feedback content, update the intent label of the candidate feedback content based on the intent information.

[0139] In the embodiments of the present application, it can be determined whether to update the intent label of the candidate feedback content according to whether the account consumption instruction for the candidate feedback content and the intent information are included in the intent label of the candidate feedback content, so as to further improve the search ability of the first system.

[0140] Among them, the account consumption instruction is determined based on the consumption behavior of the user corresponding to the account information. Among them, the consumption behavior of the user includes, but is not limited to, clicking, playing, liking, and following, or no behavior at all.

[0141] In some optional embodiments, in response to an account consumption instruction for the candidate feedback content, if the account consumption instruction indicates that the candidate feedback content is consumed and the intent label of the candidate feedback content does not include intent information, the intent information can be added to the intent label of the candidate feedback content.

[0142] Optionally, if the user clicks on the first candidate feedback content in the candidate feedback content set, it can be determined that the account consumption instruction for the first candidate feedback content is that the first candidate feedback content is consumed. When the intent label of the first candidate feedback content does not include intent information, the intent information can be added to the intent label of the candidate feedback content.

[0143] In some other optional embodiments, in response to an account consumption instruction for the candidate feedback content, if the account consumption instruction indicates that the candidate feedback content is consumed and the intent label of the candidate feedback content includes intent information, the weight data of the intent information in the intent label of the candidate feedback content can be increased.

[0144] Optionally, if the user clicks on the first candidate feedback content in the candidate feedback content set, it can be determined that the account consumption instruction for the first candidate feedback content is that the first candidate feedback content is consumed. When the intent label of the first candidate feedback content includes intent information, the weight data of the intent information in the intent label of the candidate feedback content can be increased.

[0145] In some other optional embodiments, in response to an account consumption instruction for the candidate feedback content, if the account consumption instruction indicates that the candidate feedback content is not consumed and the intent label of the candidate feedback content includes intent information, update the consumption record of the intent information in the intent label of the candidate feedback content, where the consumption record is used to delete the intent information from the intent label of the candidate feedback content when the consumption record meets the deletion condition.

[0146] Optionally, if the user does not perform any consumption behavior on the first candidate feedback content in the candidate feedback content set, when the intention tag of the candidate feedback content contains intention information, the consumption record of the intention information in the intention tag of the candidate feedback content can be updated. Among them, the consumption record includes the number of unconsumed times. At this time, the original number of unconsumed times can be incremented by one. When the consumption record meets the deletion condition, such as when the number of unconsumed times reaches the preset number threshold, the intention information can be deleted from the intention tag of the candidate feedback content. Optionally, the consumption record can be for account information, that is, each intention tag of the candidate feedback content has a consumption record corresponding to each account information. Optionally, the consumption record can be for the entire system, that is, each intention tag of the candidate feedback content has a consumption record corresponding to the entire system.

[0147] In this way, the embodiment of the present application can update the intention tag of the candidate feedback content according to the user's consumption behavior of the candidate feedback content, so as to further improve the search ability of the first system by improving the accuracy of the matching process.

[0148] Please refer to Figure 7 , which shows a schematic flowchart of an information search method provided by an embodiment of the present invention. After displaying the candidate feedback content in the candidate feedback content set, it further includes:

[0149] S213. When it is detected that the candidate feedback content set has not been consumed and the updated feedback content in the updated feedback content set associated with the updated search information corresponding to the current search information has been consumed, determine the intention information corresponding to the updated search information.

[0150] In the embodiment of the present application, when the updated search information is the subordinate information or equivalent information of the current search information, the difference between the time when the current search information is obtained and the time when the updated search information is obtained is less than or equal to the time threshold, and the updated search information and the current search information correspond to the same account information, it is determined that the updated search information is the updated search information corresponding to the current search information.

[0151] Optionally, the subordinate information means that the meaning of the current search information includes the meaning of the updated search information. For example, when the current search information is "tool" and the updated search information is "tool wrench".

[0152] Optionally, the equivalent information means that the meaning of the current search information is the same as or similar to the meaning of the updated search information. For example, when the current search information is "name" and the updated search information is "appellation".

[0153] The fact that the candidate feedback content set has not been consumed can be manifested in practical applications as:

[0154] The first case: No operation is performed on the set of candidate feedback content displayed on the screen.

[0155] The second case: The user performs a sliding switching operation on the set of candidate feedback content displayed on the screen, but does not click on any of the candidate feedback content, let alone subsequent play, like, or follow.

[0156] Therefore, when it is detected that the set of candidate feedback content has not been consumed, the updated search information corresponding to the current search information can be determined according to the judgment conditions above. Optionally, the updated search information corresponding to the current search information can be the information obtained by deleting the current search information and re-entering. Optionally, the updated search information corresponding to the current search information can be the information obtained by adding text to the current search information. Optionally, the updated search information corresponding to the current search information can be the information obtained by modifying the text of the current search information.

[0157] Subsequently, the updated search information can be input into the intent recognition model to obtain the intent information corresponding to the updated search information. The intent information corresponding to the updated search information is matched with the feedback content in the feedback content set included in the first system to obtain a new matching result. According to the new matching result and the account information, the intent information corresponding to the updated search information is determined from the feedback content set. When the updated feedback content in the updated feedback content set associated with the updated search information is consumed, the intent information corresponding to the updated search information is determined.

[0158] S215, Update and train the intent recognition model based on the current search information and the intent information corresponding to the updated search information to obtain an updated intent recognition model.

[0159] Optionally, the intent recognition model can be updated and trained based on the current search information and the intent information corresponding to the updated search information to obtain an updated intent recognition model.

[0160] In this way, the user's behavior feedback can be used to help determine more intent information of the current search information, so as to train and update the intent recognition model, enabling the intent recognition model to continuously learn and keep up with the times.

[0161] In this way, the embodiments of the present application can update the intent recognition ability of the intent recognition model according to the user's consumption behavior of the candidate feedback content, so as to further improve the search ability of the first system by improving the accuracy of the model's intent recognition.

[0162] In summary, the embodiments of the present application can not only update the intent recognition ability of the intent recognition model according to the user's consumption behavior of the candidate feedback content, but also update the intent tags of the candidate feedback content according to the user's consumption behavior of the candidate feedback content, so as to further improve the search ability of the first system.

[0163] Please refer to Figure 8 which shows a schematic structural diagram of an information search device provided by an embodiment of the present invention. The device has the function of implementing the information search method in the above method embodiment. The function can be implemented by hardware or by hardware executing corresponding software. As Figure 8 shown, the information search device 800 may include:

[0164] An information acquisition module 801, configured to acquire current search information and account information;

[0165] An intent recognition module 802, configured to perform intent recognition on the current search information based on an intent recognition model to obtain intent information corresponding to the current search information; the intent recognition model is trained by combining a preset recognition model that has not been trained with a to-be-searched data set in at least one second system; the to-be-searched data set includes to-be-searched data and search feedback data; the association degree between the search attributes of the first system and the search attributes of each second system satisfies a preset association condition;

[0166] A content matching module 803, configured to perform matching processing on the intent information corresponding to the current search information and the feedback content in the feedback content set included in the first system to obtain a matching result;

[0167] A content determination module 804, configured to determine a candidate feedback content set from the feedback content set according to the matching result and the account information.

[0168] In some possible embodiments, the content matching module is configured to:

[0169] Determine the intent label of each feedback content in the feedback content set included in the first system;

[0170] Perform matching processing on the intent information corresponding to the current search information and the intent label of each feedback content to obtain a matching result; the matching result indicates the matching degree between the current search information and each feedback content.

[0171] In some possible embodiments, the device further includes an intent label modification module, configured to:

[0172] Display the candidate feedback content in the candidate feedback content set;

[0173] In response to an account consumption instruction for the candidate feedback content, update the intent label of the candidate feedback content based on the intent information.

[0174] In some possible embodiments, the intent label modification module is configured to:

[0175] In response to an account consumption instruction for candidate feedback content, if the account consumption instruction indicates that the candidate feedback content is consumed and the intent tag of the candidate feedback content does not contain intent information, add the intent information to the intent tag of the candidate feedback content;

[0176] Or;

[0177] In response to an account consumption instruction for candidate feedback content, if the account consumption instruction indicates that the candidate feedback content is consumed and the intent tag of the candidate feedback content contains intent information, increase the weight data of the intent information in the intent tag of the candidate feedback content.

[0178] In some possible embodiments, an intent tag modification module is configured to:

[0179] In response to an account consumption instruction for candidate feedback content, if the account consumption instruction indicates that the candidate feedback content is not consumed and the intent tag of the candidate feedback content contains intent information, update the consumption record of the intent information in the intent tag of the candidate feedback content;

[0180] The consumption record is used to delete the intent information from the intent tag of the candidate feedback content when the consumption record meets the deletion condition.

[0181] In some possible embodiments, the apparatus further includes an intent recognition model update module, configured to:

[0182] When it is detected that the candidate feedback content set is not consumed and the updated feedback content in the updated feedback content set associated with the updated search information corresponding to the current search information is consumed, determine the intent information corresponding to the updated search information;

[0183] Update and train the intent recognition model based on the current search information and the intent information corresponding to the updated search information to obtain an updated intent recognition model.

[0184] In some possible embodiments, an intent recognition model update module is configured to:

[0185] When the updated search information is a lower-level information or an equivalent information of the current search information, the difference between the time when the current search information is obtained and the time when the updated search information is obtained is less than or equal to a time threshold, and the updated search information and the current search information correspond to the same account information, determine that the updated search information is the updated search information corresponding to the current search information.

[0186] In some possible embodiments, the apparatus further includes an intent recognition model training module, configured to:

[0187] Determine the search attributes of the first system and the search attributes of the alternative system;

[0188] Determine a second system from alternative systems based on the search attributes of the first system and the search attributes of the alternative systems;

[0189] Obtain the data to be searched corresponding to the first system;

[0190] Obtain the search feedback data corresponding to the data to be searched from the second system;

[0191] Obtain the intent key data of the search feedback data;

[0192] Use the intent key data in the search feedback data as the annotation data of the data to be searched corresponding to the search feedback data to train a preset recognition model to obtain an intent recognition model.

[0193] In some possible embodiments, the intent recognition model training module is used for:

[0194] Use the key information determination model to perform key information determination on each search feedback data to obtain the key information corresponding to each search feedback data;

[0195] Perform universality statistical processing on the key information corresponding to each search feedback data to obtain a key information statistical result;

[0196] Determine the intent key data of multiple search feedback data based on the key information statistical result.

[0197] It should be noted that for the device provided in the above embodiments, when implementing its functions, only the division of the above function modules is used for illustration. In actual applications, the above functions can be allocated to different function modules according to needs, that is, the internal structure of the device is divided into different function modules to complete all or part of the functions described above. In addition, the device provided in the above embodiments and the method embodiments belong to the same concept, and the specific implementation process can be seen in the method embodiments, which will not be repeated here.

[0198] In the embodiments of the present application, the term "module" or "unit" refers to a computer program with a predetermined function or a part of a computer program, which works together with other related parts to achieve a predetermined goal, and can be fully or partially implemented by using software, hardware (such as a processing circuit or a memory), or a combination thereof. Similarly, a processor (or multiple processors or memories) can be used to implement one or more modules or units. In addition, each module or unit can be a part of an overall module or unit including the functions of the module or unit.

[0199] An embodiment of the present invention provides an electronic device, which includes a processor and a memory. At least one instruction or at least one program segment is stored in the memory, and the at least one instruction or the at least one program segment is loaded and executed by the processor to implement the information search method provided in the above method embodiment.

[0200] The memory can be used to store software programs and modules. By running the software programs and modules stored in the memory, the processor can execute various functional applications and determine clustering results. The memory mainly includes a program storage area and a data storage area. Among them, the program storage area can store an operating system, application programs required for functions, etc.; the data storage area can store data created according to the use of the device. In addition, the memory can include high-speed random access memory and can also include non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other volatile solid-state storage devices. Correspondingly, the memory can also include a memory controller to provide the processor with access to the memory.

[0201] The method embodiment provided by the embodiment of the present invention can be executed on a computer terminal, a server, or a similar computing device. Taking running on a server as an example, Figure 9 is a hardware structure block diagram of a server for running an information search method provided by an embodiment of the present invention. As Figure 9 shown, the server 2000 may vary greatly due to different configurations or performances, and may include one or more central processing units (CPUs) 2010 (the processor 2010 may include, but is not limited to, a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 2030 for storing data, and one or more storage media 2020 for storing application programs 2023 or data 2022 (such as one or more mass storage devices). Among them, the memory 2030 and the storage media 2020 can be transient storage or persistent storage. The program stored in the storage media 2020 can include one or more modules, and each module can include a series of instruction operations on the server. Further, the central processor 2010 can be set to communicate with the storage media 2020 and execute a series of instruction operations in the storage media 2020 on the server 2000. The server 2000 can also include one or more power supplies 2060, one or more wired or wireless network interfaces 2050, one or more input / output interfaces 2040, and / or one or more operating systems 2021, such as Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM, and so on.

[0202] The input / output interface 2040 can be used to receive or send data via a network. Specific examples of the above-mentioned network may include a wireless network provided by a communication provider of the server 2000. In one example, the input / output interface 2040 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the input / output interface 2040 can be a Radio Frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0203] Those of ordinary skill in the art can understand that Figure 9 The structure shown is only schematic and does not limit the structure of the above-mentioned electronic device. For example, the server 2000 may also include more or fewer components than those shown Figure 9 shown, or have a different configuration from that Figure 9 shown.

[0204] Embodiments of the present invention also provide a computer-readable storage medium, which can be disposed in an electronic device to store at least one instruction or at least one segment of a program related to implementing an information search method. The at least one instruction or the at least one segment of the program is loaded and executed by the processor to implement the information search method provided by the above-mentioned method embodiments.

[0205] Optionally, in this embodiment, the above-mentioned storage medium may include, but is not limited to: various media that can store program codes such as USB flash drives, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), mobile hard disks, magnetic disks, or optical discs.

[0206] It should be noted that: the above-mentioned sequence of embodiments of the present invention is only for description and does not represent the superiority or inferiority of the embodiments. And the above description of specific embodiments of this specification is provided. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be executed in a different order from that in the embodiments and still achieve the desired results. Additionally, the processes depicted in the figures do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0207] An embodiment of the present invention also provides a computer program product or a computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the above-mentioned information search method.

[0208] Each embodiment in this specification is described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other. Each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and for the relevant parts, reference can be made to the description of the method embodiment.

[0209] Those of ordinary skill in the art can understand that all or part of the steps for implementing the above embodiments can be completed by hardware, or can be completed by a program instructing relevant hardware. The program can be stored in a computer-readable storage medium. The storage medium mentioned above can be a read-only memory, a magnetic disk, an optical disk, or the like.

[0210] The above are only the preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included in the protection scope of the present invention.

Claims

1. An information search method, characterized in that, Applied to a first system, the method includes: Obtain current search information and account information; Perform intent recognition on the current search information based on an intent recognition model to obtain intent information corresponding to the current search information; the intent recognition model is obtained by training a preset recognition model that has not been trained in combination with a to-be-searched data set in at least one second system; the to-be-searched data set includes to-be-searched data and search feedback data; the correlation degree between the search attributes of the first system and the search attributes of each second system satisfies a preset correlation condition; Perform matching processing on the intent information corresponding to the current search information and the feedback content in the feedback content set included in the first system to obtain a matching result; Determine a candidate feedback content set from the feedback content set according to the matching result and the account information.

2. The information search method according to claim 1, wherein The performing matching processing on the intent information corresponding to the current search information and the feedback content in the feedback content set included in the first system to obtain a matching result includes: Determine the intent label of each feedback content in the feedback content set included in the first system; Perform matching processing on the intent information corresponding to the current search information and the intent label of each feedback content to obtain a matching result; the matching result indicates the matching degree between the current search information and each feedback content.

3. The information search method according to claim 2, characterized in that After determining the candidate feedback content set from the feedback content set according to the matching result and the account information, it further includes: Display the candidate feedback content in the candidate feedback content set; In response to an account consumption instruction for the candidate feedback content, update the intent label of the candidate feedback content based on the intent information.

4. The information search method according to claim 3, wherein The updating the intent label of the candidate feedback content based on the intent information in response to an account consumption instruction for the candidate feedback content includes: In response to an account consumption instruction for the candidate feedback content, if the account consumption instruction indicates that the candidate feedback content is consumed and the intent label of the candidate feedback content does not include the intent information, add the intent information to the intent label of the candidate feedback content; Or; In response to an account consumption instruction for the candidate feedback content, if the account consumption instruction indicates that the candidate feedback content is consumed and the intent label of the candidate feedback content includes the intent information, increase the weight data of the intent information in the intent label of the candidate feedback content.

5. The information search method according to claim 3, characterized in that, The updating the intent label of the candidate feedback content based on the intent information in response to an account consumption instruction for the candidate feedback content includes: In response to an account consumption instruction for the candidate feedback content, if the account consumption instruction indicates that the candidate feedback content is not consumed and the intent label of the candidate feedback content includes the intent information, update the consumption record of the intent information in the intent label of the candidate feedback content; The consumption record is used to delete the intent information from the intent label of the candidate feedback content when the consumption record meets the deletion condition.

6. The information search method according to claim 3, wherein After displaying the candidate feedback content in the candidate feedback content set, it further includes: When it is detected that the candidate feedback content set is not consumed and the updated feedback content in the updated feedback content set associated with the updated search information corresponding to the current search information is consumed, determine the intent information corresponding to the updated search information; Based on the current search information and the intent information corresponding to the updated search information, update and train the intent recognition model to obtain an updated intent recognition model.

7. The information search method according to claim 6, characterized in that The method further includes: When the updated search information is subordinate information or equivalent information of the current search information, the difference between the time when the current search information is obtained and the time when the updated search information is obtained is less than or equal to a time threshold, and the updated search information and the current search information correspond to the same account information, determine that the updated search information is the updated search information corresponding to the current search information.

8. The information search method according to any one of claims 1-7, characterized in that, Before the intent information corresponding to the current search information is obtained by performing intent recognition on the current search information based on the intent recognition model, it further includes: Determine the search attributes of the first system and the search attributes of the alternative system; Based on the search attributes of the first system and the search attributes of the alternative system, determine a second system from the alternative systems; Obtain the data to be searched corresponding to the first system; Obtain the search feedback data corresponding to the data to be searched from the second system; Obtain the intent key data of the search feedback data; Use the intent key data in the search feedback data as the annotation data corresponding to the data to be searched in the search feedback data to train the preset recognition model to obtain the intent recognition model.

9. The information search method according to claim 8, wherein The number of search feedback data corresponding to the data to be searched is multiple. Obtaining the intent key data in the search feedback data includes: Use the key information determination model to determine the key information for each search feedback data to obtain the key information corresponding to each search feedback data; Perform a universality statistical process on the key information corresponding to each search feedback data to obtain a key information statistical result; Based on the key information statistical result, determine the intent key data of the multiple search feedback data.

10. An information search device, characterized in that, Applied to the first system, the device includes: An information acquisition module, configured to acquire the current search information and the account information; An intent recognition module, configured to perform intent recognition on the current search information based on the intent recognition model to obtain the intent information corresponding to the current search information; the intent recognition model is obtained by training the untrained preset recognition model in combination with the data set to be searched in at least one second system; the data set to be searched includes the data to be searched and the search feedback data; the association degree between the search attributes of the first system and the search attributes of each second system satisfies a preset association condition; A content matching module, configured to perform a matching process on the intent information corresponding to the current search information and the feedback content in the feedback content set included in the first system to obtain a matching result; A content determination module, configured to determine a candidate feedback content set from the feedback content set according to the matching result and the account information.

11. An electronic device, characterized in that, It includes a processor and a memory, and at least one instruction or at least one program segment is stored in the memory. The at least one instruction or the at least one program segment is loaded and executed by the processor to implement the information search method according to any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that, At least one instruction or at least one program segment is stored in the computer-readable storage medium. The at least one instruction or the at least one program segment is loaded and executed by a processor to implement the information search method according to any one of claims 1 to 9.

Images