CN115269512A

CN115269512A - Object recommendation method, device and storage medium for realizing IA by combining RPA and AI

Info

Publication number: CN115269512A
Application number: CN202210888405.8A
Authority: CN
Inventors: 于孟萱
Original assignee: Beijing Laiye Network Technology Co Ltd; Laiye Technology Beijing Co Ltd
Current assignee: Beijing Laiye Network Technology Co Ltd; Laiye Technology Beijing Co Ltd
Priority date: 2022-07-26
Filing date: 2022-07-26
Publication date: 2022-11-01
Also published as: WO2024022354A1

Abstract

The present disclosure relates to the field of computer technologies, and in particular, to an object recommendation method, an object recommendation device, and a storage medium for implementing IA by combining RPA and AI. The object recommendation method comprises the following steps: acquiring a bid document and a bid document set through a Robot Process Automation (RPA) system; acquiring first knowledge representation information corresponding to bidding documents on a first knowledge map by adopting a target knowledge representation model, and acquiring second knowledge representation information corresponding to at least one bidding document in a bidding document set, wherein the first knowledge map corresponds to the bidding documents and the bidding document set; respectively acquiring a matching degree set between the first knowledge representation information and the at least one second knowledge representation information by adopting a matching network model, and determining a target bidding document in the at least one bidding document based on the matching degree set; and recommending the target bidding object corresponding to the target bidding document. By the method and the device, accuracy of object recommendation can be improved.

Description

Object recommendation method, device and storage medium for realizing IA by combining RPA and AI

Technical Field

The present application relates to the field of computer technologies, and in particular, to an object recommendation method, an object recommendation device, and a storage medium for implementing IA by combining RPA and AI.

Background

Robot Process Automation (RPA) is a Process task that simulates human operations on a computer through specific robot software and automatically executes according to rules.

Artificial Intelligence (AI) is a technical science that studies and develops theories, methods, techniques and application systems for simulating, extending and expanding human Intelligence.

Intelligent Automation (IA) is a generic name for a series of technologies from robot Process Automation To artificial intelligence, and combines RPA with a variety of AI technologies such as Optical Character Recognition (OCR), intelligent Character Recognition (ICR), process Mining (Process Mining), deep Learning (Deep Learning, DL), machine Learning (Machine Learning, ML), natural Language Processing (NLP), speech Recognition (Automatic Speech Recognition, ASR), speech synthesis (Text Speech, TTS), computer Vision (Computer Vision, CV), to create a thought, learning, and adaptive end-To-end Process flow, covering from discovery, process coverage, to data collection through Automatic and continuous data collection, understanding data, and optimizing the meaning of the whole Process flow using data management and whole Process flow.

With the development of scientific technology, terminal technology matures day by day, and convenience of production and life of users is improved. In a terminal application scene, a terminal can recommend user preference information to a user according to historical user preference information or interaction information between commodities and user commodities. However, in the bidding application scenario, only depending on the historical user preference information or the interaction information may result in a problem that the accuracy of object recommendation is not high due to lack of consideration on the background knowledge of bidding.

Disclosure of Invention

The embodiment of the application provides an object recommendation method, device and storage medium for realizing IA by combining RPA and AI, so as to solve the problems in the related art, and the technical scheme is as follows:

in a first aspect, an embodiment of the present application provides an object recommendation method combining an RPA and an AI, including:

acquiring a bid document and a bid document set through a Robot Process Automation (RPA) system;

acquiring first knowledge representation information corresponding to the bid-inviting document and second knowledge representation information corresponding to at least one bid document in the bid document set on a first knowledge map by adopting a target knowledge representation model, wherein the first knowledge map corresponds to the bid-inviting document and the bid document set;

respectively acquiring a matching degree set between the first knowledge representation information and the at least one second knowledge representation information by adopting a matching network model, and determining a target bidding document in the at least one bidding document based on the matching degree set;

and recommending the target bidding object corresponding to the target bidding document.

Optionally, before the obtaining, by using the target knowledge representation model, the first knowledge representation information corresponding to the bid-on document on the first knowledge graph, the method further includes:

acquiring bid inviting documents and bid document sets in a target document format through a Robot Process Automation (RPA) system;

acquiring a bid document and a triple information set corresponding to the bid document set by adopting a triple extraction technology;

and establishing a first knowledge graph corresponding to the bidding document and the bidding document set based on the triple information set.

Optionally, the obtaining, by using a triple extraction technique, a triple information set corresponding to the bid document and the bid document set includes:

acquiring a first triple information set corresponding to a bidding document and a second triple information set corresponding to a bidding document set by adopting a triple extraction technology;

performing word segmentation processing on the first triple information set and the second triple set by adopting a crust word segmentation model to obtain a processed first triple information set and a processed second triple set;

and performing entity boundary character string auditing treatment on the processed first triple information set and the processed second triple information set to obtain triple information sets corresponding to the bidding documents and the bidding document sets.

Optionally, before the obtaining of the first knowledge representation information corresponding to the bid document and the obtaining of the second knowledge representation information corresponding to at least one bid document in the bid document by using the target knowledge representation model, the method further includes:

acquiring a bidding training document set;

acquiring a knowledge graph corresponding to a bidding training document set;

and based on a distributed representation learning reasoning technology, carrying out vector space mapping on at least one knowledge pair in the knowledge map, and establishing a target knowledge representation model.

obtaining model performance parameters corresponding to each initial knowledge representation model in the initial knowledge representation model set;

and determining a target knowledge representation model in the initial knowledge representation model set based on the model performance parameters, wherein the target knowledge representation model is a distributed knowledge inference model.

Optionally, before the matching network model is adopted to obtain the matching degree set of the first knowledge representation information and the at least one second knowledge representation information, the method further includes:

acquiring bidding attribute information corresponding to the bidding document and bidding attribute information corresponding to any bidding document in the bidding document set;

adjusting the first knowledge representation information by adopting a user representation learning module and bidding attribute information to obtain adjusted first knowledge representation information;

adjusting second knowledge representation information corresponding to any bidding document by adopting a user representation learning module and the bidding attribute information to obtain adjusted second knowledge representation information;

and traversing the second knowledge representation information set to obtain the adjusted second knowledge representation information set.

Optionally, determining a target bid document of the at least one bid document based on the set of matching degrees includes:

acquiring the highest matching degree in the matching degree set;

and obtaining the bidding document corresponding to the highest matching degree from at least one bidding document, and determining the bidding document as the target bidding document.

In a second aspect, an embodiment of the present application provides an object recommendation device combining an RPA and an AI, including:

the system comprises a document acquisition unit, a bidding document collection unit and a bidding document collection unit, wherein the document acquisition unit is used for acquiring a bidding document and the bidding document collection through a Robot Process Automation (RPA) system;

the information acquisition unit is used for acquiring first knowledge representation information corresponding to the bid-inviting document on a first knowledge map by adopting a target knowledge representation model, and acquiring second knowledge representation information corresponding to at least one bid document in a bid document set, wherein the first knowledge map corresponds to the bid-inviting document and the bid document set;

the document determining unit is used for respectively acquiring a matching degree set between the first knowledge representation information and the at least one second knowledge representation information by adopting a matching network model, and determining a target bidding document in the at least one bidding document based on the matching degree set;

and the object recommending unit is used for recommending the target bidding object corresponding to the target bidding document.

Optionally, the apparatus further includes a target document obtaining unit, a set extracting unit, and a map establishing unit, configured to, before obtaining, on the first knowledge map, first knowledge representation information corresponding to the bid document by using the target knowledge representation model:

the system comprises a target document acquisition unit, a bidding document acquisition unit and a bidding document collection unit, wherein the target document acquisition unit is used for acquiring a bidding document and a bidding document collection in a target document format through a Robot Process Automation (RPA) system;

the system comprises a set extraction unit, a bid document acquisition unit and a bid document extraction unit, wherein the set extraction unit is used for acquiring a bid document and a triple information set corresponding to the bid document set by adopting a triple extraction technology;

and the map establishing unit is used for establishing a first knowledge map corresponding to the bid document and the bid document set based on the triple information set.

Optionally, the set acquiring unit includes a set acquiring subunit, a word segmentation processing subunit, and an auditing processing subunit, and the set acquiring unit is configured to, when acquiring the triple information sets corresponding to the bid documents and the bid document sets by using a triple extraction technique:

the system comprises a set acquisition subunit, a bid document collection unit and a bid document collection unit, wherein the set acquisition subunit is used for acquiring a first triple information set corresponding to a bid document and a second triple information set corresponding to a bid document set by adopting a triple extraction technology;

the word segmentation processing subunit is configured to perform word segmentation processing on the first triple information set and the second triple set by using a crust word segmentation model to obtain a processed first triple information set and a processed second triple set;

and the auditing processing subunit is used for auditing the entity boundary character strings of the processed first triple information set and the processed second triple information set to obtain triple information sets corresponding to the bidding documents and the bidding document sets.

Optionally, the apparatus further includes a set obtaining unit, a map obtaining unit, and a model establishing unit, configured to, before obtaining first knowledge representation information corresponding to the bid document and second knowledge representation information corresponding to at least one bid document in the bid documents by using the target knowledge representation model:

the system comprises a set acquisition unit, a bid-attracting training document set acquisition unit and a bid-attracting training document acquisition unit, wherein the set acquisition unit is used for acquiring the bid-attracting training document set;

the map acquisition unit is used for acquiring a knowledge map corresponding to the bidding training document set;

and the model establishing unit is used for carrying out vector space mapping on at least one knowledge pair in the knowledge map based on a distributed representation learning inference technology to establish a target knowledge representation model.

Optionally, the apparatus further includes a parameter obtaining unit and a model determining unit, configured to, before obtaining first knowledge representation information corresponding to the bid document and obtaining second knowledge representation information corresponding to at least one bid document in the bid document by using the target knowledge representation model:

the parameter acquisition unit is used for acquiring model performance parameters corresponding to each initial knowledge representation model in the initial knowledge representation model set;

and the model determining unit is used for determining a target knowledge representation model in the initial knowledge representation model set based on the model performance parameters, and the target knowledge representation model is a distributed knowledge inference model.

Optionally, the apparatus further includes an attribute information obtaining unit, an information adjusting unit, and a set traversing unit, configured to, before obtaining the matching degree set of the first knowledge representation information and the at least one second knowledge representation information by using the matching network model:

the attribute information acquisition unit is used for acquiring bid attribute information corresponding to the bid document and bid attribute information corresponding to any bid document in the bid document set;

the information adjusting unit is used for adjusting the first knowledge representation information by adopting the user representation learning module and the bidding attribute information to obtain the adjusted first knowledge representation information;

the information adjusting unit is also used for adjusting second knowledge representation information corresponding to any bidding document by adopting the user representation learning module and the bidding attribute information to obtain the adjusted second knowledge representation information;

and the set traversing unit is used for traversing the second knowledge representation information set to obtain the adjusted second knowledge representation information set.

Optionally, the document determining unit includes a matching degree obtaining sub-unit and a document matching sub-unit, and the document determining unit is configured to, when determining a target bid document in the at least one bid document based on the matching degree set:

the matching degree obtaining subunit is used for obtaining the highest matching degree in the matching degree set;

and the document matching subunit is used for acquiring the bidding document corresponding to the highest matching degree from the at least one bidding document and determining the bidding document as the target bidding document.

In a third aspect, an embodiment of the present application provides a terminal combining an RPA and an AI, where the terminal includes: a memory and a processor. Wherein the memory and the processor are in communication with each other via an internal connection path, the memory is configured to store instructions, the processor is configured to execute the instructions stored by the memory, and the processor is configured to perform the method of any of the above aspects when the processor executes the instructions stored by the memory.

In a fourth aspect, embodiments of the present application provide a computer-readable storage medium, which stores a computer program, and when the computer program runs on a computer, the method in any one of the above-mentioned aspects is executed.

The advantages or beneficial effects in the above technical solution at least include:

in one or related embodiments, a bid document and a set of bid documents are obtained by a robotic process automation RPA system; acquiring first knowledge representation information corresponding to the bid-inviting document and second knowledge representation information corresponding to at least one bid document in the bid document set on a first knowledge map by adopting a target knowledge representation model, wherein the first knowledge map corresponds to the bid-inviting document and the bid document set; respectively acquiring a matching degree set between the first knowledge representation information and the at least one second knowledge representation information by adopting a matching network model, and determining a target bidding document in the at least one bidding document based on the matching degree set; and recommending the target bidding object corresponding to the target bidding document. Therefore, the target knowledge representation model is adopted to obtain the knowledge representation information corresponding to the bid document and the bid document on the first knowledge map, and recommendation is performed based on the knowledge representation information corresponding to the bid document and the bid document.

The foregoing summary is provided for the purpose of description only and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments, and features described above, further aspects, embodiments, and features of the present application will be readily apparent by reference to the drawings and following detailed description.

Drawings

In the drawings, like reference numerals refer to the same or similar parts or elements throughout the several views unless otherwise specified. The figures are not necessarily to scale. It is appreciated that these drawings depict only some embodiments in accordance with the disclosure and are not to be considered limiting of its scope.

Fig. 1 is a schematic diagram illustrating a background of an object recommendation method for implementing IA in combination with RPA and AI according to an embodiment of the present application;

fig. 2 is a schematic diagram illustrating a background architecture of an object recommendation method for implementing IA in combination with RPA and AI according to an embodiment of the present application;

FIG. 3 is a flowchart illustrating an object recommendation method for implementing IA in conjunction with RPA and AI according to an embodiment of the present application;

FIG. 4 is a flowchart illustrating an object recommendation method for implementing IA in conjunction with RPA and AI according to an embodiment of the present application;

FIG. 5 illustrates a flow diagram of a format conversion according to an embodiment of the present application;

FIG. 6 illustrates a first knowledge-graph representation of an embodiment of the present application;

FIG. 7 is a flowchart illustrating an object recommendation method for implementing IA in conjunction with RPA and AI according to an embodiment of the present application;

FIG. 8 is a flowchart illustrating an object recommendation method for implementing IA in conjunction with RPA and AI according to an embodiment of the present application;

fig. 9 is a schematic structural diagram of an object recommendation apparatus for implementing IA in combination with RPA and AI according to an embodiment of the present application;

fig. 10 is a schematic structural diagram of an object recommendation device for implementing IA in combination with RPA and AI according to an embodiment of the present application;

fig. 11 is a schematic structural diagram of an object recommendation apparatus for implementing IA in combination with RPA and AI according to an embodiment of the present application;

fig. 12 is a schematic structural diagram of an object recommendation device for implementing IA in combination with RPA and AI according to an embodiment of the present application;

fig. 13 is a schematic structural diagram of an object recommendation apparatus for implementing IA in combination with RPA and AI according to an embodiment of the present application;

fig. 14 is a schematic structural diagram of an object recommendation apparatus for implementing IA in combination with RPA and AI according to an embodiment of the present application;

fig. 15 is a schematic structural diagram of an object recommendation apparatus for implementing IA in combination with RPA and AI according to an embodiment of the present application;

fig. 16 shows a block diagram of a terminal according to an embodiment of the present application.

Detailed Description

Reference will now be made in detail to embodiments of the present application, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the accompanying drawings are illustrative and are only for the purpose of explaining the present application and are not to be construed as limiting the present application.

In the description of the present application, the term "plurality" means two or more.

In the description of this application, the term "RPA" refers to the automatic execution of flow tasks according to rules by a specific "robot software", simulating the operation of a human on a computer. The RPA system comprises at least three components: development tools, operation tools and a control center. The Intelligent automation platform can realize seamless integration of multiple capabilities such as RPA (Intelligent Document Processing, IDP), conversational AI (Conversational AI, coAI), process Mining (Process Mining), has five major functions of 'business understanding', 'Process creation', 'anywhere operation', 'centralized control' and 'man-machine cooperation', realizes end-to-end Intelligent automation of business processes for enterprises, replaces manual operation, further improves business efficiency and accelerates digital transformation.

In the intelligent automation platform, three components of the RPA system are named as a process creator, a process robot and a robot commander respectively. The process creator is a programming tool for process development, and performs specific steps of interface automation operation, AI identification, data reading and writing and the like in the process. The process creator allows easy assembly of an automated process meeting business requirements by adopting a flow chart and low code mode and adopting mouse dragging steps.

And after the RPA process is written, the RPA is deployed in a process robot. The operation may be initiated manually as desired, or automatically when certain trigger conditions are met. The tasks can be arranged and the process can be traced back.

The robot commander is a platform for uniformly managing a plurality of process robots in an enterprise, can rapidly issue tasks in batches, and provides data, certificates, files and the like required by operation for the process robots. The running state of the process robot can be monitored in real time or the history of the process robot can be reviewed.

Artificial Intelligence (AI) capabilities specially designed for RPA are also provided in the intelligent automation platform, and these AI capabilities also form a fourth component of the intelligent automation platform, which is called an intelligent document processing platform. The intelligent document Processing platform is a Processing platform created based on deep learning algorithms such as Optical Character Recognition (OCR), natural Language Processing (NLP) and the like, provides functions such as document identification, classification, element extraction, verification, comparison, error correction and the like, and realizes automation of daily document Processing work of enterprises.

In the description of the present application, the term "bidding document" refers to an offer invitation document, which is a main basis of a project bidding activity, in which a bidder issues and informs potential bidders of information on a project requirement, rules of the bidding activity, contract conditions, and the like.

In the description of the present application, the term "bid document" refers to a responsive document that a bidder should submit to a bid document for compilation.

In the description of the present application, the term "bid object" refers to an object that should be bid upon documentation requiring preparation of a bid document.

In the description of the present application, the term "knowledge representation model" refers to a model for representing semantic information of images, text, speech, etc. as a low-dimensional dense entity vector. Types of knowledge representation models include, but are not limited to, distance models (SE), single Layer Neural Network models (SLM), energy models (SME), transfer Distance models (relational Distance models), bilinear function based models (Bilinear function models), tensor Neural Network models (NTN), matrix decomposition models, translation models, and the like.

In the description of the present application, the term "knowledge representation information" refers to low-dimensional dense entity vector information corresponding to semantic information acquired using a knowledge representation model.

In the description of the present application, the term "knowledge graph (knowledgegraph)" refers to a semantic network with entities, concepts as nodes and semantic relationships as edges. The knowledge map is a series of different graphs for displaying the relation between knowledge development process and structure, and is used to describe knowledge resource and its carrier, mine, analyze, constitute and display knowledge and their mutual relation.

In the description of the present application, the term "matching network model" refers to a model for mapping an object into an embedding space, which also encapsulates the label distribution, then projects the test object into the same embedding space using different architectures, and then measures the similarity using cosine similarity, achieving classification and detection effects.

In the description of the present application, the term "document format" refers to the particular encoding scheme used by a computer to store textual information. Document formats include, but are not limited to, text txt Format, HTML Format, word Format, portable Document Format (PDF) Format, and the like.

In the description of the present application, the term "triple extraction technique" refers to a technique for jointly extracting triple information of entity + relations, including multiple relation extraction between entities.

In the description of the present application, the term "triplet" is a set of references such as ((x, y), z) (that is to say, a triplet is an even whose first projection is also an even), often abbreviated as (x, y, z). The triple is a concept in a common basic course-data structure of computer profession, and is mainly a compression mode for storing a sparse matrix.

In the description of the present application, the term "triplet information" is information used to identify entities and relationships between entities. The triplet information does not refer to a fixed information. For example, when the triplet information corresponding to at least one data changes, the triplet information may also change correspondingly.

In the description of the present application, the term "entity" refers to something that generally refers to as a basis for all attributes and everything primitive that can exist independently. An entity is the most basic element in a knowledge-graph.

In the description of the present application, the term "ending segmentation model" refers to a model for performing segmentation processing on chinese text. The Chinese text word segmentation model can perform the functions of word segmentation, part-of-speech tagging, keyword extraction and the like on Chinese texts, and supports a user-defined dictionary.

In the description of the present application, the term "word segmentation process" refers to a process of recombining successive word sequences into a word sequence according to a certain specification.

In the description of the present application, the term "entity boundary string auditing process" refers to a process for auditing whether triplet information is information of an instruction class.

In the description of the present application, the term "model performance parameters" refers to configuration variables inside the model. The model performance parameters may be estimated using data or learned using data. Model performance parameters are needed for model prediction. The model performance parameter values may define the model function. The model performance parameters are typically saved as part of the learning model. Model performance parameters can be estimated using an optimization algorithm, which is an efficient search for possible values of the parameters.

In the description of the present application, the term "distributed representation learning inference technique" refers to a technique of representing an entity vector in a low-dimensional dense vector space, and then performing calculation and inference.

With the development of scientific technology, terminal technology is mature day by day, and convenience of production and life of users is improved. In a terminal application scene, a user can perform object recommendation on data input by the user through an object recommendation application program.

According to some embodiments, fig. 1 is a schematic background diagram illustrating an object recommendation method for implementing IA by combining RPA and AI according to an embodiment of the present application. As shown in fig. 1, a user may click on an object recommendation application of a terminal, and when the terminal detects that the user clicks on the object recommendation application, the terminal may present an object recommendation interface. The user may enter a bid document based on the object recommendation interface. Then, when the terminal detects that the user clicks an object recommendation key, the terminal may perform object recommendation for the bid document input by the user.

Fig. 2 is a schematic diagram of a background architecture of an object recommendation method combining RPA and AI according to an embodiment of the present application. As shown in fig. 2, the terminal 11 may upload a bid-on document issued by a user to the server 13 through the network 12. When the server 13 receives the bid document, the server 13 may determine a bid document matching the bid document based on the historical user preference information and transmit the bid document matching the bid document to the terminal 11 through the network 12, and when the terminal receives the bid document transmitted from the server 13, the terminal may display the bid document on a display interface.

In some embodiments, in the bidding scene, the interaction information between the past winning enterprises and items, the attribute information of the enterprises and items, the background knowledge of the enterprises and items, including the enterprise qualification, the enterprise investment amount and the like, need to be understood, so as to recommend the bidding information improved by the user. However, in the related art, when the terminal performs user recommendation, only historical user preference information and interaction information are used, so that the accuracy of the terminal performing user recommendation is not high.

It is readily understood that the terminal includes, but is not limited to: wearable devices, handheld devices, personal computers, tablet computers, in-vehicle devices, smart phones, computing devices or other processing devices connected to a wireless modem, and the like. The terminal devices in different networks may be called different names, for example: user equipment, access terminal, subscriber unit, subscriber station, mobile station, remote terminal, mobile device, user terminal, wireless communication device, user agent or user equipment, cellular telephone, cordless telephone, personal Digital Assistant (PDA), fifth generation mobile communication technology (5G) network, or a terminal device in a future evolution network, and the like. The terminal can be installed with an operating system, which is an operating system capable of running in the terminal, is a program for managing and controlling terminal hardware and terminal applications, and is an indispensable system application in the terminal. The operating system includes, but is not limited to, android, IOS, windows Phone (WP), and Ubuntu mobile operating system.

These and other aspects of embodiments of the present application will be apparent from and elucidated with reference to the following description and drawings. In the description and drawings, particular embodiments of the application have been specifically applied to illustrate some ways of implementing the principles of the embodiments of the application, but it will be understood that the scope of the embodiments of the application is not limited thereto. On the contrary, the embodiments of the application include all changes, modifications and equivalents coming within the spirit and terms of the claims appended hereto.

An object recommendation method for implementing IA in conjunction with RPA and AI according to an embodiment of the present application is described below with reference to the drawings.

Fig. 3 is a flowchart illustrating an object recommendation method for implementing IA in conjunction with RPA and AI according to an embodiment of the present application, and as shown in fig. 3, the method may include the following steps:

step S101: acquiring a bid document and a bid document set through a Robot Process Automation (RPA) system;

according to some embodiments, a set of bid documents refers to a collective of at least one bid document. The set of bid documents is not specific to a particular set of bid documents. The set of bid documents may change accordingly, for example, when the bid document content changes. For example, when the number of bid documents included in the set of bid documents changes, the set of bid documents may also change accordingly.

It is easy to understand that when the terminal performs object recommendation, the terminal can acquire bid documents and bid document sets through the robot process automation RPA system.

Step S102: acquiring first knowledge representation information corresponding to bidding documents on a first knowledge map by adopting a target knowledge representation model, and acquiring second knowledge representation information corresponding to at least one bidding document in a bidding document set;

according to some embodiments, the target knowledge representation model refers to a trained model for obtaining knowledge representation information on the first knowledge-graph. The object knowledge representation model does not refer specifically to a fixed model. For example, the target knowledge representation model may change when the first knowledge-graph changes. When the terminal acquires a model modification instruction for the target knowledge representation model, the target knowledge representation model may also be changed.

In some embodiments, the first knowledge-graph refers to a knowledge-graph corresponding to a bid document and a set of bid documents. The first knowledge-graph does not refer to a fixed graph. For example, the first knowledge-graph may change when a bidding document changes. The first knowledge-graph may also change when the set of bid documents changes.

In some embodiments, the first knowledge representation information refers to corresponding knowledge representation information of the bid document. The first knowledge representation information does not refer to a certain fixed information. For example, when a bid document changes, the first knowledge representation information may change. When the first knowledge-graph changes, the first knowledge representation information may also change.

In some embodiments, the second knowledge representation information refers to knowledge representation information corresponding to the bid document. The second knowledge representation information does not refer to a certain fixed information. For example, the second knowledge representation information may change when a bid document changes. The second knowledge representation information may also change when the first knowledge-graph changes.

It is easy to understand that when the terminal acquires the bid document and the bid document set, the terminal may acquire the first knowledge representation information corresponding to the bid document on the first knowledge map by using the target knowledge representation model. In addition, the terminal can also acquire second knowledge representation information corresponding to at least one bid document in the bid documents.

Step S103: respectively acquiring a matching degree set between the first knowledge representation information and the at least one second knowledge representation information by adopting a matching network model, and determining a target bidding document in the at least one bidding document based on the matching degree set;

according to some embodiments, the degree of matching refers to a degree of matching between the first knowledge representation information and the second knowledge representation information. The matching degree is not specific to a certain fixed matching degree. For example, the degree of matching may change when the first knowledge representation information changes. When the second knowledge representation information changes, the matching degree may also change.

In some embodiments, the set of degrees of matching refers to a set of at least one degree of matching between the first knowledge representation information and the at least one second knowledge representation information. The matching degree set does not refer to a fixed set. For example, the set of degrees of matching may change when the amount of second knowledge representation information changes. When the degree of matching changes, the set of degrees of matching may also change.

In some embodiments, the target bid document refers to a bid document that matches the bid document obtained in the set of bid documents based on the set of degrees of matching. The target bid document does not refer specifically to a fixed document. For example, the target bid document can change when the set of bid documents changes. When a bid document changes, the target bid document may also change.

It is easy to understand that when the terminal acquires the first knowledge representation information and the at least one second knowledge representation information, the terminal may acquire a matching degree set between the first knowledge representation information and the at least one second knowledge representation information respectively by using a matching network model. Further, the terminal may determine a target bid document of the at least one bid document based on the set of matching degrees.

Step S104: and recommending the target bidding object corresponding to the target bidding document.

According to some embodiments, the target bid object refers to a bid object corresponding to the target bid document. The target bid object does not refer specifically to a fixed object. For example, when a target bid document changes, the target bid object may change.

It is easy to understand that when the terminal acquires the target bidding document, the terminal may recommend the target bidding object corresponding to the target bidding document.

In the embodiment of the application, a bidding document and a bidding document set are obtained through a Robot Process Automation (RPA) system; acquiring first knowledge representation information corresponding to bidding documents on a first knowledge map by adopting a target knowledge representation model, and acquiring second knowledge representation information corresponding to at least one bidding document in a bidding document set; respectively acquiring a matching degree set between the first knowledge representation information and the at least one second knowledge representation information by adopting a matching network model, and determining a target bidding document in the at least one bidding document based on the matching degree set; and recommending the target bidding object corresponding to the target bidding document. Therefore, the target knowledge representation model is adopted to obtain the knowledge representation information corresponding to the bid document and the bid document on the first knowledge map, and recommendation is performed based on the knowledge representation information corresponding to the bid document and the bid document.

Fig. 4 is a flowchart illustrating an object recommendation method for implementing IA in conjunction with RPA and AI according to an embodiment of the present application, and as shown in fig. 4, the method may include the following steps:

step S201: acquiring a bid document and a bid document set through a Robot Process Automation (RPA) system;

according to some embodiments, when the terminal acquires the bidding document and the bid document set based on the RPA, the terminal may grab the bidding document and the bid document set in the information extraction source and a document format corresponding to the bidding document and a document format corresponding to at least one bid document in the bid document set through the RPA system.

In some embodiments, the information extraction source refers to a source of the bid document and bid document set. The information extraction source does not refer to a fixed information. The information extraction sources include, but are not limited to, public resource trading websites, industry core bidding websites, regional bidding websites, and the like. The number of the information extraction sources is multiple, so that the full coverage of information data can be ensured, and the latest bid document and bid document set can be mastered at any time.

In some embodiments, when the terminal captures the bid document and the bid document set in the information extraction source through the RPA system, the terminal may capture a web page through a website address in the information extraction source. When the terminal captures the webpage, the terminal can extract the bid document and the bid document set from the webpage source code, and the document format corresponding to the bid document and the document format corresponding to at least one bid document in the bid document set.

For example, when the terminal captures a webpage of the industry core bidding website, the terminal may extract a bidding document in a word format corresponding to the industry core bidding website from the webpage source code of the industry core bidding website. When the terminal grabs the regional bidding website, the terminal can also extract the bidding document in the PDF format corresponding to the regional bidding website from the webpage source code of the regional bidding website.

Step S202: acquiring bid inviting documents and bid document sets in a target document format through a Robot Process Automation (RPA) system;

according to some embodiments, the target document format refers to a document format corresponding to the acquired bid document and a document format corresponding to at least one bid document in the set of bid documents selected by the RPA system. The target document format does not feature a fixed format. For example, when the document format selected by the RPA system changes, the target document format may change accordingly. When the terminal acquires the format modification instruction for the target document format, the target document format may also be changed.

In some embodiments, when the terminal acquires the bid document and the bid document set through the robot process automation RPA system, if the terminal determines that a document format corresponding to the bid document and a document format corresponding to at least one bid document in the bid document set are not in the target document format, the terminal may perform format conversion on the bid document and the bid document set through the RPA system, so as to obtain the bid document and the bid document set in the target document format.

In some embodiments, when the terminal performs format conversion on the bid document and the bid document set through the RPA system, the terminal may perform format conversion on the bid document using a document format conversion tool built in the RPA system. For example, when the terminal acquires a bid-up document in a word format, the terminal may call a word underlying macro language (VBA) using a python win32 library to convert the bid-up document in the word format into a bid-up document in a PDF format.

For example, if the target document format set by the terminal is PDF format, if the terminal captures the bid-calling document in the information extraction source through the RPA system, the bid-calling document a in word format is captured. The terminal may convert the format of the bid-adding document a into a bid-adding document in PDF format by using a document format conversion tool built in the RPA system, as shown in fig. 5.

It is easy to understand that, when the terminal obtains the bid document and the bid document set through the robot process automation RPA system, the terminal can obtain the bid document and the bid document set in the target document format through the robot process automation RPA system.

Step S203: acquiring a bid document and a triple information set corresponding to the bid document set by adopting a triple extraction technology;

according to some embodiments, a triplet information set refers to a set of at least one triplet information aggregate. The triplet information set does not refer specifically to a fixed set. For example, when a tender document changes, the set of triplet information may change. When the set of bid documents changes, the set of triplet information may also change.

For example, the triplet information included in the triplet information set includes, but is not limited to, item _ type, item _ bid _ announce _ date, item _ bid _ company, item _ funding source, item _ home _ location, item _ bid _ scope, item _ time limit, item _ build _ size, item _ investment amount, item _ business qualification, item _ personnel _ qualification, item _ business credit _ registration _ requirement, item _ performance _ qualification, bid _ project, bid _ company, item _ quote, bid _ project, bid _ company, item _ credit, item _ business _ performance, this _ bid _ score, bid _ company, whether item _ bid, item _ float _ rate, item _ award, and so on.

According to some embodiments, when the terminal acquires the triplet information sets corresponding to the bid documents and the bid document sets by using a triplet extraction technique, the triplet extraction technique adopted by the terminal includes, but is not limited to, using a multi-head selection technique, a multi-round question-answering technique, a block-by-block extraction technique, and the like.

In some embodiments, when the terminal acquires the triple information sets corresponding to the bid document and the bid document set by using a multi-head selection technology, the terminal may use a Conditional Random Field (CRF) to solve an entity identification task, and regard relationship extraction as a multi-head selection problem, thereby implementing a function of identifying multiple potential relationships of each entity.

In some embodiments, when the terminal acquires a triple information set corresponding to a bid document and a bid document set by using a multi-turn question-answering technology, the terminal may regard entity-relationship extraction as a multi-turn question-answering task, construct a question model for the entities and the relationships, and answer a question template by using text information (text to be extracted) to obtain the entity-relationships. On the question-answering task, a span (span) -based answer extraction mode can be adopted.

In some embodiments, when the terminal acquires the triple information sets corresponding to the bidding documents and the bidding document sets by using the module-based extraction technology, the terminal may divide the entity-relationship extraction into two modules. First, head-Entity extraction (HE) is extracted, and then Tail-Entity and Relation extraction (TER) is jointly extracted, so that Entity overlapping conditions can be reduced.

According to some embodiments, when the terminal acquires the triplet information sets corresponding to the bidding documents and the bidding document sets by using the triplet extraction technology, the terminal may acquire the first triplet information set corresponding to the bidding documents and the second triplet information set corresponding to the bidding document sets by using the triplet extraction technology. Furthermore, the terminal may perform word segmentation processing on the first triple information set and the second triple information set by using a crust word segmentation model to obtain a processed first triple information set and a processed second triple information set. Finally, the terminal can perform entity boundary character string auditing processing on the processed first triple information set and the processed second triple information set, so as to obtain triple information sets corresponding to the bidding documents and the bidding document sets. Therefore, the terminal can remove the messy characters in the triple information set, only retain the preset types of triple information, such as numbers, english, chinese characters and the like, and remove the repeated triple information in the triple information set. And further, the information quality of the triple information set can be improved, and the object recommendation accuracy can be improved.

In some embodiments, the first set of triplet information refers to a set of triplet information to which the bid document corresponds. The first triplet information set does not refer to a fixed set. For example, when a tender document changes, the first set of triplet information may change. When a triplet information changes, the first set of triplet information may also change.

In some embodiments, the second set of triplet information refers to the set of triplet information to which the set of bid documents corresponds. The second triplet set does not refer to a fixed set. For example, the second set of triplet information may change when the set of bid documents changes. When a triplet information changes, the second set of triplet information may also change.

In some embodiments, when the terminal performs a participle process on the first triple information set and the second triple information set by using a resultant participle model, the terminal may label the triple information in the first triple information set and the second triple information set by using a part-of-speech tagging tool pos _ seg. Furthermore, through the performance comparison of the ending word segmentation model and the part-of-speech tagging tool, useless noise entities in the first triple information set and the second triple information set can be removed.

In some embodiments, the terminal performs entity boundary string auditing on the processed first triple information set and the processed second triple information set, so that the terminal can reduce noise in the triple information set and improve the quality of the triple information set. For example, when the triplet information acquired by the terminal is "a network _ B province _ power _ limited company _ business _ state _ yes", if the terminal sets "a network _ B province _ power _ limited company" as the required triplet information and "business _ state _ yes" as the irrelevant information, the terminal may perform the physical boundary string auditing process to remove the irrelevant information.

In some embodiments, when the terminal acquires the triplet information set corresponding to the bidding document and the bidding document set by using a triplet extraction technique, the terminal may further remove the triplet information corresponding to the object that has low frequency or has not undergone bidding activity by removing the invalid features and the sparse features in the triplet information set.

It is easy to understand that when the terminal acquires the bid document and the bid document set in the target document format, the terminal may acquire the triplet information set corresponding to the bid document and the bid document set by using a triplet extraction technique.

Step S204: establishing a first knowledge graph corresponding to the bidding document and the bidding document set based on the triple information set;

according to some embodiments, when the terminal establishes the first knowledge graph corresponding to the bid document and the bid document set based on the triple information set, the terminal may acquire a triple information subset corresponding to a target entity in the triple information set by using a target deep learning model, and merge each triple information in the triple information subset, and the terminal may establish the first knowledge graph corresponding to the data set.

In some embodiments, the target deep learning model refers to a model in which a terminal acquires a triple information subset corresponding to a target entity. The target deep learning model does not refer to a fixed deep learning model. The target deep learning model may also change accordingly, for example, when the type of the deep learning model changes. When the model name of the deep learning model changes, the target deep learning model can also change correspondingly.

In some embodiments, the target entity refers to entity information required by the terminal to construct the first knowledge-graph. The target entity information does not refer to a fixed information. For example, when the triple information set changes, the target entity information may also change. When entity information changes, the target entity information may also change.

In some embodiments, the first knowledge-graph information refers to information corresponding to the first knowledge-graph. The first knowledge-graph information does not refer to a fixed information. For example, when the first knowledge-graph changes, the first knowledge-graph information may also change. When the targeted bid attribute information changes, the first knowledge-graph information may also change. When the target entity information changes, the first knowledge-graph information may also change.

For example, fig. 6 shows a first knowledge-graph representation of an embodiment of the present application. As shown in fig. 6, the "bid document E2" is a winning bid document of the "bid document E1", the responsibility of the "bid document E2" is "person G2", and the "bid document E2" is derived from "agency F2"; the "bidding document E1" is derived from the "institution F1", and the executive officer of the bidding document E1 "is the" person G1".

It is easy to understand that when the terminal acquires the triple information sets corresponding to the bid document and the bid document set, the terminal may establish the first knowledge graph corresponding to the bid document and the bid document set based on the triple information sets.

Step S205: obtaining model performance parameters corresponding to each initial knowledge representation model in the initial knowledge representation model set;

according to some embodiments, the initial knowledge representation model refers to an untrained knowledge representation model. The initial knowledge representation model does not feature some fixed model. For example, the initial knowledge representation model may change when its model performance parameters change. When the terminal acquires a model modification instruction for the initial knowledge representation model, the initial knowledge representation model may also be changed.

In some embodiments, the set of initial knowledge representation models refers to a set of at least one initial knowledge representation model aggregated. The set of initial knowledge representation models does not refer specifically to a fixed set. For example, the set of initial knowledge representation models may change when the initial knowledge representation models change. When the number of initial knowledge representation models changes, the set of initial knowledge representation models may also change.

It is easy to understand that, when the terminal acquires the target knowledge representation model, the terminal may acquire the model performance parameters corresponding to each initial knowledge representation model in the initial knowledge representation model set.

Step S206: determining a target knowledge representation model in the initial knowledge representation model set based on the model performance parameters;

according to some embodiments, when the terminal determines the target knowledge representation model in the initial knowledge representation model set based on the model performance parameters, the terminal evaluates the learning efficiency and learning effect of the initial knowledge representation model based on the model performance parameters corresponding to the initial knowledge representation model. Furthermore, the terminal can select the adopted target knowledge representation model according to the learning efficiency and the learning effect corresponding to at least one initial knowledge representation model in the initial knowledge representation model set.

For example, when the initial knowledge representation model is a transition distance model, the initial knowledge representation model may be a Translation embedding (transform for Modeling Multi-relational Data, transform) model, a Flexible Translation (TransF) model, a Learning Entity and relationship embedding (TransF) model, a hyper plane Translation (transform) model, a knowledge driven behavior Understanding (Human Activity Understanding, token) model. When the initial knowledge representation model is a bilinear model, the initial knowledge representation model may be a semantic matching model such as RESCAL, distMult, complEx, holE, simplE, analog, and the like. After the learning efficiency and the learning effect of the initial knowledge representation models are evaluated, the terminal can select to adopt a TransE model as a target knowledge representation model.

It is easy to understand that when the terminal acquires the model performance parameters corresponding to each initial knowledge representation model in the initial knowledge representation model set, the terminal may determine the target knowledge representation model in the initial knowledge representation model set based on the model performance parameters.

Step S207: acquiring first knowledge representation information corresponding to bidding documents on a first knowledge map by adopting a target knowledge representation model, and acquiring second knowledge representation information corresponding to at least one bidding document in the bidding documents;

according to some embodiments, when the terminal acquires knowledge representation information on the first knowledge graph by using the target knowledge representation model, the terminal may map symbolic representations in the triple information to a vector space and perform numerical representation based on a distributed representation learning inference method through the target knowledge representation model. Therefore, dimension disasters can be reduced, meanwhile, the implicit association between the entities and the relations in the three-tuple information can be captured, the calculation is direct, and the training speed is high.

For example, when the terminal acquires the triple information "national grid _ Jiangxi province _ Power", the terminal may represent the triple information as knowledge representation information (0, 1, 2) based on the distributed representation learning inference method.

It is easy to understand that when the terminal acquires the target knowledge representation model, the terminal may acquire the first knowledge representation information corresponding to the bid document on the first knowledge graph by using the target knowledge representation model. In addition, the terminal can also acquire second knowledge representation information corresponding to at least one bid document in the bid documents.

Step S208: respectively acquiring a matching degree set between the first knowledge representation information and the at least one second knowledge representation information by adopting a matching network model, and determining a target bidding document in the at least one bidding document based on the matching degree set;

according to some embodiments, when the terminal adopts the matching network model to respectively obtain the matching degree sets between the first knowledge representation information and the at least one second knowledge representation information, the terminal may take the first knowledge representation information as the fixed data of the matching network model and sequentially input the second knowledge representation information in the second knowledge representation information sets to the matching network model. Further, the terminal may acquire a set of matching degrees between the first knowledge representation information and the at least one second knowledge representation information, respectively.

In some embodiments, when the terminal respectively obtains the matching degree sets between the first knowledge representation information and the at least one second knowledge representation information by using the matching network model, the terminal may further refer to an attention mechanism to improve the accuracy of obtaining the matching degree sets.

In some embodiments, the attention mechanism refers to a special structure embedded in the machine learning model for automatically learning and calculating the contribution of the input data to the output data. The attention mechanism is not specifically limited to a fixed structure. For example, the attention mechanism may change when the terminal acquires a modification instruction for the attention mechanism.

Step S209: and recommending the target bidding object corresponding to the target bidding document.

According to some embodiments, when the terminal determines a target bid document among the at least one bid document based on the set of matching degrees, the terminal may determine at least two target bid documents. Furthermore, the terminal can display the at least two target bidding documents in the display page in a sequential order.

For example, the terminal may sort at least two matching degrees in the matching degree set according to the order of magnitude. Furthermore, the terminal may select the bidding document corresponding to the matching degree of the top 10 as the target bidding document, so that the terminal may display the 10 target bidding documents in the display page in order.

In some embodiments, when the terminal recommends the target bidding object corresponding to the target bidding document, the terminal may further provide interpretable reference for the target bidding object through the recommendation interpretation module, so that the recommendation result has a high interpretation degree while the object recommendation accuracy is improved.

In some embodiments, the recommendation interpretation module refers to a module for letting the user know why a target bid document is recommended to him through a colloquial and understandable interpretation. The recommendation interpretation module does not refer to a fixed module. For example, the recommendation interpretation module may change when the target bid document changes. When the terminal acquires a module modification instruction for the recommendation interpretation module, the recommendation interpretation module may also be changed.

In the embodiment of the application, the tendering document and the bidding document set in the target document format are acquired through the robot process automation RPA system, so that the matching of the RPA system to the tendering document and the bidding document set in different document formats can be improved, and the accuracy of object recommendation can be improved. And secondly, acquiring a triple information set corresponding to the bid document and the bid document set by adopting a triple extraction technology, and establishing a first knowledge graph corresponding to the bid document and the bid document set based on the triple information set, so that the accuracy of acquiring the first knowledge graph can be improved, and the accuracy of recommending a bid object can be improved. In addition, model performance parameters corresponding to each initial knowledge representation model in the initial knowledge representation model set are obtained, and the target knowledge representation model is determined in the initial knowledge representation model set based on the model performance parameters, so that the accuracy of object recommendation can be improved by selecting the required target knowledge representation model from the initial knowledge representation model set.

Fig. 7 is a flowchart illustrating an object recommendation method for implementing IA in conjunction with RPA and AI according to an embodiment of the present application, and as shown in fig. 7, the method may include the following steps:

step S301: acquiring a bid document and a bid document set through a Robot Process Automation (RPA) system;

the specific process is as described above, and is not described herein again.

Step S302: acquiring a bidding training document set;

according to some embodiments, a bid training document refers to a document used to train a knowledge representation model such that the knowledge representation model has an expression of a priori knowledge. The bid-inviting training document does not specifically refer to a fixed document. For example, the bid training document may change when the knowledge representation model changes. When a bid document and a set of bid documents change, the bid training document may also change.

In some embodiments, a set of bid training documents refers to a set of at least one bid training document aggregated. The set of bidding training documents does not refer specifically to a fixed set. For example, when a bid training document changes, the set of bid training documents may change. When the knowledge representation model changes, the set of bid training documents may also change.

It is easy to understand that when the terminal acquires the target knowledge representation model, the terminal can acquire the bid training document set.

Step S303: acquiring a knowledge graph corresponding to a bidding training document set;

according to some embodiments, when the terminal acquires the bid training document set, the terminal may acquire bid attribute information and entity information corresponding to the bid training document set. Furthermore, the terminal can acquire the attribute information of bidding and tendering and knowledge map information corresponding to the entity information through a robot process automation RPA system and an optical character recognition OCR model. Finally, the terminal can display the knowledge graph corresponding to the knowledge graph information.

In some embodiments, the bidding attribute information refers to attribute information corresponding to the set of bidding training documents. The bidding attribute information includes, but is not limited to, bidding information, industry information, regional information, and the like. For example, the bidding attribute information corresponding to the "bidding document M" in the bidding training document set is bidding information. The bidding attribute information corresponding to the "bidding document N" is bidding information.

In some embodiments, the entity information refers to information of an entity to which the set of bid training documents corresponds. Entity information includes, but is not limited to, object attribute information, data attribute information, relationship attribute information, and the like. The object attribute information refers to an entity name or an attribute name for abstracting and describing a homogeneous entity. Object attribute information includes, but is not limited to, people, institutions, time, posts, accounts, and the like. The data attribute information refers to attribute information corresponding to the entity itself. For example, attribute information corresponding to entity "person D" includes, but is not limited to, name D1, past name D2, age D3, and the like. The relationship attribute information refers to mutual relationship information between entities.

In some embodiments, the knowledge-graph information refers to information used in constructing a knowledge-graph. The knowledge-graph information does not refer to a fixed information. For example, the knowledge-graph information may change when entity information changes. When bid attribute information changes, the knowledge-graph information may also change.

It is easy to understand that when the terminal acquires the bid training document set, the terminal may acquire a knowledge graph corresponding to the bid training document set.

Step S304: based on a distributed representation learning reasoning technology, carrying out vector space mapping on at least one knowledge pair in a knowledge map, and establishing a target knowledge representation model;

according to some embodiments, knowledge refers to entity relationship information contained in a knowledge graph. The knowledge does not refer to a fixed knowledge. For example, when a knowledge-graph changes, the knowledge may change. This knowledge may change when the set of bidding training documents changes.

In some embodiments, when the terminal performs vector space mapping on at least one knowledge pair in the knowledge graph, the terminal may acquire a mapping function corresponding to an entity relationship in the knowledge graph based on a distributed representation learning inference technique. Further, the terminal may map the entity relationship to a vector space and represent the entity relationship in a numerical form. For example, knowledge of "winning bid" may be mapped to 25 and knowledge of "bidding" may be mapped to index value 68.

It is easy to understand that when the terminal acquires the knowledge graph corresponding to the bid-soliciting training document set, the terminal may perform vector space mapping on at least one knowledge pair in the knowledge graph based on a distributed representation learning inference technique. Further, the terminal may build a target knowledge representation model.

Step S305: acquiring first knowledge representation information corresponding to bidding documents on a first knowledge map by adopting a target knowledge representation model, and acquiring second knowledge representation information corresponding to at least one bidding document in the bidding documents;

the specific process is as described above, and is not described herein again.

Step S306: respectively acquiring a matching degree set between the first knowledge representation information and at least one second knowledge representation information by adopting a matching network model;

the specific process is as described above, and is not described herein again.

Step S307: acquiring the highest matching degree in the matching degree set;

for example, when the terminal acquires that the degree of matching between the first knowledge representation information and the second knowledge representation information B1 is 30, the degree of matching between the first knowledge representation information and the second knowledge representation information B2 is 50, and the degree of matching between the first knowledge representation information and the second knowledge representation information B3 is 70, the terminal may acquire that the degree of matching between the first knowledge representation information and the second knowledge representation information B3 is the highest degree of matching in the set of degrees of matching.

It is easy to understand that when the terminal acquires the set of matching degrees between the first knowledge representation information and the at least one second knowledge representation information, the terminal may acquire the highest matching degree in the set of matching degrees.

Step S308: obtaining a bidding document corresponding to the highest matching degree from at least one bidding document, and determining the bidding document as a target bidding document;

for example, when the terminal acquires that the matching degree between the first knowledge representation information and the second knowledge representation information B3 is the highest matching degree in the matching degree set, the terminal may acquire the bid document B3 corresponding to the highest matching degree from at least one bid document. Further, the terminal may determine the bid document b3 as a target bid document.

It is easy to understand that when the terminal obtains the highest matching degree in the matching degree set, the terminal may obtain, from the at least one bid document, a bid document corresponding to the highest matching degree, and determine the bid document as the target bid document.

Step S309: and recommending the target bidding object corresponding to the target bidding document.

The specific process is as described above, and is not described herein again.

In the embodiment of the application, the knowledge graph corresponding to the bidding training document set is obtained by obtaining the bidding training document set, and based on the distributed representation learning inference technology, vector space mapping is performed on at least one knowledge pair in the knowledge graph to establish the target knowledge representation model, so that the accuracy of establishing the target knowledge representation model can be improved, and the accuracy of user recommendation can be improved. And secondly, acquiring the highest matching degree in the matching degree set, acquiring a bid document corresponding to the highest matching degree in at least one bid document, and determining the bid document as a target bid document, so that the accuracy of determining the target bid document can be improved by determining the bid document with the highest matching degree as the target bid document.

Fig. 8 is a flowchart illustrating an object recommendation method for implementing IA in conjunction with RPA and AI according to an embodiment of the present application, and as shown in fig. 8, the method may include the following steps:

step S401: acquiring a bid document and a bid document set through a Robot Process Automation (RPA) system;

the specific process is as described above, and is not described herein again.

Step S402: acquiring first knowledge representation information corresponding to bidding documents on a first knowledge map by adopting a target knowledge representation model, and acquiring second knowledge representation information corresponding to at least one bidding document in the bidding documents;

the specific process is as described above, and is not described herein again.

Step S403: acquiring bidding attribute information corresponding to the bidding document and bidding attribute information corresponding to any bidding document in the bidding document set;

according to some embodiments, the bidding attribute information refers to attribute information corresponding to the bidding document. The bidding attribute information does not refer to a fixed information. The bidding attribute information includes, but is not limited to, bidding information, industry information, regional information, and the like.

In some embodiments, the bid attribute information refers to attribute information corresponding to the bid document. The bid attribute information includes, but is not limited to, bid information, industry information, regional information, and the like.

It is easy to understand that when the terminal acquires the first knowledge representation information corresponding to the bid document and acquires the second knowledge representation information corresponding to at least one bid document in the bid document, the terminal may acquire the bid attribute information corresponding to the bid document and the bid attribute information corresponding to any bid document in the bid document set.

Step S404: adjusting the first knowledge representation information by adopting a user representation learning module and bidding attribute information to obtain adjusted first knowledge representation information;

according to some embodiments, a user representation learning module refers to a module for automatically learning valid features from raw input data and converting the input information into a valid feature representation. The user-indicated learning module is not specific to a fixed module. The representation form adopted by the user representation learning module includes but is not limited to local representation, distribution representation and the like.

In some embodiments, the local representation is referred to as a one-hot vector, and this discrete representation has good explanatory property, but the dimension of the one-hot vector is high and cannot be expanded, and the orthogonality among the vectors cannot be calculated.

In some embodiments, the distribution representation can be generally represented as a low-dimensional continuous dense vector, the representation capability is much stronger, and the similarity is easy to calculate.

In some embodiments, when the terminal acquires the bidding attribute information corresponding to the bidding document, the terminal may adopt the user representation learning module to perform fusion multiplication on the feature representation corresponding to the bidding attribute information and the first knowledge representation information. Further, the terminal may acquire the adjusted first knowledge representation information.

It is easy to understand that when the terminal acquires the bidding attribute information corresponding to the bidding document, the terminal may adjust the first knowledge representation information by using the user representation learning module and the bidding attribute information to obtain the adjusted first knowledge representation information.

Step S405: adjusting second knowledge representation information corresponding to any bidding document by adopting a user representation learning module and the bidding attribute information to obtain adjusted second knowledge representation information;

according to some embodiments, when the terminal acquires the bid attribute information corresponding to the bid document, the terminal may adopt the user representation learning module to perform fusion multiplication on the feature representation corresponding to the bid attribute information and the second knowledge representation information. Further, the terminal may acquire the adjusted second knowledge representation information.

It is easy to understand that when the terminal acquires the bid attribute information corresponding to any bid document in the bid document set, the terminal may adjust the second knowledge representation information corresponding to any bid document by using the user representation learning module and the bid attribute information to obtain the adjusted second knowledge representation information.

Step S406: traversing the second knowledge representation information set to obtain an adjusted second knowledge representation information set;

it is easy to understand that when the terminal acquires the bid attribute information corresponding to any bid document in the bid document set, the terminal may traverse the second knowledge representation information set to obtain the adjusted second knowledge representation information set.

Step S407: respectively acquiring a matching degree set between the adjusted first knowledge representation information and the adjusted at least one second knowledge representation information by adopting a matching network model, and determining a target bidding document in the at least one bidding document based on the matching degree set;

the specific process is as described above, and is not described herein again.

Step S408: and recommending the target bidding object corresponding to the target bidding document.

The specific process is as described above, and is not described herein again.

In the embodiment of the application, the bidding attribute information corresponding to the bidding document and the bidding attribute information corresponding to any bidding document in the bidding document set are obtained, the user representation learning module and the bidding attribute information are adopted to adjust the first knowledge representation information to obtain the adjusted first knowledge representation information, and the second knowledge representation information set is traversed to obtain the adjusted second knowledge representation information set; therefore, the accuracy of object recommendation can be improved by adopting a multi-model fusion learning method to perform representation learning and fine adjustment on the knowledge representation information. Secondly, a matching network model is adopted to respectively obtain a matching degree set between the adjusted first knowledge representation information and the adjusted at least one second knowledge representation information, a target bidding document in the at least one bidding document is determined based on the matching degree set, and a target bidding object corresponding to the target bidding document is recommended.

The following are embodiments of the apparatus of the present application that may be used to perform embodiments of the method of the present application. For details which are not disclosed in the embodiments of the apparatus of the present application, reference is made to the embodiments of the method of the present application.

Please refer to fig. 9, which is a schematic structural diagram of an object recommendation device for implementing IA by combining RPA and AI according to an embodiment of the present application. The object recommendation device that implements IA in combination with RPA and AI may be implemented as all or part of a device by software, hardware, or a combination of both. The object recommendation apparatus 9000 that realizes IA by combining RPA and AI includes a document acquisition unit 9001, an information acquisition unit 9002, a document determination unit 9003, and an object recommendation unit 9004, in which:

a document acquiring unit 9001, configured to acquire a bid document and a bid document set by a robot process automation RPA system;

an information obtaining unit 9002, configured to obtain first knowledge representation information corresponding to a bid document on a first knowledge map by using a target knowledge representation model, and obtain second knowledge representation information corresponding to at least one bid document in a bid document set, where the first knowledge map corresponds to the bid document and the bid document set;

a document determining unit 9003, configured to obtain, by using a matching network model, matching degree sets between the first knowledge representation information and the at least one second knowledge representation information, respectively, and determine a target bid document of the at least one bid document based on the matching degree sets;

and an object recommending unit 9004 for recommending the target bidding object corresponding to the target bidding document.

Optionally, fig. 10 is a schematic structural diagram of an object recommendation device for implementing IA by combining RPA and AI according to an embodiment of the present application. As shown in fig. 10, the object recommendation apparatus 9000 further includes a target document obtaining unit 9005, a set extracting unit 9006 and a graph establishing unit 9007, for, before obtaining first knowledge representation information corresponding to the bid document on the first knowledge graph using the target knowledge representation model:

a target document acquiring unit 9005, configured to acquire, by using a robot process automation RPA system, a bid-solicited document and a bid document set in a target document format;

the set extraction unit 9006 is configured to obtain a bid document and a triplet information set corresponding to the bid document set by using a triplet extraction technique;

and an atlas establishing unit 9007, configured to establish a first knowledge atlas corresponding to the bidding document and the bidding document set based on the triple information set.

Optionally, fig. 11 is a schematic structural diagram of an object recommendation device that implements IA by combining RPA and AI according to an embodiment of the present application. As shown in fig. 11, the set obtaining unit 9006 includes a set obtaining subunit 9106, a participle processing subunit 9206 and an auditing processing subunit 9306, and the set obtaining unit 9006 is configured to, when obtaining a triple information set corresponding to a bid document and a bid document set by using a triple extraction technique:

the set obtaining subunit 9106 is configured to obtain, by using a triple extraction technology, a first triple information set corresponding to the bid document and a second triple information set corresponding to the bid document set;

a word segmentation processing subunit 9206, configured to perform word segmentation processing on the first triple information set and the second triple set by using a crust word segmentation model to obtain a processed first triple information set and a processed second triple set;

and an auditing processing subunit 9306, configured to perform entity boundary character string auditing processing on the processed first triple information set and the processed second triple information set, so as to obtain triple information sets corresponding to the bid document and the bid document set.

Optionally, fig. 12 is a schematic structural diagram of an object recommendation device for implementing IA by combining RPA and AI according to an embodiment of the present application. As shown in fig. 12, the object recommending apparatus 9000 further includes a set acquiring unit 9008, an atlas acquiring unit 9009 and a model establishing unit 9010, for, before acquiring first knowledge representation information corresponding to a bid document using a target knowledge representation model and second knowledge representation information corresponding to at least one bid document in the bid document:

a set acquiring unit 9008, configured to acquire a set of bidding training documents;

the map acquisition unit 9009 is used for acquiring a knowledge map corresponding to the bid-inviting and bidding training document set;

the model establishing unit 9010 is configured to perform vector space mapping on at least one knowledge pair in the knowledge graph based on a distributed representation learning inference technique, and establish a target knowledge representation model.

Optionally, fig. 13 is a schematic structural diagram of an object recommendation device that implements IA by combining RPA and AI according to an embodiment of the present application. As shown in fig. 13, the object recommendation apparatus 9000 further includes a parameter obtaining unit 9011 and a model determining unit 9012, which are configured to, before obtaining first knowledge representation information corresponding to a bid document and second knowledge representation information corresponding to at least one bid document in the bid document by using a target knowledge representation model:

a parameter obtaining unit 9011, configured to obtain a model performance parameter corresponding to each initial knowledge representation model in the initial knowledge representation model set;

and the model determining unit 9012 is configured to determine a target knowledge representation model in the initial knowledge representation model set based on the model performance parameters, where the target knowledge representation model is a distributed knowledge inference model.

Optionally, fig. 14 is a schematic structural diagram of an object recommendation device that implements IA by combining RPA and AI according to an embodiment of the present application. As shown in fig. 14, the object recommendation apparatus 9000 further includes an attribute information obtaining unit 9013, an information adjusting unit 9014, and a set traversing unit 9015, configured to, before obtaining the matching degree set of the first knowledge representation information and the at least one second knowledge representation information by using the matching network model:

an attribute information acquiring unit 9013, configured to acquire bid attribute information corresponding to a bid document and bid attribute information corresponding to any bid document in a bid document set;

the information adjusting unit 9014 is configured to adjust the first knowledge representation information by using the user representation learning module and the bid attribution information to obtain adjusted first knowledge representation information;

the information adjusting unit 9014 is further configured to adjust second knowledge representation information corresponding to any bid document by using the user representation learning module and the bid attribute information, so as to obtain adjusted second knowledge representation information;

and the set traversing unit 9015 is configured to traverse the second knowledge representation information set to obtain the adjusted second knowledge representation information set.

Optionally, fig. 15 is a schematic structural diagram of an object recommendation device that implements IA by combining RPA and AI according to an embodiment of the present application. As shown in fig. 15, the document determining unit 9003 further includes a matching degree obtaining subunit 9103 and a document matching subunit 9203, and the document determining unit 9003 is configured to, when determining a target bid document among the at least one bid document based on the set of matching degrees:

a matching degree obtaining subunit 9103, configured to obtain the highest matching degree in the matching degree set;

and the document matching subunit 9203 is configured to obtain, from the at least one bid document, a bid document corresponding to the highest matching degree, and determine the bid document as a target bid document.

The functions of each module in each apparatus in the embodiment of the present application may refer to corresponding descriptions in the above method, and are not described herein again.

In the embodiment of the application, a bidding document and a bidding document set are acquired through a document acquisition unit by a Robot Process Automation (RPA) system; the information acquisition unit acquires first knowledge representation information corresponding to the bid inviting document on a first knowledge map by adopting a target knowledge representation model, and acquires second knowledge representation information corresponding to at least one bid document in the bid document, wherein the first knowledge map corresponds to the bid inviting document and the bid document set; the document determining unit respectively acquires a matching degree set between the first knowledge representation information and the at least one second knowledge representation information by adopting a matching network model, and determines a target bidding document in the at least one bidding document based on the matching degree set; the object recommending unit recommends the object to be bid corresponding to the object bidding document. Therefore, the target knowledge representation model is adopted to obtain the knowledge representation information corresponding to the bid document and the bid document on the first knowledge map, and recommendation is performed based on the knowledge representation information corresponding to the bid document and the bid document, so that the background knowledge of bid and bid can be fully considered when the terminal performs object recommendation, and the accuracy of object recommendation can be improved.

Fig. 16 shows a block diagram of a terminal according to an embodiment of the present application. As shown in fig. 16, the terminal includes: a memory 1610 and a processor 1620, wherein the memory 1610 has stored therein computer programs that are executable on the processor 1620. The processor 1620 implements the object recommendation method in the above-described embodiments when executing the computer program. The number of the memory 1610 and the processor 1620 may be one or more.

The terminal further includes:

and a communication interface 1630, configured to communicate with an external device for data interactive transmission.

If the memory 1610, the processor 1620 and the communication interface 1630 are implemented independently, the memory 1610, the processor 1620 and the communication interface 1630 may be connected to each other through a bus and perform communication with each other. The bus may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, an Extended ISA (EISA) bus, or the like. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one thick line is shown in FIG. 16, but this is not intended to represent only one bus or type of bus.

Alternatively, in an implementation, if the memory 1610, the processor 1620 and the communication interface 1630 are integrated on a chip, the memory 1610, the processor 1620 and the communication interface 1630 may communicate with each other through an internal interface.

Embodiments of the present application provide a computer-readable storage medium storing a computer program, which when executed by a processor implements the method provided in the embodiments of the present application.

The embodiment of the present application further provides a chip, where the chip includes a processor, and is configured to call and execute the instruction stored in the memory from the memory, so that the communication device in which the chip is installed executes the method provided in the embodiment of the present application.

An embodiment of the present application further provides a chip, including: the system comprises an input interface, an output interface, a processor and a memory, wherein the input interface, the output interface, the processor and the memory are connected through an internal connection path, the processor is used for executing codes in the memory, and when the codes are executed, the processor is used for executing the method provided by the embodiment of the application.

It should be understood that the processor may be a Central Processing Unit (CPU), other general purpose processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, etc. A general purpose processor may be a microprocessor or any conventional processor or the like. It is noted that the processor may be an advanced reduced instruction set machine (ARM) architecture supported processor.

Further, optionally, the memory may include a read-only memory and a random access memory, and may further include a nonvolatile random access memory. The memory may be either volatile memory or nonvolatile memory, or may include both volatile and nonvolatile memory. The non-volatile memory may include a read-only memory (ROM), a Programmable ROM (PROM), an Erasable PROM (EPROM), an electrically Erasable EPROM (EEPROM), or a flash memory. Volatile memory can include Random Access Memory (RAM), which acts as external cache memory. By way of example, and not limitation, many forms of RAM are available. For example, static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), synchronous Dynamic Random Access Memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous SDRAM (ESDRAM), synchronous Link DRAM (SLDRAM), and direct bus RAM (DR RAM).

In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. The procedures or functions according to the present application are generated in whole or in part when the computer program instructions are loaded and executed on a computer. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium.

In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the application. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.

Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the feature. In the description of the present application, "a plurality" means two or more unless specifically limited otherwise.

Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more executable instructions for implementing specific logical functions or steps of the process. And the scope of the preferred embodiments of the present application includes other implementations in which functions may be performed out of the order shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved.

The logic and/or steps represented in the flowcharts or otherwise described herein, e.g., an ordered listing of executable instructions that can be considered to implement logical functions, can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions.

It should be understood that portions of the present application may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the various steps or methods may be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. All or part of the steps of the method of the above embodiments may be implemented by hardware that is configured to be instructed to perform the relevant steps by a program, which may be stored in a computer-readable storage medium, and which, when executed, includes one or a combination of the steps of the method embodiments.

In addition, functional units in the embodiments of the present application may be integrated into one processing module, or each unit may exist alone physically, or two or more units are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. The integrated module may also be stored in a computer-readable storage medium if it is implemented in the form of a software functional module and sold or used as a separate product. The storage medium may be a read-only memory, a magnetic or optical disk, or the like.

While the present invention has been described with reference to the preferred embodiments, it will be understood by those skilled in the art that various changes and modifications may be made without departing from the spirit and scope of the invention as defined by the appended claims. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. An object recommendation method for realizing IA by combining RPA and AI, which is characterized by comprising the following steps:

acquiring first knowledge representation information corresponding to bidding documents on a first knowledge map by adopting a target knowledge representation model, and acquiring second knowledge representation information corresponding to at least one bidding document in the bidding document set, wherein the first knowledge map corresponds to the bidding documents and the bidding document set;

2. The method of claim 1, further comprising, before the obtaining the first knowledge representation information corresponding to the bidding document on the first knowledge graph using the target knowledge representation model, the steps of:

acquiring a bid-inviting document and a bid document set in a target document format through a Robot Process Automation (RPA) system;

acquiring the bidding document and a triple information set corresponding to the bidding document set by adopting a triple extraction technology;

3. The method of claim 2, wherein the obtaining of the triplet information sets corresponding to the bidding documents and the bidding document sets by using the triplet extraction technique comprises:

and performing entity boundary character string auditing processing on the processed first triple information set and the processed second triple information set to obtain triple information sets corresponding to the bidding documents and the bidding document sets.

4. The method of claim 1, further comprising, before the obtaining first knowledge representation information corresponding to bidding documents and second knowledge representation information corresponding to at least one of the bidding documents by using the target knowledge representation model, the steps of:

acquiring a bidding training document set;

acquiring a knowledge graph corresponding to the bidding training document set;

and carrying out vector space mapping on at least one knowledge pair in the knowledge graph based on a distributed representation learning inference technology, and establishing a target knowledge representation model.

5. The method of claim 1, further comprising, before the obtaining first knowledge representation information corresponding to bidding documents and second knowledge representation information corresponding to at least one of the bidding documents by using the target knowledge representation model, the steps of:

6. The method according to claim 1, further comprising, before said obtaining the set of degrees of matching of the first knowledge representation information and the at least one second knowledge representation information using the matching network model:

adjusting the first knowledge representation information by adopting a user representation learning module and the bidding attribute information to obtain adjusted first knowledge representation information;

adjusting second knowledge representation information corresponding to any bidding document by adopting the user representation learning module and the bidding attribute information to obtain adjusted second knowledge representation information;

and traversing the second knowledge representation information set to obtain an adjusted second knowledge representation information set.

7. The method of claim 1, wherein said determining a target bid document of said at least one bid document based on said set of matching degrees comprises:

acquiring the highest matching degree in the matching degree set;

and obtaining the bidding document corresponding to the highest matching degree from the at least one bidding document, and determining the bidding document as the target bidding document.

8. An object recommendation device for realizing IA by combining RPA and AI, comprising:

the information acquisition unit is used for acquiring first knowledge representation information corresponding to a bidding document on a first knowledge map by adopting a target knowledge representation model, and acquiring second knowledge representation information corresponding to at least one bidding document in the bidding document set, wherein the first knowledge map corresponds to the bidding document and the bidding document set;

9. The apparatus according to claim 8, wherein the apparatus further comprises a target document obtaining unit, a set extracting unit and a graph establishing unit, and before the obtaining of the first knowledge representation information corresponding to the bidding document on the first knowledge graph by using the target knowledge representation model, the target document obtaining unit is configured to:

the target document acquisition unit is used for acquiring the bid-inviting document and the bid document set in the target document format through a Robot Process Automation (RPA) system;

the set extraction unit is used for acquiring the bidding document and a triple information set corresponding to the bidding document set by adopting a triple extraction technology;

the map establishing unit is used for establishing a first knowledge map corresponding to the bidding document and the bidding document set based on the triple information set.

10. The apparatus according to claim 9, wherein the set obtaining unit includes a set obtaining subunit, a word segmentation processing subunit, and an auditing processing subunit, and the set obtaining unit is configured to, when obtaining the triple information sets corresponding to the bid document and the bid document set by using a triple extraction technique:

the set acquisition subunit is configured to acquire a first triple information set corresponding to the bid document and a second triple information set corresponding to the bid document set by using a triple extraction technology;

the word segmentation processing subunit is configured to perform word segmentation processing on the first triple information set and the second triple information set by using a crust word segmentation model to obtain a processed first triple information set and a processed second triple information set;

and the auditing processing subunit is configured to perform entity boundary character string auditing processing on the processed first triple information set and the processed second triple information set to obtain triple information sets corresponding to the bid document and the bid document set.

11. The apparatus according to claim 8, wherein the apparatus further comprises a set obtaining unit, a map obtaining unit and a model establishing unit, and before the obtaining of the first knowledge representation information corresponding to the bidding document and the obtaining of the second knowledge representation information corresponding to at least one bidding document in the bidding documents by using the target knowledge representation model:

the collection acquisition unit is used for acquiring a bidding training document collection;

the map acquisition unit is used for acquiring a knowledge map corresponding to the bid-inviting and bidding training document set;

the model establishing unit is used for carrying out vector space mapping on at least one knowledge pair in the knowledge graph based on a distributed representation learning inference technology to establish a target knowledge representation model.

12. The apparatus according to claim 8, further comprising a parameter obtaining unit and a model determining unit, configured to, before the obtaining of the first knowledge representation information corresponding to the bid document and the obtaining of the second knowledge representation information corresponding to at least one of the bid documents by using the target knowledge representation model, obtain:

13. The apparatus according to claim 8, further comprising an attribute information obtaining unit, an information adjusting unit, and a set traversing unit, configured to, before obtaining the matching degree set of the first knowledge representation information and the at least one second knowledge representation information by using the matching network model:

the attribute information acquisition unit is used for acquiring bidding attribute information corresponding to the bidding document and bidding attribute information corresponding to any bidding document in the bidding document set;

the information adjusting unit is used for adjusting the first knowledge representation information by adopting a user representation learning module and the bidding attribute information to obtain adjusted first knowledge representation information;

the information adjusting unit is further configured to adjust second knowledge representation information corresponding to any bid document by using the user representation learning module and the bid attribute information to obtain adjusted second knowledge representation information;

14. A terminal for realizing IA by combining RPA and AI, comprising: a processor and a memory, the memory storing instructions therein, the instructions being loaded and executed by the processor to implement the method of any of claims 1 to 7.

15. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the method according to any one of claims 1 to 7.