WO2022222037A1 - 一种基于图神经网络推理的可解释推荐方法 - Google Patents

一种基于图神经网络推理的可解释推荐方法 Download PDF

Info

Publication number
WO2022222037A1
WO2022222037A1 PCT/CN2021/088464 CN2021088464W WO2022222037A1 WO 2022222037 A1 WO2022222037 A1 WO 2022222037A1 CN 2021088464 W CN2021088464 W CN 2021088464W WO 2022222037 A1 WO2022222037 A1 WO 2022222037A1
Authority
WO
WIPO (PCT)
Prior art keywords
user
item
representation
user behavior
graph
Prior art date
Application number
PCT/CN2021/088464
Other languages
English (en)
French (fr)
Inventor
吕子钰
乔宇
Original Assignee
中国科学院深圳先进技术研究院
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 中国科学院深圳先进技术研究院 filed Critical 中国科学院深圳先进技术研究院
Priority to PCT/CN2021/088464 priority Critical patent/WO2022222037A1/zh
Publication of WO2022222037A1 publication Critical patent/WO2022222037A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90Details of database functions independent of the retrieved data types
    • G06F16/95Retrieval from the web
    • G06F16/953Querying, e.g. by the use of web search engines
    • G06F16/9535Search customisation based on user profiles and personalisation

Definitions

  • the present invention relates to the field of computer technology, and more particularly, to an interpretable recommendation method based on graph neural network reasoning.
  • recommender systems have been widely used in e-commerce, news recommendation, video recommendation and other fields, gradually changing people's lifestyles.
  • an online bookstore recommends books that users like based on their historical behavior information. If users can understand why the book is recommended with an easy-to-understand explanation, it can greatly improve the effectiveness of the recommendation (helping users make decisions quickly) and the persuasiveness of the recommendation (improving the possibility of users buying the book) ), which is the interpretability of recommender systems.
  • the explainable recommendation system presents the reasons and explanations of the recommendation while recommending items to users, which can not only improve the transparency of the recommendation system, improve the user's trust and acceptance of the recommendation system, but also improve the user's satisfaction with the recommended products and services. .
  • Explainable recommender systems have important research significance and application value in fields such as e-commerce, finance, medical care, and law where interpretability and transparency are critical.
  • explainable recommendation methods mainly include two categories: explainable recommendation methods based on explanation generation and explainable recommendation methods based on knowledge graphs.
  • many interpretable recommendation methods based on explanation generation have been proposed, which can generate textual recommendation explanations in the form of real human comments.
  • a research result proposes a convolutional neural network model combined with a dual attention mechanism to learn user preference representation and item feature representation from user/item reviews, and use local and global attention mechanisms to select informative words to assist score prediction , to enhance the interpretability of the model.
  • a research result proposes a neural scoring regression model to predict user-item ratings, while using a recurrent neural network such as Long Short-Term Memory (LSTM) based on an encoder-decoder framework.
  • LSTM Long Short-Term Memory
  • GRU Gate Recurrent Unit
  • Another research result uses matrix factorization for score prediction, and proposes a sequence-to-sequence explanation generative model based on generative adversarial networks.
  • interpretable recommendation schemes use text information extraction to generate recommended explanations, or use knowledge graph reasoning paths to enhance model interpretability, and achieve certain results, there are still problems in user behavior analysis modeling and interpretation quality. certain deficiencies.
  • interpretable recommendation methods based on explanation generation directly use generative models to generate recommended explanations. Although they can generate text explanations that are easy for humans to understand in the form of explanations, due to the problem of data sparsity, the generated explanations are of low quality and poor readability.
  • the interpretable recommendation method based on knowledge graph uses the facts and associations introduced from common knowledge bases such as DBPedia and FreeBase to construct knowledge graphs.
  • information dissemination and reasoning based on graph models can improve the recommendation performance, it introduces redundant noise.
  • Data it is difficult to deal with unrelated entities, which affects the recommendation efficiency and results.
  • methods based on knowledge graphs need to manually pre-set paths, rules, etc., and lack domain knowledge, resulting in homogenization of recommended interpretation results.
  • the purpose of the present invention is to overcome the above-mentioned defects of the prior art and provide an interpretable recommendation method based on graph neural network reasoning.
  • the technical solution of the present invention is to provide an interpretable recommendation method based on graph neural network reasoning.
  • the method includes the following steps:
  • the interaction matrix of user behavior construct a multi-relational user behavior graph, in which the node set of the graph is all users and all items, and the edges are user-item interaction relationships;
  • For the user behavior graph use the user behavior preference understanding model to learn high-level associations and propagate user behavior preferences, and obtain a vector representation of user, item, and user-item associations;
  • the fusion of user preference state representation, item state representation, and user-item association representation output by the user behavior preference understanding model is used as the input of the explanation generation model, and combined with the text comment set, the relevant explanation of the item recommended for a given user is obtained. .
  • the advantage of the present invention is that, for the problem of interpretable recommendation, an interpretable recommendation method based on graph neural network reasoning is proposed, and knowledge reasoning is used to deeply understand user behavior patterns and mine potential intentions of user behaviors, thereby realizing Comprehensive, fine-grained analysis and modeling of user behavior.
  • the present invention takes into account the accuracy and interpretability of the recommendation, provides personalized recommendations to users while generating high-quality and easy-to-understand recommendation explanations, and generates recommendation explanations (recommendation reasons) while providing recommendation results, thereby improving the recommendation system.
  • the transparency of the recommendation system and users' trust and acceptance of the recommendation system and improve the user's satisfaction with the recommended products and services.
  • FIG. 1 is a flowchart of an interpretable recommendation method based on graph neural network reasoning according to an embodiment of the present invention
  • FIG. 2 is a frame diagram of an interpretable recommendation method based on graph neural network reasoning according to an embodiment of the present invention.
  • the problem of the present invention can be defined as: given a set of users U, a set of items V, a user-item interaction matrix R and a set of text reviews D, for a given user u, recommend the top-N items (the ones that the user has not interacted with before). item), while generating an explanation Y for the recommended item.
  • the interpretable recommendation method based on graph neural network reasoning realizes comprehensive user behavior modeling by constructing a user behavior graph and using graph neural network reasoning to propagate multi-hop information.
  • the method includes four parts, which are user behavior graph construction, user behavior preference understanding based on graph neural network inference, user behavior interaction prediction, and explanation generation model.
  • the provided interpretable recommendation method based on graph neural network reasoning includes the following steps.
  • Step S110 for the interaction matrix of user behavior, construct a user behavior graph.
  • a multi-relational user behavior graph G is constructed, where the node set (node) of the graph G is all users and all items U ⁇ V. Edges are user-item interactions.
  • a graph G the graph embedding representation of the initial nodes and edges using random variables.
  • the graph embedding representation of user node is denoted as e u
  • the graph embedding representation of item node is denoted as e v
  • the edge representation of user-item interaction relationship is denoted as r uv .
  • Step S120 using a graph neural network to infer user behavior preference understanding.
  • a user behavior preference understanding model based on graph attention network which can not only learn high-level association information, but also use a multi-hop propagation mechanism to deeply and comprehensively understand user behavior preferences.
  • a complete vector representation of users, items, and user-item associations can be obtained. Specifically include the following steps:
  • Step S121 information dissemination of the user behavior graph attention network.
  • the information dissemination mechanism that is, the information dissemination between a node and its neighbor nodes (directly connected nodes).
  • its first-order neighbor node set is According to the first-order associativity of the graph structure, from the first-order neighbor nodes.
  • the information propagated to node e i is defined as:
  • Step S122 information fusion of the user behavior graph attention network.
  • Step S123 the user behavior graph attention network is updated.
  • the graph neural network employs multi-hop reasoning in a recursive manner to model the higher-order connectivity information of the graph structure. Therefore, the fused representation is subsequently updated to the new representation of the node. For example, at the kth update, the node is expressed as:
  • the representations of nodes and edges are finally output, that is, user state representation e u , item state representation ev , and user-item association representation r uv .
  • step S120 aiming at the deficiencies of existing methods in user behavior modeling, it is proposed to construct a user behavior preference understanding model based on graph neural network reasoning, so as to obtain a high-quality user behavior preference representation.
  • Step S130 using a neural network collaborative filtering framework to predict user behavior interaction.
  • a neural network collaborative filtering framework based on a neural network collaborative filtering framework, user representation, item representation, and user-item association representation are integrated, and a neural collaboration layer is used to implement recommendation prediction.
  • a three-layer neural collaborative filtering model (Neural CF layers) is applied to achieve user-item interaction prediction.
  • the network structure of the model follows a tower-like structure, where the lowest layer is the widest, and each subsequent layer has fewer neurons.
  • the output of each layer in the model is used as the input to the next layer.
  • the input of the neural network coordination layer of the first layer is the splicing of the user state representation e u , the item state representation ev , and the user-item association representation r uv , that is, [e u , e v , r uv ].
  • a hidden vector is generated, which is expressed as:
  • W l and b l are the weight matrix and bias matrix of the lth layer
  • a rectified linear unit ReLU can be used for the activation function used.
  • the hidden vector h l is mapped to the predicted score Expressed as:
  • W is the weight matrix and ⁇ is the sigmoid function.
  • is the sigmoid function.
  • Step S140 using the explanation generation model to obtain an intuitive text explanation for the recommended item.
  • an explanation generator based on an encoder-decoder framework is designed, a copy mechanism is introduced to extract relevant information from user/item source text reviews, and a high-quality, easy-to-understand recommendation explanation is generated by combining the generation mode and the copy mode.
  • the copy mode it is to calculate the probability of the word appearing in the original comment, and when recommending the item v j to the user ui , the copy source of the word is the set of words of the user comment document of ui and the item comment document of v j .
  • a recommendation explanation Y that recommends the item v j to the user ui is generated.
  • ij [y ij1 ,y ij2 , ..y ijT ], T is the interpretation length.
  • the method proposed by the present invention is an end-to-end method, so the feature representation learning, recommendation prediction and explanation generation model are trained and optimized by means of joint learning.
  • This end-to-end interpretable recommendation framework reduces the difficulty of training. , speeding up the training process while reducing memory requirements,
  • the present invention has at least the following advantages:
  • the proposed interpretable recommendation method based on graph neural network reasoning provides high-performance recommendation results while generating high-quality, easy-to-understand recommendation explanations to achieve both recommendation effectiveness and A recommended framework for interpretability.
  • a deep-level user behavior understanding model is proposed, which uses the high-order correlation information propagation in user behavior to perform deep-level user behavior reasoning, and realizes comprehensive and deep-level user behavior modeling and understanding.
  • Obtaining high-quality representations in three aspects: user, item, and user-item can effectively improve recommendation performance.
  • the present invention innovatively proposes a user behavior understanding model based on graph neural network reasoning.
  • User-item association representation realizes comprehensive and fine-grained user behavior modeling, and solves the problem of insufficiency of current user behavior analysis and modeling.
  • the present invention may be a system, method and/or computer program product.
  • the computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of the present invention.
  • a computer-readable storage medium may be a tangible device that can hold and store instructions for use by the instruction execution device.
  • the computer-readable storage medium may be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing.
  • Non-exhaustive list of computer readable storage media include: portable computer disks, hard disks, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM) or flash memory), static random access memory (SRAM), portable compact disk read only memory (CD-ROM), digital versatile disk (DVD), memory sticks, floppy disks, mechanically coded devices, such as printers with instructions stored thereon Hole cards or raised structures in grooves, and any suitable combination of the above.
  • RAM random access memory
  • ROM read only memory
  • EPROM erasable programmable read only memory
  • flash memory static random access memory
  • SRAM static random access memory
  • CD-ROM compact disk read only memory
  • DVD digital versatile disk
  • memory sticks floppy disks
  • mechanically coded devices such as printers with instructions stored thereon Hole cards or raised structures in grooves, and any suitable combination of the above.
  • Computer-readable storage media are not to be construed as transient signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (eg, light pulses through fiber optic cables), or through electrical wires transmitted electrical signals.
  • the computer readable program instructions described herein may be downloaded to various computing/processing devices from a computer readable storage medium, or to an external computer or external storage device over a network such as the Internet, a local area network, a wide area network, and/or a wireless network.
  • the network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and/or edge servers.
  • a network adapter card or network interface in each computing/processing device receives computer-readable program instructions from a network and forwards the computer-readable program instructions for storage in a computer-readable storage medium in each computing/processing device .
  • the computer program instructions for carrying out the operations of the present invention may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine dependent instructions, microcode, firmware instructions, state setting data, or instructions in one or more programming languages.
  • Source or object code written in any combination including object-oriented programming languages, such as Smalltalk, C++, Python, etc., and conventional procedural programming languages, such as the "C" language or similar programming languages.
  • the computer readable program instructions may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer, or entirely on the remote computer or server implement.
  • the remote computer may be connected to the user's computer through any kind of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (eg, using an Internet service provider through the Internet connect).
  • LAN local area network
  • WAN wide area network
  • custom electronic circuits such as programmable logic circuits, field programmable gate arrays (FPGAs), or programmable logic arrays (PLAs)
  • FPGAs field programmable gate arrays
  • PDAs programmable logic arrays
  • Computer readable program instructions are executed to implement various aspects of the present invention.
  • These computer readable program instructions may be provided to a processor of a general purpose computer, special purpose computer or other programmable data processing apparatus to produce a machine that causes the instructions when executed by the processor of the computer or other programmable data processing apparatus , resulting in means for implementing the functions/acts specified in one or more blocks of the flowchart and/or block diagrams.
  • These computer readable program instructions can also be stored in a computer readable storage medium, these instructions cause a computer, programmable data processing apparatus and/or other equipment to operate in a specific manner, so that the computer readable medium on which the instructions are stored includes An article of manufacture comprising instructions for implementing various aspects of the functions/acts specified in one or more blocks of the flowchart and/or block diagrams.
  • Computer readable program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other equipment to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other equipment to produce a computer-implemented process , thereby causing instructions executing on a computer, other programmable data processing apparatus, or other device to implement the functions/acts specified in one or more blocks of the flowcharts and/or block diagrams.
  • each block in the flowchart or block diagrams may represent a module, segment, or portion of instructions, which comprises one or more functions for implementing the specified logical function(s) executable instructions.
  • the functions noted in the blocks may occur out of the order noted in the figures. For example, two blocks in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved.
  • each block of the block diagrams and/or flowchart illustrations, and combinations of blocks in the block diagrams and/or flowchart illustrations can be implemented in dedicated hardware-based systems that perform the specified functions or actions , or can be implemented in a combination of dedicated hardware and computer instructions. It is well known to those skilled in the art that implementation in hardware, implementation in software, and implementation in a combination of software and hardware are all equivalent.

Landscapes

  • Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

本发明公开了一种基于图神经网络推理的可解释推荐方法。该方法包括:针对用户行为的交互矩阵,构建多关系用户行为图;对于用户行为图,利用用户行为偏好理解模型学习高阶的关联关系并传播用户行为偏好,得到用户、物品、用户-物品间关联的向量表示;将用户行为偏好理解模型输出的用户状态表示、物品状态表示、用户-物品关联表示输入至用户行为偏好理解模型,获得给定用户的物品推荐;将用户行为偏好理解模型输出的用户偏好状态表示、物品状态表示、用户-物品关联表示的融合作为解释生成模型的输入,并结合文本评论集合,获得为给定用户所推荐物品的相关解释。本发明能够在提供高性能的推荐结果的同时生成高质量、易于理解的推荐解释。

Description

一种基于图神经网络推理的可解释推荐方法 技术领域
本发明涉及计算机技术领域,更具体地,涉及一种基于图神经网络推理的可解释推荐方法。
背景技术
推荐系统作为缓解信息过载问题的重要工具,在电子商务、新闻推荐、视频推荐等领域得到了广泛的应用,逐渐改变着人们的生活方式。例如,网上书店根据用户的历史行为信息推荐用户喜欢的书籍。如果能用通俗易懂的解释让用户了解到为什么会推荐这本书,可以极大地提高推荐的有效性(帮助用户快速做出决策)和推荐的说服力(提高用户购买这本书的可能性),这就是推荐系统的可解释性。可解释推荐系统在向用户推荐物品的同时呈现推荐的原因及解释,不仅能够提升推荐系统的透明度,提高用户对推荐系统的信任度和接受度,还能提升用户对推荐产品和服务的满意度。可解释推荐系统在电子商务、金融、医疗、法律等可解释性和透明度至关重要的领域具有重要的研究意义和应用价值。
在现有技术中,可解释推荐方法主要包括两类:基于解释生成的可解释推荐方法和基于知识图谱的可解释推荐方法。近年来,随着自然语言处理技术发展,提出了很多基于解释生成的可解释推荐方法,可生成类似人类真实评论形式的文本推荐解释。例如,一项研究成果提出结合双注意力机制的卷积神经网络模型,从用户/物品评论中学习用户偏好表示和物品特征表示,采用本地和全局注意力机制选取有信息量的词辅助评分预测,增强模型的解释性。又如,一项研究成果提出了神经评分回归模型预测用户-物品评分,同时基于编码器-解码器(encoder-decoder)框架利用循环神经网络如长短期记忆网络(Long Short-Term Memory,LSTM)、门循环单元(Gate Recurrent Unit,GRU)生成文本解释。另外一项研究成果利用矩阵 分解进行评分预测,同时提出基于生成对抗网络的序列到序列的解释生成模型。
对于基于知识图谱的可解释推荐方法,将知识图谱作为辅助信息引入到推荐系统受到了越来越多的关注,知识图谱能够引入多种类型的关联信息。研究者们利用知识图谱的推理路径增强模型的透明度和解释性,提出了一些基于知识图谱的可解释推荐方法。一项研究中提出端到端的知识图谱感知推荐方法RippleNet,从知识图谱的关联路径中自动发现传播路径,并利用信息传播机制发现用户层次化的潜在兴趣。另一项研究成果提出知识感知路径的循环神经网络,结合知识图谱实体和关系的表示生成路径表示,利用路径的序列化依赖关系推理用户偏好,应用推理路径提供推荐解释。
然而,尽管现有的可解释推荐方案利用文本信息提取生成推荐解释,或利用知识图谱推理路径增强模型的可解释性,取得了一定的效果,但是在用户行为分析建模和解释质量方面还存在一定的不足。此外,基于解释生成的可解释推荐方法直接采用生成模型产生推荐解释,虽然在解释形式上能够生成易于人类理解的文本解释,但是由于数据稀疏性问题,导致生成的解释质量低、可读性差。
此外,基于知识图谱的可解释推荐方法利用常识知识库如DBPedia、FreeBase等引入常识库的事实及关联构建知识图,基于图模型的信息传播和推理虽然能够提升推荐性能,但是引入了冗余噪声数据,难以处理无关联的实体,影响了推荐效率和结果。并且,基于知识图谱的方法需人工预先设定路径、规则等,缺少领域知识,导致推荐解释结果同质化。
发明内容
本发明的目的是克服上述现有技术的缺陷,提供一种基于图神经网络推理的可解释推荐方法。
本发明的技术方案是提供一种基于图神经网络推理的可解释推荐方法。该方法包括以下步骤:
针对用户行为的交互矩阵,构建多关系用户行为图,其中图的节点集合为所有用户和所有物品,边为用户-物品交互关系;
对于所述用户行为图,利用用户行为偏好理解模型学习高阶的关联关系并传播用户行为偏好,得到用户、物品、用户-物品间关联的向量表示;
将所述用户行为偏好理解模型输出的用户状态表示、物品状态表示、用户-物品关联表示输入至用户行为偏好理解模型,获得给定用户的物品推荐;
将所述用户行为偏好理解模型输出的用户偏好状态表示、物品状态表示、用户-物品关联表示的融合作为解释生成模型的输入,并结合文本评论集合,获得为给定用户所推荐物品的相关解释。
与现有技术相比,本发明的优点在于,针对可解释性推荐问题,提出了基于图神经网络推理的可解释推荐方法,利用知识推理深入理解用户行为模式,挖掘用户行为潜在意图,从而实现对用户行为全面地、细粒度地分析和建模。并且,本发明兼顾推荐准确性和可解释性,向用户提供个性化推荐的同时生成高质量、易于用户理解的推荐解释,在提供推荐结果的同时生成推荐解释(推荐理由),提升了推荐系统的透明度以及用户对推荐系统的信任度和接受度,并提升了用户对推荐产品和服务的满意度。
通过以下参照附图对本发明的示例性实施例的详细描述,本发明的其它特征及其优点将会变得清楚。
附图说明
被结合在说明书中并构成说明书的一部分的附图示出了本发明的实施例,并且连同其说明一起用于解释本发明的原理。
图1是根据本发明一个实施例的基于图神经网络推理的可解释推荐方法的流程图;
图2是根据本发明一个实施例的基于图神经网络推理的可解释推荐方法的框架图。
具体实施方式
现在将参照附图来详细描述本发明的各种示例性实施例。应注意到:除非另外具体说明,否则在这些实施例中阐述的部件和步骤的相对布置、数字表达式和数值不限制本发明的范围。
以下对至少一个示例性实施例的描述实际上仅仅是说明性的,决不作为对本发明及其应用或使用的任何限制。
对于相关领域普通技术人员已知的技术、方法和设备可能不作详细讨论,但在适当情况下,所述技术、方法和设备应当被视为说明书的一部分。
在这里示出和讨论的所有例子中,任何具体值应被解释为仅仅是示例性的,而不是作为限制。因此,示例性实施例的其它例子可以具有不同的值。
应注意到:相似的标号和字母在下面的附图中表示类似项,因此,一旦某一项在一个附图中被定义,则在随后的附图中不需要对其进行进一步讨论。
在以下描述的实施例中,设数据集中具有m个用户,记为u;n个物品,记为v。用户与物品的交互行为表示为<u i,v j,r ij,d ij>,其中r ij表示用户u i对物品v j的评分,d ij表示用户u i对物品v j的文本评论,用户-物品交互矩阵表示为R,文本评论集合表示为D,所有文本评论构成的词典表示为V g
本发明的问题可定义为:给定用户集合U,物品集合V,用户-物品交互矩阵R和文本评论集合D,对于给定的用户u,推荐top-N个物品(该用户未交互过的物品),同时生成推荐物品的解释Y。
简言之,本发明提供的基于图神经网络推理的可解释推荐方法,通过构建用户行为图,利用图神经网络推理进行多跳信息传播,实现全面地用户行为建模。总体上,该方法包括四部分内容,分别是用户行为图构建、基于图神经网络推理的用户行为偏好理解、用户行为交互预测、解释生成模型。
具体地,结合图1和图2所示,所提供的基于图神经网络推理的可解释推荐方法包括以下步骤。
步骤S110,对于用户行为的交互矩阵,构建用户行为图。
针对用户行为交互矩阵R,构建多关系用户行为图G,其中图G的节 点集合(node)为所有用户和所有物品U∪V。边为用户-物品交互关系。对于图G,利用随机变量初始节点和边的图嵌入表示。用户节点的图嵌入表示记为e u,物品节点的图嵌入表示记为e v,用户-物品交互关系边表示记为r uv。通过构建用户行为图,有利于后续的用户行为理解。
步骤S120,利用图神经网络推理出用户行为偏好理解。
在一个实施例中,提出基于图注意力网络的用户行为偏好理解模型,不仅可以学习高阶的关联信息,还可以利用多跳(multi-hop)传播机制深层次地、全面地理解用户行为偏好,可得到完整的用户、物品、用户-物品间关联的向量表示。具体包括以下步骤:
步骤S121,用户行为图注意力网络的信息传播。
定义信息传播机制,即节点与它邻居节点间(直接连接节点)的信息传播。对于节点e i,它的一阶邻居节点集合为
Figure PCTCN2021088464-appb-000001
根据图结构的一阶关联性,从一阶邻居节点
Figure PCTCN2021088464-appb-000002
传播到节点e i的信息定义为:
Figure PCTCN2021088464-appb-000003
其中,
Figure PCTCN2021088464-appb-000004
测量了一对关系节点间的影响程度,δ为非线性的激活函数,可以选用ReLU或者tanh函数等,W t,W r,W e,W n为学习参数,b是偏置参数。
Figure PCTCN2021088464-appb-000005
同时考虑了关联的节点e i、e j和它们间关联关系的影响r ij,γ为正则化后的影响度量值。因此邻居节点向当前节点e i传播的信息是权重化地邻居节点表示组合
Figure PCTCN2021088464-appb-000006
步骤S122,用户行为图注意力网络的信息融合。
在得到(一阶)邻接节点的信息传播后,融合节点自身的嵌入表示e i和来自邻接节点的传播信息
Figure PCTCN2021088464-appb-000007
Figure PCTCN2021088464-appb-000008
其中
Figure PCTCN2021088464-appb-000009
是所使用的融合函数,可采用图卷积网络的非线性加权求和等融合方式。
步骤S123,用户行为图注意力网络的更新。
在一个实施例中,图神经网络采用递归方式利用多跳(multi-hop)推理对图结构的高阶连接信息进行建模。因此后续用融合后的表示更新为节 点的新的表示。例如,在第k次更新,节点
Figure PCTCN2021088464-appb-000010
的表示为:
Figure PCTCN2021088464-appb-000011
通过上述过程,最终输出节点和边的表示,即用户状态表示e u、物品状态表示e v、用户-物品关联表示r uv
在此步骤S120中,针对现有方法在用户行为建模方面的不足,提出基于图神经网络推理构建用户行为偏好理解模型,以获得高质量的用户行为偏好表示。
步骤S130,利用神经网络协同过滤框架对用户行为交互进行预测。
在一个实施例中,基于神经网络协同过滤框架,融合用户表示、物品表示、用户-物品关联表示,采用神经协同层实现推荐预测。
例如,应用三层神经协同过滤模型(Neural CF layers)实现用户-物品交互预测。模型的网络结构服从塔状结构,其中最低层最宽,每个接下来的层会有更少的神经元。模型中每一层的输出会作为下一层的输入。具体地,第一层的神经网络协同层的输入为用户状态表示e u、物品状态表示e v、用户-物品关联表示r uv的拼接,即[e u,e v,r uv]。在每一神经网络协同层,生成隐藏的向量,表示为:
Figure PCTCN2021088464-appb-000012
其中W l和b l为第l层的权重矩阵和偏置矩阵,
Figure PCTCN2021088464-appb-000013
为使用的激活函数,可使用修正线性单元ReLU。在最后一层,隐藏向量h l被映射为预测评分
Figure PCTCN2021088464-appb-000014
表示为:
Figure PCTCN2021088464-appb-000015
其中W为权重矩阵,σ为sigmoid函数。对于用户u,根据预测评分排序
Figure PCTCN2021088464-appb-000016
推荐top-N个物品,即前N个推荐物品,N是预先设定的整数。
步骤S140,利用解释生成模型,获得针对推荐物品的直观文本解释。
在一个实施例中,设计基于编码器-解码器框架的解释生成器,引入拷贝机制从用户/物品源文本评论中抽取相关信息,结合生成模式和拷贝模式生成高质量、易于理解的推荐解释。
具体地,应用循环神经网络GRU作为解释生成器,并引入拷贝机制从用户原有评论中抽取信息,结合两种模式(生成模式和拷贝模式)生成直观文本解释Y=[y 1,y 2,..y T](词序列),易于用户阅读理解。
对于生成模式,利用GRU作为解码器,其中解码器的初始状态
Figure PCTCN2021088464-appb-000017
为用户偏好状态表示e u、物品状态表示e v、用户-物品关联表示r uv的融合,用户u i推荐物品v j时生成的解释词y ijt为w t的概率为P gen(y ijt=w t),表示为:
Figure PCTCN2021088464-appb-000018
其中,
Figure PCTCN2021088464-appb-000019
为在t时刻GRU的隐藏状态,w t是词的向量表示,w t来自整个词典V g
对于拷贝模式,是计算出现在原评论中的词的概率,在向用户u i推荐物品v j时解释词的拷贝源为u i的用户评论文档和v j的物品评论文档的词的集合
Figure PCTCN2021088464-appb-000020
在拷贝模式下,解释词y ijt为w t的概率表示为P copy(y ijt=w t):
Figure PCTCN2021088464-appb-000021
优选地,结合生成模式和拷贝模式,构成解释生成模型,最终生成目标词的概率为P(y ijt=w t),参照公式(6),生成向用户u i推荐物品v j的推荐解释Y ij=[y ij1,y ij2,..y ijT],T为解释长度。
P(y ijt=w t)=P gen(y ijt=w t)+P copy(y ijt=w t)      (6)
通过设计结合生成模式和拷贝机制的解释生成模型,相对于现有的基于解释生成的可解释推荐方法能够获得高质量的、易于用户理解的推荐解释,解决了目前解释质量低、可读性差的问题。
需说明的是,本发明提出的方法是端到端的方法,因此特征表示学习、推荐预测和解释生成模型采用联合学习的方式训练和优化,这种端到端的可解释推荐框架,降低了训练难度,加速了训练过程,同时减少了内存需求,
为了验证本发明的有效性和先进性,利用提出的基于图神经网络推理的可解释方法,对Amazon-Electronics和Amazon-Music Instrument这两个数据集进行了大量实验。实验结果表明,本发明始终优于可解释性推荐当前最好方法,展现出巨大优越性,具有十分广阔的应用前景。
相对于现有技术,本发明至少具有以下优势:
(1)针对可解释性推荐任务,所提出的基于图神经网络推理的可解 释推荐方法,在提供高性能的推荐结果的同时生成高质量、易于人类理解的推荐解释,实现兼顾推荐有效性和可解释性的推荐框架。
(2)基于图神经网络推理提出了深层次的用户行为理解模型,利用用户行为中的高阶关联信息传播执行深层次的用户行为推理,实现全面的、深层次的用户行为建模及理解,在用户、物品、用户-物品三方面获得高质量的表征,有效地提高推荐性能。
(3)在解释生成模型中,引入了拷贝机制,提出了生成模式和拷贝模式相结合的解释生成模型,有利于生成高质量、易于人类理解的推荐解释。
综上所述,本发明创新地提出一种基于图神经网络推理的用户行为理解模型,通过构建用户行为图,基于图神经网络推理进行多跳信息传播,学习深层次、高质量的用户、物品、用户-物品关联表示,实现全面的、细粒度的用户行为建模,解决了目前用户行为分析建模的不足的问题。
应理解的是,在不违背本发明精神和范围的前提下,本领域技术人员可对上述实施例进行适当的改变或变型,例如,采用LSTM替代GRU网络,或者采用其他的激活函数用于非线性处理等。
本发明可以是系统、方法和/或计算机程序产品。计算机程序产品可以包括计算机可读存储介质,其上载有用于使处理器实现本发明的各个方面的计算机可读程序指令。
计算机可读存储介质可以是可以保持和存储由指令执行设备使用的指令的有形设备。计算机可读存储介质例如可以是――但不限于――电存储设备、磁存储设备、光存储设备、电磁存储设备、半导体存储设备或者上述的任意合适的组合。计算机可读存储介质的更具体的例子(非穷举的列表)包括:便携式计算机盘、硬盘、随机存取存储器(RAM)、只读存储器(ROM)、可擦式可编程只读存储器(EPROM或闪存)、静态随机存取存储器(SRAM)、便携式压缩盘只读存储器(CD-ROM)、数字多功能盘(DVD)、记忆棒、软盘、机械编码设备、例如其上存储有指令的打孔卡或凹槽内凸起结构、以及上述的任意合适的组合。这里所使用的计算机可读存储介质不被解释为瞬时信号本身,诸如无线电波或者其他自由 传播的电磁波、通过波导或其他传输媒介传播的电磁波(例如,通过光纤电缆的光脉冲)、或者通过电线传输的电信号。
这里所描述的计算机可读程序指令可以从计算机可读存储介质下载到各个计算/处理设备,或者通过网络、例如因特网、局域网、广域网和/或无线网下载到外部计算机或外部存储设备。网络可以包括铜传输电缆、光纤传输、无线传输、路由器、防火墙、交换机、网关计算机和/或边缘服务器。每个计算/处理设备中的网络适配卡或者网络接口从网络接收计算机可读程序指令,并转发该计算机可读程序指令,以供存储在各个计算/处理设备中的计算机可读存储介质中。
用于执行本发明操作的计算机程序指令可以是汇编指令、指令集架构(ISA)指令、机器指令、机器相关指令、微代码、固件指令、状态设置数据、或者以一种或多种编程语言的任意组合编写的源代码或目标代码,所述编程语言包括面向对象的编程语言—诸如Smalltalk、C++、Python等,以及常规的过程式编程语言—诸如“C”语言或类似的编程语言。计算机可读程序指令可以完全地在用户计算机上执行、部分地在用户计算机上执行、作为一个独立的软件包执行、部分在用户计算机上部分在远程计算机上执行、或者完全在远程计算机或服务器上执行。在涉及远程计算机的情形中,远程计算机可以通过任意种类的网络—包括局域网(LAN)或广域网(WAN)—连接到用户计算机,或者,可以连接到外部计算机(例如利用因特网服务提供商来通过因特网连接)。在一些实施例中,通过利用计算机可读程序指令的状态信息来个性化定制电子电路,例如可编程逻辑电路、现场可编程门阵列(FPGA)或可编程逻辑阵列(PLA),该电子电路可以执行计算机可读程序指令,从而实现本发明的各个方面。
这里参照根据本发明实施例的方法、装置(系统)和计算机程序产品的流程图和/或框图描述了本发明的各个方面。应当理解,流程图和/或框图的每个方框以及流程图和/或框图中各方框的组合,都可以由计算机可读程序指令实现。
这些计算机可读程序指令可以提供给通用计算机、专用计算机或其它可编程数据处理装置的处理器,从而生产出一种机器,使得这些指令在通 过计算机或其它可编程数据处理装置的处理器执行时,产生了实现流程图和/或框图中的一个或多个方框中规定的功能/动作的装置。也可以把这些计算机可读程序指令存储在计算机可读存储介质中,这些指令使得计算机、可编程数据处理装置和/或其他设备以特定方式工作,从而,存储有指令的计算机可读介质则包括一个制造品,其包括实现流程图和/或框图中的一个或多个方框中规定的功能/动作的各个方面的指令。
也可以把计算机可读程序指令加载到计算机、其它可编程数据处理装置、或其它设备上,使得在计算机、其它可编程数据处理装置或其它设备上执行一系列操作步骤,以产生计算机实现的过程,从而使得在计算机、其它可编程数据处理装置、或其它设备上执行的指令实现流程图和/或框图中的一个或多个方框中规定的功能/动作。
附图中的流程图和框图显示了根据本发明的多个实施例的系统、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上,流程图或框图中的每个方框可以代表一个模块、程序段或指令的一部分,所述模块、程序段或指令的一部分包含一个或多个用于实现规定的逻辑功能的可执行指令。在有些作为替换的实现中,方框中所标注的功能也可以以不同于附图中所标注的顺序发生。例如,两个连续的方框实际上可以基本并行地执行,它们有时也可以按相反的顺序执行,这依所涉及的功能而定。也要注意的是,框图和/或流程图中的每个方框、以及框图和/或流程图中的方框的组合,可以用执行规定的功能或动作的专用的基于硬件的系统来实现,或者可以用专用硬件与计算机指令的组合来实现。对于本领域技术人员来说公知的是,通过硬件方式实现、通过软件方式实现以及通过软件和硬件结合的方式实现都是等价的。
以上已经描述了本发明的各实施例,上述说明是示例性的,并非穷尽性的,并且也不限于所披露的各实施例。在不偏离所说明的各实施例的范围和精神的情况下,对于本技术领域的普通技术人员来说许多修改和变更都是显而易见的。本文中所用术语的选择,旨在最好地解释各实施例的原理、实际应用或对市场中的技术改进,或者使本技术领域的其它普通技术人员能理解本文披露的各实施例。本发明的范围由所附权利要求来限定。

Claims (10)

  1. 一种基于图神经网络推理的可解释推荐方法,包括以下步骤:
    步骤S1:针对用户行为的交互矩阵,构建多关系用户行为图,其中图的节点集合为所有用户和所有物品,边为用户-物品交互关系;
    步骤S2:对于所述用户行为图,利用用户行为偏好理解模型学习高阶的关联关系并传播用户行为偏好,得到用户、物品、用户-物品间关联的向量表示;
    步骤S3:将所述用户行为偏好理解模型输出的用户状态表示、物品状态表示、用户-物品关联表示输入至用户行为偏好理解模型,获得给定用户的物品推荐;
    步骤S4:将所述用户行为偏好理解模型输出的用户偏好状态表示、物品状态表示、用户-物品关联表示的融合作为解释生成模型的输入,并结合文本评论集合,获得为给定用户所推荐物品的相关解释。
  2. 根据权利要求1所述的方法,其特征在于,步骤S2包括:
    对于用户行为图G,根据图结构的一阶关联性,从一阶邻居节点
    Figure PCTCN2021088464-appb-100001
    传播到节点e i的信息定义为:
    Figure PCTCN2021088464-appb-100002
    Figure PCTCN2021088464-appb-100003
    Figure PCTCN2021088464-appb-100004
    其中,
    Figure PCTCN2021088464-appb-100005
    是一对关系节点e j和e i间的影响程度,δ是非线性激活函数,W t,W r,W e,W n是相关项权重,b是偏置,r ij表示节点间关联关系的影响,γ为正则化后的影响度量值,邻居节点向当前节点e i传播的信息是权重化地邻居节点表示组合
    Figure PCTCN2021088464-appb-100006
    节点e i的一阶邻居节点集合为
    Figure PCTCN2021088464-appb-100007
    在得到一阶邻接节点的信息传播后,融合节点自身的嵌入表示e i和来自邻接节点的传播信息
    Figure PCTCN2021088464-appb-100008
    表示为
    Figure PCTCN2021088464-appb-100009
    其中
    Figure PCTCN2021088464-appb-100010
    为所使用的融合函数;
    采用递归方式利用多跳推理对图结构的高阶连接信息进行建模,最终 输出节点和边的表示,标记为用户状态e u、物品状态e v、用户-物品关联表示r uv
  3. 根据权利要求1所述的方法,其特征在于,步骤S3包括:
    所述用户行为偏好理解模型采用塔状结构的神经协同过滤模型,其第一层的神经网络协同层的输入为用户状态表示、物品状态表示、用户-物品关联表示的拼接,对于每一神经网络协同层,生成隐藏的向量所示为:
    Figure PCTCN2021088464-appb-100011
    其中,W l和b l为第l层的权重矩阵和偏置矩阵,
    Figure PCTCN2021088464-appb-100012
    是激活函数,h l-1是第l-1层输出的向量;
    在神经协同过滤模型的最后一层,隐藏向量h l被映射为推荐物品的预测评分
    Figure PCTCN2021088464-appb-100013
    对于给定用户,根据预测评分
    Figure PCTCN2021088464-appb-100014
    排序并推荐物品。
  4. 根据权利要求1所述的方法,其中,在步骤S4中,结合拷贝模式从用户原有评论中抽取信息,生成给定用户u i推荐物品v j的推荐解释Y ij=[y ij1,y ij2,..y ijT],表示为:
    P(y ijt=w t)=P gen(y ijt=w t)+P copy(y ijt=w t)
    其中,T是解释长度,P gen(y ijt=w t)是利用生成模式获得的用户u i推荐物品v j时生成的解释词y ijt为w t的概率,P copy(y ijt=w t)是采用拷贝机制获得的用户u i推荐物品v j时生成的解释词y ijt为w t的概率。
  5. 根据权利要求4所述的方法,其特征在于,对于生成模式,采用循环神经网络GRU作为解码器,该解码器的初始状态
    Figure PCTCN2021088464-appb-100015
    为用户偏好状态表示e u、物品状态表示e v、用户-物品关联表示r uv的融合,用户u i推荐物品v j时生成的解释词y ijt为w t的概率表示为:
    Figure PCTCN2021088464-appb-100016
    Figure PCTCN2021088464-appb-100017
    其中,
    Figure PCTCN2021088464-appb-100018
    是在t时刻GRU的隐藏状态,w t是词的向量表示,w t来自词典V g
  6. 根据权利要求5所述的方法,其中,在拷贝模式下,用户u i推荐物品v j时生成的解释词y ijt为w t的概率表示为:
    Figure PCTCN2021088464-appb-100019
    其中,V c是u i的用户评论文档和v j的物品评论文档的词的集合。
  7. 根据权利要求3所述的方法,其特征在于,所述在神经协同过滤模型的最后一层,隐藏向量h l被映射为推荐物品的预测评分
    Figure PCTCN2021088464-appb-100020
    表示为:
    Figure PCTCN2021088464-appb-100021
    其中,W为权重矩阵,σ为sigmoid函数,h l是隐藏向量,l是层索引。
  8. 根据权利要求2所述的方法,其特征在于,融合函数
    Figure PCTCN2021088464-appb-100022
    是非线性加权求和。
  9. 一种计算机可读存储介质,其上存储有计算机程序,其中,该程序被处理器执行时实现根据权利要求1至8中任一项所述方法的步骤。
  10. 一种计算机设备,包括存储器和处理器,在所述存储器上存储有能够在处理器上运行的计算机程序,其特征在于,所述处理器执行所述程序时实现权利要求1至8中任一项所述的方法的步骤。
PCT/CN2021/088464 2021-04-20 2021-04-20 一种基于图神经网络推理的可解释推荐方法 WO2022222037A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/CN2021/088464 WO2022222037A1 (zh) 2021-04-20 2021-04-20 一种基于图神经网络推理的可解释推荐方法

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2021/088464 WO2022222037A1 (zh) 2021-04-20 2021-04-20 一种基于图神经网络推理的可解释推荐方法

Publications (1)

Publication Number Publication Date
WO2022222037A1 true WO2022222037A1 (zh) 2022-10-27

Family

ID=83723643

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/088464 WO2022222037A1 (zh) 2021-04-20 2021-04-20 一种基于图神经网络推理的可解释推荐方法

Country Status (1)

Country Link
WO (1) WO2022222037A1 (zh)

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116308683A (zh) * 2023-05-17 2023-06-23 武汉纺织大学 基于知识图谱的服装品牌定位推荐方法、设备及存储介质
CN116564408A (zh) * 2023-04-28 2023-08-08 上海科技大学 基于知识图谱推理的合成致死基因对预测方法、装置、设备及介质
CN116645174A (zh) * 2023-07-27 2023-08-25 山东省人工智能研究院 基于解耦多行为表征学习的个性化推荐方法
CN116720004A (zh) * 2023-08-09 2023-09-08 腾讯科技(深圳)有限公司 推荐理由生成方法、装置、设备及存储介质
CN117150150A (zh) * 2023-10-31 2023-12-01 中国科学技术大学 一种基于图信号处理的群体推荐方法
CN117171448A (zh) * 2023-08-11 2023-12-05 哈尔滨工业大学 一种基于图神经网络的多行为社会化推荐方法及系统
CN117216417A (zh) * 2023-11-07 2023-12-12 北京智谱华章科技有限公司 融合知识信息和协同信息的推荐方法、装置、设备及介质
CN117540111A (zh) * 2024-01-09 2024-02-09 安徽农业大学 一种基于图神经网络的偏好感知社会化推荐方法
CN117556224A (zh) * 2024-01-12 2024-02-13 国网四川省电力公司电力科学研究院 一种电网设施抗震风险评估系统、方法及存储介质
CN117992680A (zh) * 2024-04-02 2024-05-07 华侨大学 基于改进协同过滤及惩罚标签评分的多样化推荐方法
CN118013131A (zh) * 2024-04-10 2024-05-10 南昌大学 基于评分增强的图神经网络数据推荐方法、装置及介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10664929B2 (en) * 2018-06-19 2020-05-26 International Business Machines Corporation Extracting data for professional event and participant grouping recommendations
CN111259238A (zh) * 2020-01-13 2020-06-09 山西大学 一种基于矩阵分解的事后可解释性推荐方法及装置
CN111523047A (zh) * 2020-04-13 2020-08-11 中南大学 基于图神经网络的多关系协同过滤算法
CN111666496A (zh) * 2020-06-09 2020-09-15 中森云链(成都)科技有限责任公司 一种基于评论文本的组推荐方法

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10664929B2 (en) * 2018-06-19 2020-05-26 International Business Machines Corporation Extracting data for professional event and participant grouping recommendations
CN111259238A (zh) * 2020-01-13 2020-06-09 山西大学 一种基于矩阵分解的事后可解释性推荐方法及装置
CN111523047A (zh) * 2020-04-13 2020-08-11 中南大学 基于图神经网络的多关系协同过滤算法
CN111666496A (zh) * 2020-06-09 2020-09-15 中森云链(成都)科技有限责任公司 一种基于评论文本的组推荐方法

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
GUODONG WU, ZHA ZHIKANG; TU LIJING; TAO HONG; SONG FUGENG: "Research advances in graph neural network recommendation CAAI Transactions on Intelligent Systems", CAAI TRANSACTIONS ON INTELLIGENT SYSTEMS, vol. 15, no. 1, 31 January 2020 (2020-01-31), pages 14 - 24, XP055979620, ISSN: 1673-4785, DOI: 10.11992/tis.201908034 *
LI ZHAONING: "Research and Implementation of Recommendation Technology Based on Multi-scene Session Data", MASTER THESIS, TIANJIN POLYTECHNIC UNIVERSITY, CN, no. 7, 15 July 2017 (2017-07-15), CN , XP055979623, ISSN: 1674-0246 *
LIU MENGJUAN, WANG WEI; LI YANG-XI; LUO XU-CHENG; QIN ZHI-GUANG: "AttentionRank+: A Graph-Based Recommendation Combining Attention Relationship and Multi-Behaviors", CHINESE JOURNAL OF COMPUTERS, BEIJING., CN, vol. 40, no. 3, 31 March 2017 (2017-03-31), CN , pages 634 - 648, XP055979626, ISSN: 0254-4164 *

Cited By (21)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116564408A (zh) * 2023-04-28 2023-08-08 上海科技大学 基于知识图谱推理的合成致死基因对预测方法、装置、设备及介质
CN116564408B (zh) * 2023-04-28 2024-03-01 上海科技大学 基于知识图谱推理的合成致死基因对预测方法、装置、设备及介质
CN116308683A (zh) * 2023-05-17 2023-06-23 武汉纺织大学 基于知识图谱的服装品牌定位推荐方法、设备及存储介质
CN116645174A (zh) * 2023-07-27 2023-08-25 山东省人工智能研究院 基于解耦多行为表征学习的个性化推荐方法
CN116645174B (zh) * 2023-07-27 2023-10-17 山东省人工智能研究院 基于解耦多行为表征学习的个性化推荐方法
CN116720004B (zh) * 2023-08-09 2023-12-15 腾讯科技(深圳)有限公司 推荐理由生成方法、装置、设备及存储介质
CN116720004A (zh) * 2023-08-09 2023-09-08 腾讯科技(深圳)有限公司 推荐理由生成方法、装置、设备及存储介质
CN117171448A (zh) * 2023-08-11 2023-12-05 哈尔滨工业大学 一种基于图神经网络的多行为社会化推荐方法及系统
CN117171448B (zh) * 2023-08-11 2024-05-28 哈尔滨工业大学 一种基于图神经网络的多行为社会化推荐方法及系统
CN117150150B (zh) * 2023-10-31 2024-02-09 中国科学技术大学 一种基于图信号处理的群体推荐方法
CN117150150A (zh) * 2023-10-31 2023-12-01 中国科学技术大学 一种基于图信号处理的群体推荐方法
CN117216417A (zh) * 2023-11-07 2023-12-12 北京智谱华章科技有限公司 融合知识信息和协同信息的推荐方法、装置、设备及介质
CN117216417B (zh) * 2023-11-07 2024-02-20 北京智谱华章科技有限公司 融合知识信息和协同信息的推荐方法、装置、设备及介质
CN117540111A (zh) * 2024-01-09 2024-02-09 安徽农业大学 一种基于图神经网络的偏好感知社会化推荐方法
CN117540111B (zh) * 2024-01-09 2024-03-26 安徽农业大学 一种基于图神经网络的偏好感知社会化推荐方法
CN117556224B (zh) * 2024-01-12 2024-03-22 国网四川省电力公司电力科学研究院 一种电网设施抗震风险评估系统、方法及存储介质
CN117556224A (zh) * 2024-01-12 2024-02-13 国网四川省电力公司电力科学研究院 一种电网设施抗震风险评估系统、方法及存储介质
CN117992680A (zh) * 2024-04-02 2024-05-07 华侨大学 基于改进协同过滤及惩罚标签评分的多样化推荐方法
CN117992680B (zh) * 2024-04-02 2024-06-07 华侨大学 基于改进协同过滤及惩罚标签评分的多样化推荐方法
CN118013131A (zh) * 2024-04-10 2024-05-10 南昌大学 基于评分增强的图神经网络数据推荐方法、装置及介质
CN118013131B (zh) * 2024-04-10 2024-06-11 南昌大学 基于评分增强的图神经网络数据推荐方法、装置及介质

Similar Documents

Publication Publication Date Title
WO2022222037A1 (zh) 一种基于图神经网络推理的可解释推荐方法
Yang et al. Aspect-based sentiment analysis with alternating coattention networks
Wang et al. On the robustness of chatgpt: An adversarial and out-of-distribution perspective
Wu et al. Ai-generated content (aigc): A survey
Schwarzenberg et al. Layerwise relevance visualization in convolutional text graph classifiers
CN113505244B (zh) 基于深度学习的知识图谱构建方法、系统、设备及介质
CN110457480B (zh) 基于交互式注意力机制的细粒度情感分类模型的构建方法
JP2023539532A (ja) テキスト分類モデルのトレーニング方法、テキスト分類方法、装置、機器、記憶媒体及びコンピュータプログラム
CN108363695B (zh) 一种基于双向依赖语法树表征的用户评论属性抽取方法
JP7149993B2 (ja) 感情分析モデルの事前トレーニング方法、装置及び電子機器
Zhang et al. An emotional classification method of Chinese short comment text based on ELECTRA
US11995405B2 (en) Multi-lingual model training method, apparatus, electronic device and readable storage medium
CN111931054B (zh) 一种基于改进残差结构的序列推荐方法和系统
Hui et al. Few-shot relation classification by context attention-based prototypical networks with BERT
WO2022052744A1 (zh) 会话信息处理方法、装置、计算机可读存储介质及设备
CN111598118A (zh) 一种视觉问答任务实现方法及系统
Yenduri et al. Heuristic-assisted bert for twitter sentiment analysis
Hassija et al. Unleashing the potential of conversational AI: Amplifying Chat-GPT’s capabilities and tackling technical hurdles
CN113158045B (zh) 一种基于图神经网络推理的可解释推荐方法
Wei et al. Sentiment classification of tourism reviews based on visual and textual multifeature fusion
CN112740200A (zh) 用于基于共指消解的端到端深度强化学习的系统和方法
US20240046128A1 (en) Dynamic causal discovery in imitation learning
Nan et al. Speaker-oriented latent structures for dialogue-based relation extraction
Cao et al. Dual‐Channel Reasoning Model for Complex Question Answering
US20230229859A1 (en) Zero-shot entity linking based on symbolic information

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21937282

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21937282

Country of ref document: EP

Kind code of ref document: A1