CN113158045A - Interpretable recommendation method based on graph neural network reasoning - Google Patents
Interpretable recommendation method based on graph neural network reasoning Download PDFInfo
- Publication number
- CN113158045A CN113158045A CN202110426669.7A CN202110426669A CN113158045A CN 113158045 A CN113158045 A CN 113158045A CN 202110426669 A CN202110426669 A CN 202110426669A CN 113158045 A CN113158045 A CN 113158045A
- Authority
- CN
- China
- Prior art keywords
- user
- representation
- article
- user behavior
- graph
- Prior art date
- Legal status (The legal status 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 status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 53
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 31
- 230000006399 behavior Effects 0.000 claims abstract description 61
- 238000010586 diagram Methods 0.000 claims abstract description 19
- 230000003993 interaction Effects 0.000 claims abstract description 15
- 239000011159 matrix material Substances 0.000 claims abstract description 14
- 230000004927 fusion Effects 0.000 claims abstract description 9
- 230000006870 function Effects 0.000 claims description 18
- 230000015654 memory Effects 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 7
- 230000001537 neural effect Effects 0.000 claims description 7
- 238000001914 filtration Methods 0.000 claims description 6
- 230000004913 activation Effects 0.000 claims description 5
- 230000000306 recurrent effect Effects 0.000 claims description 5
- 230000014509 gene expression Effects 0.000 claims description 2
- 230000000644 propagated effect Effects 0.000 claims description 2
- 238000012552 review Methods 0.000 claims description 2
- 230000007246 mechanism Effects 0.000 description 9
- 238000012545 processing Methods 0.000 description 8
- 241000282414 Homo sapiens Species 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 238000011160 research Methods 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 2
- 239000002585 base Substances 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000000835 fiber Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000008447 perception Effects 0.000 description 2
- 230000001902 propagating effect Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- RYGMFSIKBFXOCR-UHFFFAOYSA-N Copper Chemical compound [Cu] RYGMFSIKBFXOCR-UHFFFAOYSA-N 0.000 description 1
- -1 DBPedia Substances 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 229910052802 copper Inorganic materials 0.000 description 1
- 239000010949 copper Substances 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 239000012458 free base Substances 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses an interpretable recommendation method based on graph neural network reasoning. The method comprises the following steps: constructing a multi-relation user behavior diagram aiming at an interaction matrix of user behaviors; for the user behavior graph, learning a high-order incidence relation by using a user behavior preference understanding model and transmitting user behavior preference to obtain vector representation of user, article and user-article correlation; inputting the user state representation, the article state representation and the user-article association representation output by the user behavior preference understanding model into the user behavior preference understanding model to obtain article recommendation of a given user; and taking the fusion of the user preference state representation, the item state representation and the user-item association representation output by the user behavior preference understanding model as the input of the interpretation generation model, and combining the text comment set to obtain the relevant interpretation of the recommended item for the given user. The present invention is capable of generating high quality, easy to understand recommendations while providing high performance recommendations.
Description
Technical Field
The invention relates to the technical field of computers, in particular to an interpretable recommendation method based on graph neural network reasoning.
Background
The recommendation system is used as an important tool for relieving the information overload problem, is widely applied to the fields of electronic commerce, news recommendation, video recommendation and the like, and gradually changes the life style of people. For example, an online bookstore recommends books that the user likes based on the user's historical behavior information. If the user can be made aware of why the book is recommended by a colloquial and understandable explanation, the effectiveness of the recommendation (helping the user make a quick decision) and the persuasiveness of the recommendation (improving the possibility of the user purchasing the book) can be greatly improved, which is the interpretability of the recommendation system. The reason and explanation of recommendation can be presented when the recommendation system recommends the articles to the user, so that the transparency of the recommendation system can be improved, the trust and acceptance of the user on the recommendation system can be improved, and the satisfaction of the user on recommended products and services can be improved. The interpretable recommendation system has important research significance and application value in fields with important interpretability and transparency, such as electronic commerce, finance, medical treatment, law and the like.
In the prior art, the interpretable recommendation method mainly includes two types: interpretable recommendation methods based on interpretation generation and interpretable recommendation methods based on knowledge-graphs. In recent years, with the development of natural language processing technology, many interpretable recommendation methods based on interpretation generation have been proposed, which can generate text recommendation interpretations in a form similar to human real comments. For example, one research result proposes a convolutional neural network model combined with a double attention mechanism, learns user preference representation and item feature representation from user/item comments, and adopts a local and global attention mechanism to select a word with information content to assist in scoring prediction, so as to enhance the model interpretability. As another example, one study outcome presented a neural scoring regression model to predict user-item scores while generating text interpretations based on an encoder-decoder framework using a Recurrent neural network such as Long Short-Term Memory network (LSTM), Gate Recurrent Unit (GRU). Another study performed scoring predictions using matrix decomposition while proposing a sequence-to-sequence interpretation generation model based on generation of the countermeasure network.
For the interpretable recommendation method based on the knowledge graph, the knowledge graph is introduced into a recommendation system as auxiliary information and receives more and more attention, and the knowledge graph can introduce various types of associated information. Researchers use the reasoning path of the knowledge graph to enhance the transparency and the interpretability of the model, and provide some interpretable recommendation methods based on the knowledge graph. An end-to-end knowledge graph perception recommendation method RippleNet is provided in a research, a propagation path is automatically found from the associated path of the knowledge graph, and the information propagation mechanism is used for finding the hierarchical potential interest of a user. And the other research result provides a recurrent neural network of the knowledge perception path, generates path representation by combining knowledge graph entities and the representation of the relationship, utilizes the serialized dependency relationship of the path to infer user preference, and applies the inference path to provide recommendation explanation.
However, although the existing interpretable recommendation scheme has a certain effect by utilizing text information extraction to generate recommendation interpretations or utilizing interpretability of a knowledge graph reasoning path enhanced model, certain defects exist in the aspects of user behavior analysis modeling and interpretation quality. In addition, the interpretable recommendation method based on interpretation generation directly adopts a generation model to generate the recommendation interpretation, and although the text interpretation which is easy to understand by human can be generated in the interpretation form, the generated interpretation has low quality and poor readability due to the problem of data sparsity.
In addition, the interpretable recommendation method based on the knowledge graph utilizes the fact that a common knowledge base such as DBPedia, FreeBase and the like is introduced into the common knowledge base and is associated to construct a knowledge graph, and the recommendation performance can be improved through information propagation and reasoning based on a graph model, but redundant noise data is introduced, unrelated entities are difficult to process, and recommendation efficiency and results are influenced. In addition, the method based on the knowledge graph needs to manually preset paths, rules and the like, and lacks domain knowledge, so that the recommendation and interpretation result is homogenized.
Disclosure of Invention
The present invention aims to overcome the defects of the prior art and provide an interpretable recommendation method based on graph neural network reasoning.
The technical scheme of the invention is to provide an interpretable recommendation method based on graph neural network reasoning. The method comprises the following steps:
aiming at an interaction matrix of user behaviors, constructing a multi-relationship user behavior graph, wherein a node set of the graph comprises all users and all articles, and the edges are user-article interaction relationships;
for the user behavior diagram, learning a high-order incidence relation by using a user behavior preference understanding model and transmitting user behavior preference to obtain vector representation of user, article and user-article correlation;
inputting the user state representation, the article state representation and the user-article association representation output by the user behavior preference understanding model into the user behavior preference understanding model to obtain article recommendation of a given user;
and taking the fusion of the user preference state representation, the item state representation and the user-item association representation output by the user behavior preference understanding model as the input of an interpretation generation model, and combining a text comment set to obtain the relevant interpretation of the item recommended for the given user.
Compared with the prior art, the invention has the advantages that the interpretable recommendation method based on graph neural network reasoning is provided for the interpretable recommendation problem, the user behavior mode is deeply understood by knowledge reasoning, and the potential intention of the user behavior is mined, so that the user behavior is comprehensively analyzed and modeled in a fine-grained manner. Moreover, the invention gives consideration to the recommendation accuracy and interpretability, generates the recommendation explanation with high quality and easy understanding for the user while providing personalized recommendation for the user, generates the recommendation explanation (recommendation reason) while providing the recommendation result, improves the transparency of the recommendation system and the trust and acceptance of the user on the recommendation system, and improves the satisfaction of the user on the recommended products and services.
Other features of the present invention and advantages thereof will become apparent from the following detailed description of exemplary embodiments thereof, which proceeds with reference to the accompanying drawings.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description, serve to explain the principles of the invention.
FIG. 1 is a flow diagram of an interpretable recommendation method based on graph neural network inference, according to one embodiment of the invention;
FIG. 2 is a block diagram of an interpretable recommendation method based on graph neural network inference, according to one embodiment of the invention.
Detailed Description
Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that: the relative arrangement of the components and steps, the numerical expressions and numerical values set forth in these embodiments do not limit the scope of the present invention unless specifically stated otherwise.
The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the invention, its application, or uses.
Techniques, methods, and apparatus known to those of ordinary skill in the relevant art may not be discussed in detail but are intended to be part of the specification where appropriate.
In all examples shown and discussed herein, any particular value should be construed as merely illustrative, and not limiting. Thus, other examples of the exemplary embodiments may have different values.
It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus, once an item is defined in one figure, further discussion thereof is not required in subsequent figures.
In the embodiment described below, let m users in a data set, denoted as u; n items, denoted v. The interaction behavior of the user with the article is expressed as<ui，vj，rij，dij>Wherein r isijRepresenting user uiFor article vjScore of (d)ijRepresenting user uiFor article vjThe user-item interaction matrix is represented as R, the text comment set is represented as D, and the dictionary formed by all the text comments is represented as Vg。
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 comments D, for a given user U, a top-N number of items (items that the user has not interacted) are recommended, while an interpretation Y of the recommended items is generated.
In short, the interpretable recommendation method based on graph neural network reasoning provided by the invention realizes comprehensive user behavior modeling by constructing the user behavior graph and carrying out multi-hop information propagation by utilizing the graph neural network reasoning. Generally, the method comprises four parts, namely user behavior graph construction, user behavior preference understanding based on graph neural network inference, user behavior interaction prediction and interpretation generation model.
Specifically, as shown in fig. 1 and fig. 2, the interpretable recommendation method based on graph neural network reasoning provided includes the following steps.
And step S110, constructing a user behavior diagram for the interaction matrix of the user behavior.
And aiming at the user behavior interaction matrix R, constructing a multi-relation user behavior graph G, wherein node sets (nodes) of the graph G are all users and all articles U & ltU & gt V & lt/U & gt. The edges are user-item interactions. For graph G, a graph-embedded representation of nodes and edges is initialized with random variables. Graph-embedded representation of user nodes is denoted as euGraph-embedded representation of item nodes is denoted as evUser-item interaction edge representation is denoted as ruv. And the user behavior graph is constructed, so that the follow-up user behavior understanding is facilitated.
And step S120, reasoning out the user behavior preference understanding by using the graph neural network.
In one embodiment, a user behavior preference understanding model based on a graph attention network is provided, not only high-order association information can be learned, but also user behavior preferences can be understood comprehensively and deeply by utilizing a multi-hop (multi-hop) propagation mechanism, and complete vector representation of users, articles and user-article associations can be obtained. The method specifically comprises the following steps:
step S121, the user behavior diagram pays attention to information propagation of the network.
An information propagation mechanism is defined, namely information propagation between a node and its neighbor nodes (directly connected nodes). For node eiIts first-order neighbor node set is
According to the first-order relevance of the graph structure, the first-order neighbor nodes
Propagate to node eiThe information of (2) is defined as:
wherein,
measuring the influence degree between a pair of relation nodes, wherein delta is a nonlinear activation function, such as a ReLU function or a tanh function, Wt，Wr，We，WnTo learn the parameters, b is a bias parameter.
While taking into account the associated node ei、ejAnd influence of the relationship between them rijAnd gamma is the normalized influence metric value. So that the neighbor node goes to the current node eiThe propagated information is a weighted combination of the representations of the neighboring nodes
And step S122, fusing the information of the user behavior diagram attention network.
After obtaining the information propagation of the (first order) neighboring nodes, the embedded representation e of the fusion node itselfiAnd propagation information from neighboring nodes
Is composed of
Wherein
The fusion function used may be a fusion mode such as nonlinear weighted summation of a graph convolution network.
Step S123, the user behavior diagram attention network is updated.
In one embodiment, the graph neural network models the higher-order connectivity information of the graph structure using multi-hop (multi-hop) inference in a recursive manner. And thus subsequently updated with the fused representation to a new representation of the node. For example, at the kth update, the node
Is expressed as:
through the above process, the representation of the nodes and edges, i.e. the user state representation e, is finally outputuArticle status representation evUser-item association representation ruv。
In this step S120, in order to overcome the shortcomings of the existing method in user behavior modeling, it is proposed to construct a user behavior preference understanding model based on graph neural network inference to obtain a high-quality user behavior preference representation.
And S130, predicting the user behavior interaction by utilizing a neural network collaborative filtering framework.
In one embodiment, based on a neural network collaborative filtering framework, a user representation, an article representation and a user-article association representation are fused, and a neural collaboration layer is adopted to realize recommendation prediction.
For example, 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, with the lowest layer being widest and fewer neurons per next layer. Each in the modelThe output of one layer will serve as the input to the next layer. Specifically, the input of the neural network co-layer of the first layer is a user state representation euArticle status representation evUser-item association representation ruvSplicing of, i.e. [ e ]u，ev，ruv]. At each neural network co-layer, a hidden vector is generated, represented as:
wherein WlAnd blThe weight matrix and the bias matrix for the l-th layer,
for the activation function used, a modified linear unit ReLU may be used. At the last layer, the vector h is hiddenlIs mapped to a prediction score
Expressed as:
wherein W is a weight matrix and sigma is a sigmoid function. For user u, sorting according to prediction score
And recommending top-N items, namely the first N recommended items, wherein N is a preset integer.
And step S140, obtaining the visual text explanation aiming at the recommended article by utilizing the explanation generation model.
In one embodiment, an interpretation generator based on an encoder-decoder framework is designed, a copy mechanism is introduced to extract relevant information from user/item source text comments, and a high-quality and easily-understood recommendation interpretation is generated by combining a generation mode and a copy mode.
In particular, a recurrent neural network GRU is applied as a solutionAn explanation generator is introduced, a copy mechanism is introduced to extract information from original comments of a user, and two modes (a generation mode and a copy mode) are combined to generate a visual text explanation Y ═ Y1，y2，..yT](word sequence) that is easy for the user to read and understand.
For the generation mode, GRU is used as a decoder, wherein the initial state of the decoder
Representing e for user preference statusuArticle status representation evUser-item association representation ruvFusion of (1), user uiRecommending items vjInterpreter y generated at the timeijtIs wtHas a probability of Pgen(yijt＝wt) Expressed as:
wherein,
for hidden states of GRU at time t, wt is a vector representation of a word, wt coming from the entire lexicon Vg。
For the copy mode, the probability of the word appearing in the original comment is calculated, and the user u is presented with the probabilityiRecommending items vjThe copy source of the interpretive word is uiV andjof item review documents
In copy mode, the interpreter yijtIs wtIs represented as Pcopy(yijt＝wt)：
Preferably, the generation model is constructed by combining the generation model and the copy model, and the probability of generating the target word is P (y)ijtWt), referring to equation (6), the user u is generatediRecommending items vjRecommended explanation of (Y)ij＝[yij1，yij2，..yijT]And T is an interpretation length.
P(yijt＝wt)＝Pgen(yijt＝wt)+Pcopy(yijt＝wt) (6)
By designing an interpretation generation model combining a generation mode and a copy mechanism, compared with the conventional interpretable recommendation method based on interpretation generation, the method can obtain high-quality recommendation interpretation which is easy to understand by a user, and solves the problems of low interpretation quality and poor readability at present.
It should be noted that the method provided by the invention is an end-to-end method, so that the feature representation learning, recommendation prediction and interpretation generation model is trained and optimized in a joint learning mode, the end-to-end interpretable recommendation frame reduces the training difficulty, accelerates the training process, reduces the memory requirement,
to verify the effectiveness and advancement of the present invention, a number of experiments were conducted on both Amazon-Electronics and Amazon-Music instruments data sets using the proposed interpretable method based on graph neural network reasoning. The experimental result shows that the method is always superior to the best method of interpretability recommendation, shows great superiority and has very wide application prospect.
Compared with the prior art, the invention has at least the following advantages:
(1) aiming at the interpretable recommendation task, the provided interpretable recommendation method based on graph neural network reasoning generates a recommendation explanation with high quality and easy understanding for human beings while providing a high-performance recommendation result, and realizes a recommendation framework which gives consideration to both recommendation effectiveness and interpretability.
(2) A deep-level user behavior understanding model is provided based on graph neural network reasoning, high-level associated information in user behaviors is utilized to propagate and execute the deep-level user behavior reasoning, comprehensive and deep-level user behavior modeling and understanding are achieved, high-quality representations are obtained in three aspects of users, articles and user-articles, and recommendation performance is effectively improved.
(3) In the interpretation and generation model, a copy mechanism is introduced, and the interpretation and generation model combining the generation mode and the copy mode is provided, so that the generation of high-quality recommended interpretation easy to understand by human is facilitated.
In conclusion, the invention innovatively provides a user behavior understanding model based on graph neural network reasoning, which realizes comprehensive and fine-grained user behavior modeling by constructing a user behavior graph, carrying out multi-hop information propagation based on graph neural network reasoning, learning deep-level and high-quality user, article and user-article associated representation, and solving the problem of the deficiency of the conventional user behavior analysis modeling.
It is to be understood that appropriate changes or modifications may be made by those skilled in the art to the above-described embodiments without departing from the spirit and scope of the present invention, for example, using LSTM instead of GRU networks, or using other activation functions for non-linear processing, etc.
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 embodied therewith for causing a processor to implement various aspects of the present invention.
The computer readable storage medium may be a tangible device that can hold and store the instructions for use by the instruction execution device. The computer readable storage medium may be, for example, but not limited to, an electronic memory device, a magnetic memory device, an optical memory device, an electromagnetic memory device, a semiconductor memory device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of the computer readable storage medium would include the following: a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a Static Random Access Memory (SRAM), a portable compact disc read-only memory (CD-ROM), a Digital Versatile Disc (DVD), a memory stick, a floppy disk, a mechanical coding device, such as punch cards or in-groove projection structures having instructions stored thereon, and any suitable combination of the foregoing. Computer-readable storage media as used herein is not to be construed as transitory signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through a waveguide or other transmission medium (e.g., optical pulses through a fiber optic cable), or electrical signals transmitted through electrical wires.
The computer-readable program instructions described herein may be downloaded from a computer-readable storage medium to a respective computing/processing device, or to an external computer or external storage device via 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. The network adapter card or network interface in each computing/processing device receives computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in a computer-readable storage medium in the respective computing/processing device.
The computer program instructions for carrying out operations of the present invention may be assembler instructions, Instruction Set Architecture (ISA) instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including an object oriented programming language such as Smalltalk, C + +, Python, or the like, and conventional procedural programming languages, such as the "C" programming 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. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider). In some embodiments, aspects of the present invention are implemented by personalizing an electronic circuit, such as a programmable logic circuit, a Field Programmable Gate Array (FPGA), or a Programmable Logic Array (PLA), with state information of computer-readable program instructions, which can execute the computer-readable program instructions.
Aspects of the present invention are described herein with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer-readable program instructions.
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, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions/acts specified in the flowchart and/or block diagram block or blocks. These computer-readable program instructions may also be stored in a computer-readable storage medium that can direct a computer, programmable data processing apparatus, and/or other devices to function in a particular manner, such that the computer-readable medium storing the instructions comprises an article of manufacture including instructions which implement the function/act specified in the flowchart and/or block diagram block or blocks.
The computer readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other devices to cause a series of operational steps to be performed on the computer, other programmable apparatus or other devices to produce a computer implemented process such that the instructions which execute on the computer, other programmable apparatus or other devices implement the functions/acts specified in the flowchart and/or block diagram block or blocks.
The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of instructions, which comprises one or more executable instructions for implementing the specified logical function(s). In some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown 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. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions. It is well known to those skilled in the art that implementation by hardware, by software, and by a combination of software and hardware are equivalent.
Having described embodiments of the present invention, the foregoing description is intended to be exemplary, not exhaustive, and not limited to the embodiments disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen in order to best explain the principles of the embodiments, the practical application, or improvements made to the technology in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein. The scope of the invention is defined by the appended claims.
Claims (10)
1. An interpretable recommendation method based on graph neural network reasoning comprises the following steps:
step S1: aiming at an interaction matrix of user behaviors, constructing a multi-relationship user behavior graph, wherein a node set of the graph comprises all users and all articles, and the edges are user-article interaction relationships;
step S2: for the user behavior diagram, learning a high-order incidence relation by using a user behavior preference understanding model and transmitting user behavior preference to obtain vector representation of user, article and user-article correlation;
step S3: inputting the user state representation, the article state representation and the user-article association representation output by the user behavior preference understanding model into the user behavior preference understanding model to obtain article recommendation of a given user;
step S4: and taking the fusion of the user preference state representation, the item state representation and the user-item association representation output by the user behavior preference understanding model as the input of an interpretation generation model, and combining a text comment set to obtain the relevant interpretation of the item recommended for the given user.
2. The method according to claim 1, wherein step S2 includes:
for the user behavior graph G, according to the first-order relevance of the graph structure, the first-order neighbor nodes
Propagate to node eiThe information of (2) is defined as:
wherein,
is a pair of relational nodes ejAnd eiDegree of influence between, delta being a non-linear activation function, Wt，Wr，We，WnIs the correlation term weight, b is the offset, rijExpressing the influence of the incidence relation between the nodes, wherein gamma is the influence metric value after regularization, and the neighbor node points to the current node eiThe propagated information is a weighted combination of the representations of the neighboring nodes
Node eiIs a first-order neighbor node set
After information propagation of the first-order adjacent nodes is obtained, the embedded expression e of the fusion node is selfiAnd propagation information from neighboring nodes
Is shown as
Wherein
As a fusion function used;
modeling the high-order connection information of the graph structure by using multi-hop inference in a recursive mode, and finally outputting the representation of nodes and edges, wherein the representation is marked as a user state euArticle state evUser-item association representation ruv。
3. The method according to claim 1, wherein step S3 includes:
the user behavior preference understanding model adopts a neural collaborative filtering model with a tower-shaped structure, the input of a first layer of neural network collaborative layer is the splicing of user state representation, article state representation and user-article association representation, and for each neural network collaborative layer, the generation of a hidden vector is shown as follows:
wherein, WlAnd blThe weight matrix and the bias matrix for the l-th layer,
is an activation function, hl-1Is the vector of the l-1 layer output;
at the last layer of the neural collaborative filtering model, the vector h is hiddenlPredicted scores mapped to recommended items
Scoring based on prediction for a given user
The items are sorted and recommended.
4. The method of claim 1, wherein in step S4, the given user u is generated by extracting information from the user' S original comments in conjunction with the copy modeiRecommending items vjRecommended explanation of (Y)ij＝[yij1，yij2，..yijT]Expressed as:
P(yijt＝wt)＝Pgen(yijt＝wt)+Pcopy(yijt＝wt)
wherein T is the interpretation length, Pgen(yijt＝wt) Is a user u obtained by using a generation patterniRecommending items vjInterpreter y generated at the timeijtIs wtProbability of (P)copy(yijt＝wt) Is a user u obtained by a copy mechanismiRecommendation materialArticle vjInterpreter y generated at the timeijtIs wtThe probability of (c).
5. Method according to claim 4, characterized in that for the generation mode a recurrent neural network GRU is used as decoder, the initial state of which is the initial state
Representing e for user preference statusuArticle status representation evUser-item association representation ruvFusion of (1), user uiRecommending items vjInterpreter y generated at the timeijtIs wtThe probability of (d) is expressed as:
wherein,
is a hidden state of GRU at time t, wtIs a vector representation of a word, wtFrom a dictionary Vg。
6. The method of claim 5, wherein in copy mode, user uiRecommending items vjInterpreter y generated at the timeijtIs wtThe probability of (d) is expressed as:
wherein, VcIs uiV andjof item review documents。
7. The method of claim 3, wherein the hidden vector h is hidden at the last layer of the neural collaborative filtering modellPredicted scores mapped to recommended items
Expressed as:
wherein W is a weight matrix, sigma is a sigmoid function, and hlIs the concealment vector and l is the layer index.
8. The method of claim 2, wherein the fusion function
A non-linear weighted sum.
9. A computer-readable storage medium, on which a computer program is stored, which program, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 8.
10. A computer device comprising a memory and a processor, on which memory a computer program is stored which is executable on the processor, characterized in that the steps of the method of any of claims 1 to 8 are implemented when the processor executes the program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110426669.7A CN113158045B (en) | 2021-04-20 | 2021-04-20 | Interpretable recommendation method based on graph neural network reasoning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110426669.7A CN113158045B (en) | 2021-04-20 | 2021-04-20 | Interpretable recommendation method based on graph neural network reasoning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113158045A true CN113158045A (en) | 2021-07-23 |
| CN113158045B CN113158045B (en) | 2022-11-01 |
Family
ID=76867723
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110426669.7A Active CN113158045B (en) | 2021-04-20 | 2021-04-20 | Interpretable recommendation method based on graph neural network reasoning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113158045B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113987261A (en) * | 2021-11-08 | 2022-01-28 | 烟台大学 | Video recommendation method and system based on dynamic trust perception |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109754317A (en) * | 2019-01-10 | 2019-05-14 | 山东大学 | Merge interpretation clothes recommended method, system, equipment and the medium of comment |
| CN109933721A (en) * | 2019-02-01 | 2019-06-25 | 中森云链(成都)科技有限责任公司 | A kind of interpretable recommended method merging user concealed article preference and implicit trust |
| CN110060132A (en) * | 2019-04-24 | 2019-07-26 | 吉林大学 | Interpretable Method of Commodity Recommendation based on fine-grained data |
| WO2019183191A1 (en) * | 2018-03-22 | 2019-09-26 | Michael Bronstein | Method of news evaluation in social media networks |
| CN112084407A (en) * | 2020-09-08 | 2020-12-15 | 辽宁工程技术大学 | Collaborative filtering recommendation method fusing graph neural network and attention mechanism |
| CN112650929A (en) * | 2020-12-31 | 2021-04-13 | 安徽农业大学 | Graph neural network recommendation method integrating comment information |
-
2021
- 2021-04-20 CN CN202110426669.7A patent/CN113158045B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019183191A1 (en) * | 2018-03-22 | 2019-09-26 | Michael Bronstein | Method of news evaluation in social media networks |
| CN109754317A (en) * | 2019-01-10 | 2019-05-14 | 山东大学 | Merge interpretation clothes recommended method, system, equipment and the medium of comment |
| CN109933721A (en) * | 2019-02-01 | 2019-06-25 | 中森云链(成都)科技有限责任公司 | A kind of interpretable recommended method merging user concealed article preference and implicit trust |
| CN110060132A (en) * | 2019-04-24 | 2019-07-26 | 吉林大学 | Interpretable Method of Commodity Recommendation based on fine-grained data |
| CN112084407A (en) * | 2020-09-08 | 2020-12-15 | 辽宁工程技术大学 | Collaborative filtering recommendation method fusing graph neural network and attention mechanism |
| CN112650929A (en) * | 2020-12-31 | 2021-04-13 | 安徽农业大学 | Graph neural network recommendation method integrating comment information |
Non-Patent Citations (1)
| Title |
|---|
| 王子冬: ""基于评论文本的推荐算法研究"", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113987261A (en) * | 2021-11-08 | 2022-01-28 | 烟台大学 | Video recommendation method and system based on dynamic trust perception |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113158045B (en) | 2022-11-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2022222037A1 (en) | Interpretable recommendation method based on graph neural network inference | |
| Wang et al. | On the robustness of chatgpt: An adversarial and out-of-distribution perspective | |
| Hadi et al. | Large language models: a comprehensive survey of its applications, challenges, limitations, and future prospects | |
| Seo et al. | Comparative study of deep learning-based sentiment classification | |
| Ling et al. | Domain specialization as the key to make large language models disruptive: A comprehensive survey | |
| Liu et al. | Detection of spam reviews through a hierarchical attention architecture with N-gram CNN and Bi-LSTM | |
| CN110457480B (en) | Construction method of fine granularity emotion classification model based on interactive attention mechanism | |
| CN108363695B (en) | User comment attribute extraction method based on bidirectional dependency syntax tree representation | |
| JP2023539532A (en) | Text classification model training method, text classification method, device, equipment, storage medium and computer program | |
| EP4411567A1 (en) | Item recommendation method and apparatus, and storage medium | |
| Ouyang et al. | Dialogue graph modeling for conversational machine reading | |
| CN111598118B (en) | Visual question-answering task implementation method and system | |
| CN111931054B (en) | Sequence recommendation method and system based on improved residual error structure | |
| CN111914185A (en) | Graph attention network-based text emotion analysis method in social network | |
| Mishra et al. | Please be polite: Towards building a politeness adaptive dialogue system for goal-oriented conversations | |
| US20220171941A1 (en) | Multi-lingual model training method, apparatus, electronic device and readable storage medium | |
| Varshney et al. | Commonsense and named entity aware knowledge grounded dialogue generation | |
| Hassija et al. | Unleashing the potential of conversational AI: Amplifying Chat-GPT’s capabilities and tackling technical hurdles | |
| Dongbo et al. | Intelligent chatbot interaction system capable for sentimental analysis using hybrid machine learning algorithms | |
| CN113158045B (en) | Interpretable recommendation method based on graph neural network reasoning | |
| CN111967941A (en) | Method for constructing sequence recommendation model and sequence recommendation method | |
| Zhu et al. | Self-directed machine learning | |
| Xu et al. | Netrl: Task-aware network denoising via deep reinforcement learning | |
| CN113761910A (en) | Comment text fine-grained emotion analysis method integrating emotional characteristics | |
| Gou et al. | Topic model for personalized end-to-end task-oriented dialogue |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |