AI intelligent customer service interaction method for foreign student renting and related components

Abstract

The application relates to an AI intelligent customer service interaction method for overseas student house renting and related components, which is applied to the technical field of overseas student house renting, and the method comprises the following steps: obtaining user consultation information and a user portrait, wherein the user consultation information at least comprises one of text, voice and an image; identifying the user consultation information, converting the identified content into standard terminology, obtaining current user demand, wherein the current user demand comprises a demand type, and the demand type comprises a house source recommendation type, a consultation type and a contract risk early warning type; matching potential user demand based on the user portrait and the current user demand; determining user intention based on the current user demand and the potential user demand; and determining demand response information based on the user intention, the demand type and an overseas student house renting vertical knowledge graph. The application has the effects of improving the efficiency and reliability of overseas student house renting response.

Description

Technical Field

[0001] This application relates to the technical field of overseas student housing rental, and in particular to an AI-powered intelligent customer service interaction method and related components for overseas student housing rental. Background Technology

[0002] With the continuous development of the overseas study market, the number of international students is constantly increasing, leading to a growing demand for rental housing. International students face numerous challenges in the rental process, making the development of the rental market crucial for meeting their housing needs and ensuring the stability of their studies and lives. Good rental services can help international students better adapt to life abroad. With the widespread adoption of internet technology, online rentals are gradually becoming the mainstream trend, providing international students with more choices and convenience.

[0003] In traditional overseas student housing scenarios, the following methods are commonly used to address students' housing inquiries and other related issues: First, there's human customer service, where students communicate with customer service personnel via phone, email, or in-person visits. Customer service staff use their experience and knowledge to answer questions and recommend properties. Second, there are static website displays, where housing platforms showcase housing information and surrounding amenities, allowing students to browse and find properties that meet their needs. Third, there's a simple keyword search function, where students enter keywords such as area or price, and the system filters relevant properties based on those keywords.

[0004] However, these technologies all have significant drawbacks. Human customer service is inefficient, struggling to handle a large volume of inquiries from international students simultaneously, and customer service personnel have limited knowledge and experience, potentially failing to fully and accurately meet the needs of international students. Static webpage displays and simple keyword search functions lack a precise understanding and matching of the personalized needs of international students, failing to provide targeted housing recommendations and detailed answers based on their specific circumstances, and thus failing to effectively meet the diverse and personalized rental needs of international students. Summary of the Invention

[0005] To improve the efficiency and reliability of responses to international students' housing rental requests, this application provides an AI-powered intelligent customer service interaction method and related components for international students' housing rental needs.

[0006] Firstly, this application provides an AI-powered intelligent customer service interaction method for international students renting apartments, employing the following technical solution: An AI-powered intelligent customer service interaction method for international students renting apartments includes: Obtain user consultation information and user profiles, wherein the user consultation information includes at least one of text, voice and image; The user consultation information is identified, and the identified content is converted into standard terminology to obtain the current user needs. The current user needs include the type of needs, which includes property recommendation type, consultation type, and contract risk warning type. Match potential user needs based on the user profile and the current user needs; Determine user intent based on the current user needs and the potential user needs; Demand response information is determined based on the user intent, the demand type, and the vertical knowledge graph of international student housing rentals.

[0007] By employing the aforementioned technical solutions, user inquiry information can be automatically identified and converted into standard terminology. This allows for the rapid and accurate determination of current user needs. Matching potential user needs with current user profiles and further identifying user intent improves the reliability of user intent. Combining user intent, need type, and a vertical knowledge graph of international student housing determines the appropriate response information. This approach accurately understands the diverse and personalized housing needs of international students, providing targeted housing recommendations, consultations, and contract risk warnings. It addresses the issues of low efficiency and lack of accurate matching in traditional overseas student housing services, thereby improving the efficiency and reliability of responses to international student housing inquiries.

[0008] Optionally, the step of identifying the user consultation information and converting the identified content into standard terminology to obtain the current user needs includes: The text is identified using natural language processing technology to obtain the first identified content; The speech is recognized using automatic speech recognition technology to obtain the second recognized content; The image is recognized based on a preset visual language model to obtain the third recognition content; The first identification content, the second identification content, and the third identification content are summarized to obtain the identification content; Based on a pre-set fault-tolerant library of rental terms for international students, the non-standard terms in the identified content are converted into standard terms to obtain the current user requirements.

[0009] By adopting the above technical solution, using natural language processing technology, automatic speech recognition technology and preset visual language models to identify and summarize text, speech and images respectively, it is possible to comprehensively and accurately identify user consultation information. Then, by using a preset error-tolerant library of terms for international students renting rooms, non-standard terms in the identified content are converted into standard terms, which can convert user consultation information into standard current user needs, making it easier to accurately match potential user needs and determine user intent in the future.

[0010] Optionally, matching potential user needs based on the user profile and the current user needs includes: Extract the core profile information corresponding to the preset core dimensions in the user profile; Based on the core profile information, the current user needs, and preset mining factors, the first candidate potential needs are matched from the preset rule base for mining potential housing needs of international students. Determine whether there is a conflict between the first candidate potential needs and / or between the first candidate potential needs and the current user needs; If there is a conflict between the first candidate potential needs and / or between the first candidate potential needs and the current user needs, then the conflict is reconciled based on the weights corresponding to the various preset mining factors to obtain the second candidate potential needs; The user profile matching degree of the second candidate potential demand is scored to obtain the first score; The second candidate potential demand is scored based on its relevance to the current demand, resulting in a second score. Based on the first score and the second score, the potential demand score for each of the second candidate potential demands is determined; The second candidate potential needs are sorted from high to low according to the potential need scores to obtain the candidate potential need ranking results; The first preset number of second candidate potential needs in the candidate potential needs ranking results are determined as the potential user needs.

[0011] By adopting the above technical solution, core profile information is extracted from user profiles, and first candidate potential needs are matched with current user needs. Conflicting first candidate potential needs are reconciled to obtain second candidate potential needs. Then, the user profile matching degree and the relevance of the current needs are scored to obtain a potential need score. The second candidate potential needs are ranked according to the potential need score, and the potential user needs are finally determined. Compared with traditional rental methods, this method can more accurately understand and match the personalized needs of international students, and improve the efficiency and reliability of the rental response for international students.

[0012] Optionally, before determining the demand response information based on the user intent, the demand type, and the vertical knowledge graph of international student housing rentals, the method further includes: The vertical knowledge graph for international students' housing rental is constructed, which includes entities and the relationships between entities. The entities include housing resources, universities, communities, transportation stations, living facilities, rental regulations, and cultural background information. The vertical knowledge graph for international students' housing rental is stored in a graph database. Obtain updated property listing information through a preset interface; The attributes of the entity are dynamically updated in real time based on the updated property information.

[0013] By adopting the above technical solutions, a vertical knowledge graph for international students' housing rentals is constructed, storing entities and relationships such as housing listings, universities, and communities in a graph database. This allows for the comprehensive integration of housing rental-related information. By obtaining updated housing listing information through preset interfaces and dynamically updating entity attributes in real time, the timeliness and accuracy of the information in the vertical knowledge graph for international students' housing rentals can be guaranteed. This provides a reliable foundation for determining demand response information based on the vertical knowledge graph for international students' housing rentals, and better meets the diverse housing rental needs of overseas students.

[0014] Optionally, the demand response information includes recommended housing information, and the determination of demand response information based on the user intent, the demand type, and a vertical knowledge graph of international student housing rentals includes: If the demand type is the housing recommendation type, then candidate housing information is matched from the vertical knowledge graph of international student housing based on the user intent and multi-hop reasoning. The candidate housing information includes recommended housing, recommendation reasons and matching degree scores. The authenticity of the candidate housing information is verified. The candidate property listings that have passed the property listing authenticity verification are sorted from high to low according to the matching degree score to obtain the property listing ranking result; The first second preset number of candidate properties in the property ranking results are determined as recommended properties. The preference display method is determined based on the user profile; The recommended housing information is displayed to the user based on the aforementioned preference display method.

[0015] By adopting the above technical solution, when the demand type is housing recommendation, candidate housing information is matched from the vertical knowledge graph of international student housing based on user intent and multi-hop reasoning. This can accurately locate housing that meets the user's needs, and the inclusion of recommendation reasons and matching scores allows users to better understand the housing. Verifying the authenticity of candidate housing information ensures the authenticity and reliability of the recommended housing. Sorting the verified housing according to the matching score and selecting the top two preset number of housings as recommended housing information prioritizes showing users the most suitable housing. Determining the preferred display method based on user profile and displaying recommended housing information can meet users' personalized display needs, improve user experience, and effectively solve the problem that traditional housing rental methods cannot accurately meet the personalized housing rental needs of international students.

[0016] Optionally, the demand response information includes consultation response content, and the determination of demand response information based on the user intent, the demand type, and the vertical knowledge graph of international student housing rentals includes: If the demand type is the consultation type, then the vertical knowledge graph of international student housing rentals is searched based on the user intent to determine the search results; If there is missing content in the search results, then obtain the landlord information corresponding to the user's intent; Based on the landlord information, determine the communication time and script; During the communication period, supplementary information corresponding to the missing content is obtained from the landlord according to the communication script. The content of the consultation response is determined based on the search results and the supplementary information. The preference display method is determined based on the user profile; The consultation response content is displayed to the user based on the aforementioned preference display method.

[0017] By adopting the above technical solution, and targeting specific consultation needs, if there are missing results after searching the vertical knowledge graph of international student housing, supplementary information can be obtained by using landlord information to determine the communication time and script, thereby improving the comprehensiveness and accuracy of consultation response content.

[0018] Optionally, determining the demand response information based on the user intent, the demand type, and the vertical knowledge graph of international student housing rentals includes: If the demand type is the contract risk warning type, then the contract terms in the user intent are compared with the corresponding entities in the vertical knowledge graph of international student housing rental to determine the risk content; Based on the aforementioned risk content, a risk warning report and modification suggestions will be generated; The required response information is determined based on the risk warning report and the proposed modifications.

[0019] By adopting the above technical solution, risk warnings can be provided for rental contracts for international students. Comparing the contract terms in the user's intent with the corresponding entities in the vertical knowledge graph of international student rentals can accurately identify risky content. Based on this, a risk warning report and modification suggestions are generated, and demand response information is determined. This helps international students promptly identify contract risks and obtain corresponding solutions, improving rental security.

[0020] Secondly, this application provides an AI-powered intelligent customer service interaction system for international students renting apartments, employing the following technical solution: An AI-powered intelligent customer service interaction system for international students renting apartments includes: The user demand acquisition module is used to acquire user consultation information and user profiles. The user consultation information includes at least one of text, voice and image. The user demand identification module is used to identify the user consultation information and convert the identified content into standard terms to obtain the current user demand. The current user demand includes demand types, which include property recommendation type, consultation type and contract risk warning type. The potential demand matching module is used to match potential user demands based on the user profile and the current user demands. The user intent determination module is used to determine the user intent based on the current user needs and the potential user needs; The user demand response module is used to determine demand response information based on the user intent, the demand type, and the vertical knowledge graph of international student housing rentals.

[0021] By employing the aforementioned technical solutions, user inquiry information can be automatically identified and converted into standard terminology. This allows for the rapid and accurate determination of current user needs. Matching potential user needs with current user profiles and further identifying user intent improves the reliability of user intent. Combining user intent, need type, and a vertical knowledge graph of international student housing determines the appropriate response information. This approach accurately understands the diverse and personalized housing needs of international students, providing targeted housing recommendations, consultations, and contract risk warnings. It addresses the issues of low efficiency and lack of accurate matching in traditional overseas student housing services, thereby improving the efficiency and reliability of responses to international student housing inquiries.

[0022] Thirdly, this application provides an electronic device that adopts the following technical solution: An electronic device includes a processor coupled to a memory; The memory stores a computer program that can be loaded by a processor and executed by the AI-powered intelligent customer service interaction method for international students renting apartments, as described in any of the first aspects.

[0023] Fourthly, this application provides a computer-readable storage medium, which adopts the following technical solution: A computer-readable storage medium storing a computer program capable of being loaded by a processor and executing the AI-powered intelligent customer service interaction method for international students renting apartments as described in any of the first aspects. Attached Figure Description

[0024] Figure 1 This is a flowchart illustrating an AI-powered intelligent customer service interaction method for international students renting apartments, as provided in an embodiment of this application.

[0025] Figure 2 This is a structural block diagram of an AI-powered intelligent customer service interaction system for international students renting apartments, provided in an embodiment of this application.

[0026] Figure 3 This is a structural block diagram of the electronic device provided in the embodiments of this application. Detailed Implementation

[0027] The present application will be further described in detail below with reference to the accompanying drawings.

[0028] This application provides an AI-powered intelligent customer service interaction method for international students renting apartments. This method can be executed by an electronic device, which can be a server or a terminal device. The server can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services. The terminal device can be a smartphone, tablet, desktop computer, etc., but is not limited to these.

[0029] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0030] Furthermore, the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article, unless otherwise specified, generally indicates that the preceding and following related objects have an "or" relationship.

[0031] like Figure 1 As shown, an AI-powered intelligent customer service interaction method for international students renting apartments is described in the following steps (S101-S105): Step S101: Obtain user consultation information and user profile.

[0032] User inquiry information includes at least one of the following: text, voice, and image. User inquiry information can also be presented in multiple forms at the same time.

[0033] Users can input text (including text or documents), voice, and images through the interactive interface. Text and voice can be information describing the user's needs, such as "I want to find an apartment near the school"; images can be pictures related to the property, such as photos of the room, photos of the surrounding environment, etc.; and user inquiry information can be obtained from the interactive interface.

[0034] User profiles can be obtained based on user interaction history and personal information filled in by users. User profiles include multiple dimensions, such as: basic attributes (country / region of study abroad, target school, study duration, visa type, language ability, etc.), preference attributes (rent budget, apartment type preference, decoration style, supporting facilities requirements, whether to accept co-renting, etc.), behavioral attributes (historical search keywords, property click records, consultation history, characteristics of rejecting / saving properties, etc.), scenario attributes (planned move-in time, rental duration, whether it is the first time renting overseas, etc.), and system usage attributes (login records, etc.).

[0035] Step S102: Identify the user's inquiry information and convert the identified content into standard terms to obtain the current user's needs, which include the type of needs.

[0036] The types of requests include property recommendations, inquiries, and contract risk warnings. For example, a property recommendation request might be for a two-bedroom apartment with a balcony near UCL in London, with a monthly rent not exceeding £1,800, and suitable for sharing. An inquiry request might be for a property with payment options such as "What are the payment methods for property A?" A contract risk warning request might be for a property with potential risks such as "Please check if there are any risks associated with this contract."

[0037] Specifically, the process involves identifying user inquiry information and converting the identified content into standard terminology to obtain the current user needs. This includes: identifying text based on natural language processing technology to obtain the first identified content; identifying speech based on automatic speech recognition technology to obtain the second identified content; identifying images based on a preset visual language model to obtain the third identified content; summarizing the first, second, and third identified content to obtain the identified content; and converting non-standard terms in the identified content into standard terms based on a preset error-tolerant library of terms for international student housing rentals to obtain the current user needs.

[0038] In this embodiment, the user-input text is recognized using Natural Language Processing (NLP) technology (e.g., BERT-BiLSTM fusion model) (including but not limited to word segmentation, semantic entity extraction, etc.) to obtain the first recognized content, which includes core semantic information. Speech is recognized using Automatic Speech Recognition (ASR) technology (e.g., FunASR, Whisper, etc.) (including but not limited to noise reduction, speech segmentation, speech-to-text processing, and extraction of emotional features from the speech) to obtain the second recognized content. A preset visual language model can be a visual language model pre-trained with a large number of labeled image samples, such as a convolutional neural network model. This preset visual language model can extract features from user-uploaded images, such as room layout, decoration style, lighting conditions, and furniture configuration. Based on this preset visual language model, the image is recognized to obtain the third recognized content. All three recognition methods support multiple languages.

[0039] The first, second, and third identified content are summarized and deduplicated to obtain the identified content. The preset terminology tolerance library for overseas students renting apartments is a dedicated terminology library built for the overseas student rental scenario. The stored content includes: standard overseas rental terminology (categorized by country of study), mapping relationship between non-standard terminology / colloquial expressions and standard terminology, mutual translation relationship of terminology in different countries, and terminology tolerance rules (such as spelling error correction, abbreviation expansion, etc.). The preset terminology tolerance library for overseas students renting apartments converts non-standard terminology in the identified content into standard terminology, and analyzes user consultation information through a preset demand type determination rule library to determine the demand type. The demand type and the standardized terminology identification content are jointly determined as the current user demand.

[0040] The pre-defined demand type determination rule library contains various demand type determination rules. For example: if a user's inquiry contains core words such as "find a house," "recommend," or "rent a house," and is accompanied by property attributes such as rent, apartment type, and location, then the demand type is property recommendation; if a user's inquiry contains core words such as "excuse me," "how to do it," or "what rules," and addresses issues such as rental policies, payment methods, and deposit refunds, then the demand type is inquiry; if a user's inquiry contains core words such as "contract," "terms," ​​"risk," or "review," and is accompanied by contract-related information, then the demand type is contract risk warning.

[0041] Step S103: Match potential user needs based on user profiles and current user needs.

[0042] Specifically, matching potential user needs based on user profiles and current user needs includes: extracting core profile information corresponding to preset core dimensions in the user profile; matching first candidate potential needs from a preset rule base for mining potential needs of international students renting apartments based on the core profile information, current user needs, and preset mining factors; determining whether there are conflicts between the first candidate potential needs and / or between the first candidate potential needs and the current user needs; if there are conflicts between the first candidate potential needs and / or between the first candidate potential needs and the current user needs, then reconciling the conflicts based on the weights corresponding to various preset mining factors to obtain second candidate potential needs; scoring the user profile matching degree of the second candidate potential needs to obtain a first score; scoring the relevance of the second candidate potential needs to the current needs to obtain a second score; determining the potential need score for each second candidate potential need based on the first score and the second score; sorting the second candidate potential needs from high to low according to the potential need score to obtain the candidate potential need ranking result; and determining the first preset number of second candidate potential needs in the candidate potential need ranking result as potential user needs.

[0043] In this embodiment, the core profile information is the profile information corresponding to the preset core dimensions in the user profile. The preset core dimensions may include basic attributes, preference attributes, behavioral attributes, scenario attributes, etc. Based on the core profile information, current user needs, and preset mining factors, the first candidate potential needs are matched from the preset rule base for mining potential needs of international students renting apartments. The preset rule base for mining potential needs of international students renting apartments is a set of rules pre-summarized based on a large amount of historical data and experience. The preset mining factors may include profile-needs correlation factors, cross-cultural adaptation factors, behavioral trajectory reverse inference factors, dynamic scenario adaptation factors, and security guarantee supplementary factors, etc.

[0044] The profile-demand correlation factor is to match the first candidate potential demand from the preset rule base for mining potential demand for international students' housing rentals based on the core profile information and the current user demand. For example, if the core profile information includes the target school as UCL and the current user demand includes "two bedrooms and one living room", then the first candidate potential demand includes roommates who are mostly UCL students.

[0045] Cross-cultural adaptation factors are based on the country / region of study abroad and match the corresponding rental culture from the pre-set rule library of potential rental needs of international students. That is, the first candidate potential needs. For example, if the country of study abroad is Japan, the first candidate potential needs include "accepting key money / gift money, no pets".

[0046] The behavioral trajectory reverse inference factor is to match the corresponding potential needs from the pre-set rule library of potential housing needs of international students by using the user's historical behavior. That is, the first candidate potential needs. For example, if a user filters housing within 10 minutes of the campus multiple times, but does not click on the housing, the first candidate potential needs include being close to the campus and cost-effective, and rejecting housing close to the campus that exceeds the budget.

[0047] The dynamic scenario adaptation factor is to combine the current time and the dynamics of the housing market to match the corresponding time-sensitive potential needs from the preset rule base of potential housing needs of international students. That is, the first candidate potential needs. For example, if the user's check-in time is September (back-to-school season), the first candidate potential needs include being able to book one month in advance and supporting short-term rental transition.

[0048] The supplementary safety guarantee factors are matched with the safety needs corresponding to the pain points of international students renting houses across borders from the pre-set rule base for mining potential needs of international students. For example, if the authenticity of the listing is not mentioned in the current user's needs, the first candidate potential needs include that the pictures are not stolen and the listing address can be verified.

[0049] Determine whether there is a conflict between any two first candidate potential needs and / or between the first candidate potential needs and the current user needs, for example, a conflict between a preference for Nordic style and affordability; if there is no conflict between any two first candidate potential needs or between the first candidate potential needs and the current user needs, then determine all first candidate potential needs as second candidate potential needs.

[0050] If there is a conflict between the first candidate potential demand and the current user demand, the first candidate potential demand that conflicts with the current user demand is deleted. For the remaining first candidate potential demands, if there is a conflict between the first candidate potential demands, the weights corresponding to various preset mining factors are obtained from the database. The first candidate potential demand with the higher corresponding weight among the conflicting first candidate potential demands is determined as the second candidate potential demand. At the same time, the second candidate potential demand also includes the first candidate potential demand that does not conflict.

[0051] The user profile matching score is applied to the second candidate potential demand to obtain a first score; the relevance score of the second candidate potential demand to the current demand is applied to obtain a second score. The first and second scores can be determined by retrieving preset user profile matching score rules and preset current demand relevance score rules from the database, applying the preset user profile matching score rules to each second candidate potential demand to obtain the first score, and applying the preset current demand relevance score rules to each second candidate potential demand to obtain the second score. The preset user profile matching score rules and preset current demand relevance score rules are pre-set by staff and are not specifically limited here. For example, the second candidate demand can be calculated... The first text similarity between potential needs and user profiles is selected. The first score is determined based on the correlation between the first text similarity and the first score in the preset user profile matching score rules. The higher the first text similarity, the higher the first score. The second text similarity between the second candidate potential needs and the current user needs is calculated. The second score is determined based on the correlation between the second text similarity and the second score in the preset current need relevance score rules. The higher the second text similarity, the higher the second score. Alternatively, a lightweight neural network model (MLP multilayer perceptron) can be used to score the user profile matching degree and the current need relevance to obtain the first score and the second score. The lightweight neural network model is a scoring model trained with a large number of scoring samples.

[0052] The potential demand score for each second candidate potential demand is calculated as follows: First score × Preset profile matching weight + Second score × Preset demand association weight. The preset profile matching weight and the preset demand association weight are both preset and are not specifically limited here. The second candidate potential demands are sorted from highest to lowest according to their potential demand scores to obtain the candidate potential demand ranking result. The first preset number of second candidate potential demands in the candidate potential demand ranking result (preset and not specifically limited here) are determined as potential user demands.

[0053] Step S104: Determine user intent based on current user needs and potential user needs.

[0054] The current user needs and potential user needs are combined to determine the user intent.

[0055] Step S105: Determine the demand response information based on user intent, demand type, and the vertical knowledge graph of international student housing rentals.

[0056] The demand response information includes recommended housing information, inquiry response content, risk warning reports corresponding to the contract, and contract modification suggestions.

[0057] Specifically, before determining the demand response information based on user intent, demand type, and the vertical knowledge graph of international student housing rentals, the method also includes: constructing a vertical knowledge graph of international student housing rentals, which includes entities and relationships between entities. Entities include housing, universities, communities, transportation stations, living facilities, rental regulations, and cultural background information. The vertical knowledge graph of international student housing rentals is stored in a graph database; obtaining housing update information through a preset interface; and dynamically updating the attributes of entities in real time based on the housing update information.

[0058] Unlike general knowledge graphs, the vertical knowledge graph for international student housing is specifically designed and tailored to the international student housing scenario. This knowledge graph stores comprehensive information needed for international students to rent housing in a structured format. It includes various entities and their relationships. Entities include housing listings, universities, communities, transportation hubs, amenities, rental regulations (distinguished by country / state / county), and cultural context clues (such as US credit score requirements and UK security deposit regulations). Relationships describe the connections between entities, such as: Housing - [Location] -> Community - [Commuting Time] -> University; Community - [Safety Rating] -> High / Medium / Low; Housing - [Applicable Regulations] -> Rental Regulations, etc. This vertical knowledge graph for international student housing is built by integrating official data from multiple countries, real estate website information, and user-generated content. The vertical knowledge graph for international student housing rentals is stored in a graph database. It can also connect bidirectionally or unidirectionally with the databases of multiple mainstream global housing platforms (such as Zillow, Rightmove, Realtor.com, Domain.com.au, etc.) via preset interfaces to periodically or event-triggeredly obtain housing update information, such as price changes, inventory status (available / booked / rented), new listings, and delistings. The system dynamically updates entity attributes in real time based on this information. The preset interfaces can be either RESTful APIs or GraphQL interfaces.

[0059] Specifically, the system determines demand response information based on user intent, demand type, and a vertical knowledge graph of international student housing. This includes: if the demand type is housing recommendation, matching candidate housing information from the vertical knowledge graph of international student housing based on user intent and multi-hop reasoning. Candidate housing information includes recommended housing, reasons for recommendation, and matching score; verifying the authenticity of candidate housing information; sorting the candidate housing information that has passed the authenticity verification according to the matching score from high to low to obtain the housing ranking result; determining the top two preset number of candidate housing information in the housing ranking result as recommended housing information; determining the preferred display method based on user profile; and displaying the recommended housing information to the user based on the preferred display method.

[0060] In this embodiment, if the demand type is a housing recommendation type, candidate housing information is matched from the vertical knowledge graph of international student housing based on the user intent and the multi-hop reasoning method corresponding to the vertical knowledge graph of international student housing. The multi-hop reasoning process can be completed by the native reasoning engine of the graph database (e.g., Neo4j). Specifically, it includes implementing multi-hop reasoning through Cypher statements and completing housing matching degree scoring through custom user functions (UDFs). For example, starting from the target school entity in the user profile, the first hop finds community entities with reasonable commuting time (e.g., less than 30 minutes), the second hop filters based on the community's safety rating and living facilities, and the third hop matches housing entities within the qualified community according to the user intent. Custom user functions include calculating the text similarity between the housing description text and the user intent text in the candidate housing information, and determining the matching degree score based on the preset relationship between text similarity and matching degree score. The content of Cypher statements and custom user functions (UDFs) is not specifically limited here.

[0061] After obtaining candidate property listings, computer vision technology is used to perform digital watermark detection, EXIF ​​metadata analysis (including shooting time, location, and device analysis), and reverse image search on the property images in the candidate listings to identify stolen images from the internet or fake images generated by AI, thereby completing the property authenticity verification. The candidate property listings that pass the authenticity verification are then sorted from highest to lowest according to their matching score, resulting in a property ranking result. The top two preset number (pre-set, not specifically limited here) of candidate property listings in the property ranking result are determined as recommended property listings. Recommended property listings include not only the matching score but also a risk score (when the risk score is below a preset threshold, it can alert the user of potential risks). The risk score is determined based on the property's historical price stability and the originality of the property listing text. For example: Price fluctuation coefficient = (Property's most...) / ... The calculation is as follows: High Rent - Lowest Rent of the Property in the Past Six Months / Average Rent of the Property in the Past Six Months; Originality Coefficient = 1 - Number of Characters with the Most Overlapping with Other Property Descriptions on the Entire Network / Total Number of Characters in the Current Property Description; Price Fluctuation Score is determined based on the correlation between the preset Price Fluctuation Coefficient and Price Fluctuation Score; Originality Score is determined based on the correlation between the preset Originality Coefficient and Originality Score; Risk Score is the weighted sum of Price Fluctuation Score and Originality Score (specific weights are preset and not specifically limited here); User profiles are analyzed using data analysis tools (e.g., Python) to determine preferred display methods (e.g., text, text-to-speech (TTS), images, video links, VR viewing entrances, etc.); Multimedia materials from the multimedia library are used to display recommended property information to users according to their preferred display methods.

[0062] The multimedia resource library stores images, videos, VR links, and other multimedia materials associated with the properties, which can be used for recommendations and displays. The multimedia resource library and the actual properties are linked through a vertical knowledge graph of international student housing rentals.

[0063] Specifically, the system determines the response information based on user intent, demand type, and a vertical knowledge graph of international student housing, including: if the demand type is a consultation, the system searches the vertical knowledge graph of international student housing based on the user intent to determine the search results; if there is missing content in the search results, the system obtains the landlord information corresponding to the user intent; the system determines the communication time and script based on the landlord information; during the communication time, the system obtains supplementary information corresponding to the missing content from the landlord according to the communication script; the system determines the consultation response content based on the search results and supplementary information; the system determines the preferred display method based on the user profile; and the system displays the consultation response content to the user based on the preferred display method.

[0064] In this embodiment, if the demand type is a consultation, the vertical knowledge graph for international student housing rentals is searched according to the user's intent to determine the search results. For example, if the user is looking for UK deposit regulations, the search results can be directly obtained from the vertical knowledge graph for international student housing rentals. If there is missing content in the search results, such as the user inquiring about the payment method for rent of property A, and this information is not included in the vertical knowledge graph for international student housing rentals, the landlord information corresponding to the user's intent (e.g., property A) is obtained. The landlord information can be obtained from the vertical knowledge graph for international student housing rentals or collected through other channels. The communication time is determined based on the working hours in the landlord information, and the corresponding standardized communication script is retrieved from the preset international student housing rental communication script library based on the landlord's location in the landlord information. During the communication time, supplementary information corresponding to the missing content is obtained from the landlord according to the communication script. The search results and supplementary information are jointly determined as the consultation response content. The consultation response content is displayed to the user according to the user's preferred display method (this process is the same as the display process for the property recommendation type mentioned above, and will not be repeated here).

[0065] Specifically, demand response information is determined based on user intent, demand type, and the vertical knowledge graph of international student housing rentals. This includes: if the demand type is a contract risk warning type, comparing the contract terms in the user intent with the corresponding entities in the vertical knowledge graph of international student housing rentals to determine the risk content; generating a risk warning report and modification suggestions based on the risk content; and determining demand response information based on the risk warning report and modification suggestions.

[0066] In this embodiment, if the demand type is a contract risk warning type, the contract terms in the user's intent are compared with the corresponding entities in the vertical knowledge graph of international student housing rental to determine the risk content. The risk warning report adopts the method of "fixed template + dynamic content filling". The preset template is retrieved from the preset template library according to the risk content, and the dynamic content is automatically retrieved and filled from the "risk item list + international student housing rental vertical knowledge graph" according to the risk content, thereby generating a risk warning report. Modification suggestions are retrieved from the contract risk warning library according to the risk content. The risk warning report and modification suggestions are jointly determined as the demand response information.

[0067] Figure 2 This is a structural block diagram of an AI-powered intelligent customer service interaction system 200 for international students renting apartments, provided as an embodiment of this application.

[0068] like Figure 2 As shown, the AI-powered intelligent customer service interaction system 200, designed for international students renting apartments, mainly includes: User demand acquisition module 201 is used to acquire user inquiry information and user profiles. User inquiry information includes at least one of text, voice and image. The user demand identification module 202 is used to identify user consultation information and convert the identified content into standard terms to obtain the current user demand. The current user demand includes demand type, which includes property recommendation type, consultation type and contract risk warning type. The potential demand matching module 203 is used to match potential user demands based on user profiles and current user demands. User intent determination module 204 is used to determine user intent based on current user needs and potential user needs; User demand response module 205 is used to determine demand response information based on user intent, demand type, and vertical knowledge graph of international student housing rental.

[0069] As an optional implementation of this embodiment, the user demand identification module 202 is specifically used to identify user consultation information and convert the identified content into standard terms to obtain the current user demand, including: identifying text based on natural language processing technology to obtain first identified content; identifying speech based on automatic speech recognition technology to obtain second identified content; identifying images based on a preset visual language model to obtain third identified content; summarizing the first identified content, the second identified content, and the third identified content to obtain identified content; and converting non-standard terms in the identified content into standard terms based on a preset error-tolerant library of terms for international student housing rentals to obtain the current user demand.

[0070] As an optional implementation of this embodiment, the potential demand matching module 203 is specifically used to match potential user demands based on user profiles and current user demands, including: extracting core profile information corresponding to preset core dimensions in the user profile; matching first candidate potential demands from a preset rule base for mining potential demands of international students renting apartments based on the core profile information, current user demands, and preset mining factors; determining whether there is a conflict between the first candidate potential demands and / or between the first candidate potential demands and the current user demands; if there is a conflict between the first candidate potential demands and / or between the first candidate potential demands and the current user demands, then reconciling the conflict based on the weights corresponding to various preset mining factors to obtain second candidate potential demands; scoring the user profile matching degree of the second candidate potential demands to obtain a first score; scoring the current demand relevance of the second candidate potential demands to obtain a second score; determining the potential demand score of each second candidate potential demand based on the first score and the second score; sorting the second candidate potential demands from high to low according to the potential demand score to obtain a candidate potential demand sorting result; and determining the first preset number of second candidate potential demands in the candidate potential demand sorting result as potential user demands.

[0071] As an optional implementation of this embodiment, the user demand response module 205 is specifically used to, before determining the demand response information based on user intent, demand type, and the vertical knowledge graph of international student housing rentals, further include: constructing a vertical knowledge graph of international student housing rentals, which includes entities and relationships between entities, and entities include housing, universities, communities, transportation stations, living facilities, rental regulations, and cultural background information, and the vertical knowledge graph of international student housing rentals is stored in a graph database; obtaining housing update information through a preset interface; and dynamically updating the attributes of entities in real time based on the housing update information.

[0072] As an optional implementation of this embodiment, the demand response information includes recommended housing information. The user demand response module 205 is specifically used to determine the demand response information based on user intent, demand type, and the vertical knowledge graph of international student housing rentals. This includes: if the demand type is housing recommendation, then matching candidate housing information from the vertical knowledge graph of international student housing rentals based on user intent and multi-hop reasoning. The candidate housing information includes recommended housing, reasons for recommendation, and matching degree score; verifying the authenticity of the candidate housing information; sorting the candidate housing information that has passed the authenticity verification according to the matching degree score from high to low to obtain the housing ranking result; determining the top two preset number of candidate housing information in the housing ranking result as recommended housing information; determining the preferred display method based on the user profile; and displaying the recommended housing information to the user based on the preferred display method.

[0073] As an optional implementation of this embodiment, the demand response information includes consultation response content. The user demand response module 205 is specifically used to determine the demand response information based on user intent, demand type, and the vertical knowledge graph of international student housing rentals. This includes: if the demand type is consultation, then searching the vertical knowledge graph of international student housing rentals based on user intent to determine the search results; if there is missing content in the search results, then obtaining the landlord information corresponding to the user intent; determining the communication time period and communication script based on the landlord information; obtaining supplementary information corresponding to the missing content from the landlord according to the communication script during the communication time period; determining the consultation response content based on the search results and supplementary information; determining the preferred display method based on the user profile; and displaying the consultation response content to the user based on the preferred display method.

[0074] As an optional implementation of this embodiment, the user demand response module 205 is specifically used to determine demand response information based on user intent, demand type, and the vertical knowledge graph of international student housing rentals, including: if the demand type is a contract risk warning type, then comparing the contract terms in the user intent with the corresponding entities in the vertical knowledge graph of international student housing rentals to determine the risk content; generating a risk warning report and modification suggestions based on the risk content; and determining demand response information based on the risk warning report and modification suggestions.

[0075] In one example, a module in any of the above systems may be one or more integrated circuits configured to implement the above methods, such as one or more application-specific integrated circuits (ASICs), or one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs), or a combination of at least two of these integrated circuit forms.

[0076] For example, when modules in a system can be implemented through a processing element scheduler, the processing element can be a general-purpose processor, such as a central processing unit (CPU) or other processor capable of calling programs. Alternatively, these modules can be integrated together to form a system-on-a-chip (SOC).

[0077] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the system and modules described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0078] Figure 3This is a structural block diagram of an electronic device 300 provided in an embodiment of this application.

[0079] like Figure 3 As shown, the electronic device 300 includes a processor 301 and a memory 302, and may further include one or more of an information input / output (I / O) interface 303, a communication component 304, and a communication bus 305.

[0080] The processor 301 controls the overall operation of the electronic device 300 to complete all or part of the steps of the AI-powered intelligent customer service interaction method for international students renting apartments. The memory 302 stores various types of data to support the operation of the electronic device 300. This data may include, for example, instructions for any application or method operating on the electronic device 300, as well as application-related data. The memory 302 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as one or more of Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0081] I / O interface 303 provides an interface between processor 301 and other interface modules, such as keyboards, mice, and buttons. These buttons can be virtual or physical. Communication component 304 is used for wired or wireless communication between electronic device 300 and other devices. Wireless communication includes Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G, or 4G, or a combination thereof. Therefore, the corresponding communication component 304 may include a Wi-Fi component, a Bluetooth component, and an NFC component.

[0082] The electronic device 300 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to execute the AI-powered intelligent customer service interaction method for international students renting apartments as described in the above embodiments.

[0083] The communication bus 305 may include a path for transmitting information between the aforementioned components. The communication bus 305 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The communication bus 305 may be divided into an address bus, a data bus, a control bus, etc.

[0084] Electronic device 300 may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital radio receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), and in-vehicle terminals (such as in-vehicle navigation terminals), as well as fixed terminals such as digital TVs and desktop computers, and may also be servers.

[0085] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described AI-powered intelligent customer service interaction method for international students renting apartments.

[0086] The computer-readable storage medium may include various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0087] The terms “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0088] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the foregoing application concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions claimed in this application.

Claims

1. An AI-powered intelligent customer service interaction method for international students renting apartments, characterized in that: include: Obtain user consultation information and user profiles, wherein the user consultation information includes at least one of text, voice and image; The user consultation information is identified, and the identified content is converted into standard terminology to obtain the current user needs. The current user needs include the type of needs, which includes property recommendation type, consultation type, and contract risk warning type. Match potential user needs based on the user profile and the current user needs; Determine user intent based on the current user needs and the potential user needs; Demand response information is determined based on the user intent, the demand type, and the vertical knowledge graph of international student housing rentals.

2. The method according to claim 1, characterized in that, The process of identifying the user's inquiry information and converting the identified content into standard terminology to obtain the current user's needs includes: The text is identified using natural language processing technology to obtain the first identified content; The speech is recognized using automatic speech recognition technology to obtain the second recognized content; The image is recognized based on a preset visual language model to obtain the third recognition content; The first identification content, the second identification content, and the third identification content are summarized to obtain the identification content; Based on a pre-set fault-tolerant library of rental terms for international students, the non-standard terms in the identified content are converted into standard terms to obtain the current user requirements.

3. The method according to claim 1, characterized in that, The matching of potential user needs based on the user profile and the current user needs includes: Extract the core profile information corresponding to the preset core dimensions in the user profile; Based on the core profile information, the current user needs, and preset mining factors, the first candidate potential needs are matched from the preset rule base for mining potential housing needs of international students. Determine whether there is a conflict between the first candidate potential needs and / or between the first candidate potential needs and the current user needs; If there is a conflict between the first candidate potential needs and / or between the first candidate potential needs and the current user needs, then the conflict is reconciled based on the weights corresponding to the various preset mining factors to obtain the second candidate potential needs; The user profile matching degree of the second candidate potential demand is scored to obtain the first score; The second candidate potential demand is scored based on its relevance to the current demand, resulting in a second score. Based on the first score and the second score, the potential demand score for each of the second candidate potential demands is determined; The second candidate potential needs are sorted from high to low according to the potential need scores to obtain the candidate potential need ranking results; The first preset number of second candidate potential needs in the candidate potential needs ranking results are determined as the potential user needs.

4. The method according to claim 1, characterized in that, Before determining the demand response information based on the user intent, the demand type, and the vertical knowledge graph of international student housing rentals, the method further includes: The vertical knowledge graph for international students' housing rental is constructed, which includes entities and the relationships between entities. The entities include housing resources, universities, communities, transportation stations, living facilities, rental regulations, and cultural background information. The vertical knowledge graph for international students' housing rental is stored in a graph database. Obtain updated property listing information through a preset interface; The attributes of the entity are dynamically updated in real time based on the updated property information.

5. The method according to claim 1, characterized in that, The demand response information includes recommended housing information. The process of determining the demand response information based on the user intent, the demand type, and a vertical knowledge graph of international student housing includes: If the demand type is the housing recommendation type, then candidate housing information is matched from the vertical knowledge graph of international student housing based on the user intent and multi-hop reasoning. The candidate housing information includes recommended housing, recommendation reasons and matching degree scores. The authenticity of the candidate housing information is verified. The candidate property listings that have passed the property listing authenticity verification are sorted from high to low according to the matching degree score to obtain the property listing ranking result; The first second preset number of candidate properties in the property ranking results are determined as the recommended properties. The preference display method is determined based on the user profile; The recommended housing information is displayed to the user based on the aforementioned preference display method.

6. The method according to claim 1, characterized in that, The demand response information includes consultation response content. Determining the demand response information based on the user intent, the demand type, and a vertical knowledge graph of international student housing rentals includes: If the demand type is the consultation type, then the vertical knowledge graph of international student housing rentals is searched based on the user intent to determine the search results; If there is missing content in the search results, then obtain the landlord information corresponding to the user's intent; Based on the landlord information, determine the communication time and script; During the communication period, supplementary information corresponding to the missing content is obtained from the landlord according to the communication script. The content of the consultation response is determined based on the search results and the supplementary information. The preference display method is determined based on the user profile; The consultation response content is displayed to the user based on the aforementioned preference display method.

7. The method according to claim 1, characterized in that, The process of determining the demand response information based on the user intent, the demand type, and the vertical knowledge graph of international student housing rentals includes: If the demand type is the contract risk warning type, then the contract terms in the user intent are compared with the corresponding entities in the vertical knowledge graph of international student housing rental to determine the risk content; Based on the aforementioned risk content, a risk warning report and modification suggestions will be generated; The required response information is determined based on the risk warning report and the proposed modifications.

8. An AI-powered intelligent customer service interaction system for international students renting apartments, characterized in that: include: The user demand acquisition module is used to acquire user consultation information and user profiles. The user consultation information includes at least one of text, voice and image. The user demand identification module is used to identify the user consultation information and convert the identified content into standard terms to obtain the current user demand. The current user demand includes demand types, which include property recommendation type, consultation type and contract risk warning type. The potential demand matching module is used to match potential user demands based on the user profile and the current user demands. The user intent determination module is used to determine the user intent based on the current user needs and the potential user needs. The user demand response module is used to determine demand response information based on the user intent, the demand type, and the vertical knowledge graph of international student housing rentals.

9. An electronic device, characterized in that, Includes a processor, which is coupled to a memory; The processor is configured to execute a computer program stored in the memory, causing the electronic device to perform the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, Includes a computer program or instructions that, when run on a computer, cause the computer to perform the method as described in any one of claims 1 to 7.

Images