Tourism information pushing method and device, electronic equipment and storage medium

Abstract

The invention provides a travel information pushing method and device, electronic equipment and a storage medium, and relates to the technical field of machine learning. The method comprises the following steps: acquiring first text travel data in a target webpage; determining prompt information of the text travel task, and generating second text travel data based on the prompt information by using the initial text travel model; generating third text travel data based on the first text travel data and the second text travel data by using a pre-trained sequence generative adversarial network; performing fine tuning on the initial text travel model based on the first text travel data, the second text travel data and the third text travel data to obtain a target text travel model; and obtaining a travel demand of the user through the text travel assistant application program, calling the target text travel model, generating text travel information corresponding to the travel demand through the target text travel model, and pushing the text travel information to the user. According to the method, the target text travel model is called through the text travel assistant application program, personalized text travel information is generated for the user, the model accuracy is improved, and the user experience is improved.

Description

Technical Field

[0001] The present invention relates to the field of machine learning technology, and in particular to a method, device, electronic device and storage medium for pushing cultural and tourism information. Background Art

[0002] With the rapid development of information technology and the continuous improvement of people's living standards, the tourism industry has shown a booming development trend. People's requirements for the enrichment, refinement and personalization of cultural and tourism information are also increasing. However, in actual travel scenarios, tourists often find it difficult to obtain comprehensive and accurate destination information. At the same time, most of the existing tourism services on the market lack pertinence and cannot fully meet the personalized needs of different tourists.

[0003] In order to solve the above problems, relevant technologies usually use big data and deep learning technology to collect and analyze massive cultural and tourism data, extract the real needs of tourists, and generate travel suggestions and services that are closer to personal preferences. However, the way the above methods process cultural and tourism data is often difficult to cope with complex and changing cultural and tourism scenarios. On the one hand, the original cultural and tourism data may contain noise and incompleteness, which affects the accuracy of the analysis results; on the other hand, it is difficult to cover all potential tourism needs by relying solely on historical cultural and tourism data. Summary of the invention

[0004] In view of this, the embodiments of the present disclosure propose a method, device, electronic device and computer-readable storage medium for pushing cultural and tourism information to solve the problems existing in related technologies, such as the possibility of noise and incompleteness in the original cultural and tourism data, which affects the accuracy of the analysis results, and the problem that the way of processing cultural and tourism data simply relying on historical cultural and tourism data is difficult to cover all potential tourism needs, and is often difficult to cope with complex and changeable cultural and tourism scenarios.

[0005] According to the first aspect of the present disclosure, a method for pushing cultural and tourism information is provided, including: obtaining first cultural and tourism data in a target web page; determining prompt information of a cultural and tourism task, and generating second cultural and tourism data based on the prompt information using an initial cultural and tourism model; wherein the initial cultural and tourism model is a pre-trained large language model; generating third cultural and tourism data based on the first cultural and tourism data and the second cultural and tourism data using a pre-trained sequence generative adversarial network; fine-tuning the initial cultural and tourism model based on the first cultural and tourism data, the second cultural and tourism data, and the third cultural and tourism data to obtain a target cultural and tourism model; obtaining travel needs input by a user through a cultural and tourism assistant application and calling the target cultural and tourism model, generating cultural and tourism information corresponding to the travel needs through the target cultural and tourism model, and pushing the cultural and tourism information to the user.

[0006] According to the second aspect of the present disclosure, a cultural and tourism information push device is provided, including: a data acquisition module, used to acquire first cultural and tourism data in a target web page; the data acquisition module is also used to determine prompt information of a cultural and tourism task, and use an initial cultural and tourism model to generate second cultural and tourism data based on the prompt information; wherein the initial cultural and tourism model is a pre-trained large language model; the data acquisition module is also used to use a pre-trained sequence generative adversarial network to generate third cultural and tourism data based on the first cultural and tourism data and the second cultural and tourism data; a model training module is used to fine-tune the initial cultural and tourism model based on the first cultural and tourism data, the second cultural and tourism data, and the third cultural and tourism data to obtain a target cultural and tourism model; an information push module is used to acquire travel needs input by a user through a cultural and tourism assistant application and call the target cultural and tourism model, generate cultural and tourism information corresponding to the travel needs through the target cultural and tourism model, and push the cultural and tourism information to the user.

[0007] According to a third aspect of the present disclosure, an electronic device is provided, comprising: a processor; and a memory storing a program, wherein the program comprises instructions, and when the instructions are executed by the processor, the processor executes the above-mentioned cultural and travel information push method.

[0008] According to a fourth aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable the computer to execute the above-mentioned cultural and tourism information pushing method.

[0009] The exemplary embodiments of the present disclosure may have some or all of the following beneficial effects: In the cultural and tourism information push method provided in the example implementation mode of the present disclosure, the first cultural and tourism data in the target web page is obtained; the prompt information of the cultural and tourism task is determined, and the second cultural and tourism data is generated based on the prompt information using the initial cultural and tourism model; wherein the above-mentioned initial cultural and tourism model is a pre-trained large language model; the third cultural and tourism data is generated based on the first cultural and tourism data and the second cultural and tourism data using the pre-trained sequence generative adversarial network; the initial cultural and tourism model is fine-tuned based on the first cultural and tourism data, the second cultural and tourism data and the third cultural and tourism data to obtain the target cultural and tourism model; the travel needs input by the user are obtained through the cultural and tourism assistant application and the target cultural and tourism model is called, and the cultural and tourism information corresponding to the travel needs is generated through the target cultural and tourism model, and the cultural and tourism information is pushed to the user. The embodiment of the present application uses the first cultural and tourism data obtained from the target web page, the second cultural and tourism data generated based on the initial cultural and tourism model, and the third cultural and tourism data generated by the sequence generative adversarial network as a training data set for fine-tuning the above-mentioned initial cultural and tourism model, enriching the source of data, reducing the impact of noise and data incompleteness, and improving the accuracy of the target cultural and tourism model obtained through training; in addition, the embodiment of the present application accepts the user's travel needs through the above-mentioned cultural and tourism assistant application and calls the target cultural and tourism model, so that the target cultural and tourism model can generate a customized travel guide according to the user's travel needs, and provide personalized services, effectively integrating tourism resources and enhancing the user's travel experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Further details, features and advantages of the present disclosure are disclosed in the following description of exemplary embodiments in conjunction with the accompanying drawings, in which: Figure 1 A flow chart of a method for pushing cultural and tourism information according to an exemplary embodiment of the present disclosure is shown; Figure 2 A schematic diagram of the structure of multi-head attention introduced in a generator of a sequence generation adversarial network in a method for pushing cultural and tourism information according to an exemplary embodiment of the present disclosure is shown; Figure 3 A schematic diagram of a network structure of an improved sequence generation adversarial network in a method for pushing cultural and tourism information according to an exemplary embodiment of the present disclosure is shown; Figure 4 A schematic diagram of the interface of a Chinese travel assistant application in a method for pushing cultural travel information according to an exemplary embodiment of the present disclosure is shown; Figure 5 A schematic diagram of the interface of a Chinese travel assistant application in a method for pushing cultural travel information according to an exemplary embodiment of the present disclosure is shown; Figure 6 A schematic block diagram of a device for pushing cultural and tourism information according to an exemplary embodiment of the present disclosure is shown; Figure 7A structural block diagram of an exemplary electronic device that can be used to implement the embodiments of the present disclosure is shown. DETAILED DESCRIPTION

[0011] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein, which are instead provided for a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0012] It should be understood that the various steps described in the method embodiments of the present disclosure may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this respect.

[0013] The term "including" and its variations used in this document are open inclusions, that is, "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one other embodiment"; the term "some embodiments" means "at least some embodiments". Relevant definitions of other terms will be given in the description below. It should be noted that the concepts of "first", "second", etc. mentioned in this disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0014] It should be noted that the modifications of "one" and "plurality" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0015] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0016] The embodiments of the present disclosure provide a method, device, electronic device and computer-readable storage medium for pushing cultural and tourism information. The following describes the scheme of the present disclosure with reference to the accompanying drawings for illustration.

[0017] The present disclosure embodiment first provides a method for pushing cultural and tourism information. Figure 1 As shown, the method for pushing cultural and tourism information specifically includes the following steps: S110: Acquire the first cultural and travel data in the target webpage; S120: Determine prompt information of the cultural tourism task, and generate second cultural tourism data based on the prompt information using an initial cultural tourism model; wherein the initial cultural tourism model is a pre-trained large language model; S130: Generate third cultural and tourism data based on the first cultural and tourism data and the second cultural and tourism data using a pre-trained sequence generation adversarial network; S140: fine-tuning the initial cultural tourism model based on the first cultural tourism data, the second cultural tourism data, and the third cultural tourism data to obtain a target cultural tourism model; S150: Obtain the travel needs input by the user through the cultural and travel assistant application and call the target cultural and travel model, generate cultural and travel information corresponding to the travel needs through the target cultural and travel model, and push the cultural and travel information to the user.

[0018] The cultural and tourism information push method provided by the example implementation method of the present disclosure uses the first cultural and tourism data obtained from the target web page, the second cultural and tourism data generated based on the initial cultural and tourism model, and the third cultural and tourism data generated by the sequence generative adversarial network as a training data set for fine-tuning the above-mentioned initial cultural and tourism model, thereby enriching the source of data, reducing the impact of noise and data incompleteness, and improving the accuracy of the target cultural and tourism model obtained through training; in addition, the embodiment of the present application accepts the user's travel needs through the above-mentioned cultural and tourism assistant application and calls the target cultural and tourism model, so that the target cultural and tourism model can generate a customized travel guide according to the user's travel needs, and provide personalized services, effectively integrating tourism resources and enhancing the user's travel experience.

[0019] Next, in another embodiment, the above steps are described in more detail.

[0020] In step S110, first cultural and travel data in the target webpage is obtained.

[0021] In the embodiment of the present application, the target webpage may be a webpage containing cultural and tourism information in a tourism information platform. For example, the cultural and tourism information may be an introduction to a scenic spot, local customs, accommodation, transportation, etc. The target webpage may be indicated by a Uniform Resource Locator (URL), and specifically, the URL may point to a detailed information page of the cultural and tourism information.

[0022] In the embodiment of the present application, the first cultural and tourism data is the data after processing the cultural and tourism information extracted from the target web page. For example, the cultural and tourism information can be automatically collected from the target web pages of major tourism information platforms using web crawler technology, and the collected cultural and tourism information is processed by data deduplication, error checking, and standardized conversion to obtain the first cultural and tourism data.

[0023] Specifically, the above-mentioned acquisition of the first cultural and tourism data in the target webpage can be implemented as follows: S1: Use web crawler technology to collect cultural and tourism information.

[0024] In the embodiment of the present application, the above-mentioned collection of cultural and tourism information using web crawler technology can be specifically implemented as follows: initialize the URL list, select the target web page to be crawled according to the characteristics of different tourism information platforms, and generate an initialized URL list. The URLs in the URL list usually point to specific scenic spots, hotels, transportation and other detailed information pages; design web page request management, and use the Scrapy framework in Python to send Hypertext Transfer Protocol (HTTP) requests to obtain web page content. For dynamically loaded content, user operations are simulated through tools such as Selenium to load JavaScript content. Among them, the above-mentioned Scrapy framework is a Python framework for crawling website data and extracting structured information; the above-mentioned Selenium is a tool for testing Web applications, which can simulate various operations of users in the browser, such as clicking buttons, entering text, selecting drop-down menus, scrolling pages, etc.; the above-mentioned JavaScript is an advanced interpreted programming language, mainly used for web page development.

[0025] S2: Analyze and extract the crawled cultural and tourism information.

[0026] In the embodiment of the present application, in the process of collecting cultural and tourism information through the web crawler technology, the web page data obtained is usually presented in the format of HyperText Markup Language (HTML) or lightweight data exchange format (JavaScript Object Notation, JSON) / eXtensible Markup Language (XML). The above parsing and extraction of the crawled cultural and tourism information can be specifically implemented as follows: use the BeautifulSoup and lxml libraries in Python to parse the web page structure and extract key information fields, such as the name of the scenic spot, geographical location, ticket price, opening hours, user evaluation, etc. Among them, the above BeautifulSoup is a Python library for parsing and processing HTML and XML documents, which provides simple functions and methods to navigate, search and modify the analysis tree, making it easier to extract data from the web page; the above lxml is a Python library for processing XML and HTML documents, which combines an efficient parser and a convenient application programming interface (Application ProgrammingInterface, API), making it easy to parse, modify and generate XML / HTML documents.

[0027] S3: Process the extracted data to obtain first cultural and tourism data.

[0028] In an embodiment of the present application, the above-mentioned data processing process can be specifically implemented as follows: data deduplication, cleaning up duplicate web page data to avoid the same information being captured multiple times; error checking, checking whether the data format meets expectations, and repairing parsing errors or missed fields; standardization conversion, unified processing of field formats, such as standardization of time, currency and other information, to ensure that data from different sources can be seamlessly integrated.

[0029] In the embodiment of the present application, the first cultural and tourism data is collected from the target web page through the above steps S1 to S3, ensuring the comprehensiveness, accuracy and consistency of the basic cultural and tourism information, and providing a high-quality data source for the subsequent training and application of cultural and tourism models.

[0030] In step S120, the prompt information of the cultural and tourism task is determined, and the initial cultural and tourism model is used to generate second cultural and tourism data based on the prompt information; wherein the initial cultural and tourism model is a pre-trained large language model.

[0031] In an embodiment of the present application, the above-mentioned cultural and tourism tasks are tasks generated based on the user's needs in cultural and tourism. Exemplarily, the cultural and tourism tasks can be scenic spot description generation tasks, travel guide generation tasks, and user evaluation summary tasks. Among them, the above-mentioned scenic spot description generation task requires the above-mentioned initial cultural and tourism model to generate a detailed scenic spot introduction based on the input keywords, including historical background, cultural characteristics, geographical location and other aspects; the above-mentioned travel guide generation task requires the above-mentioned initial cultural and tourism model to generate a travel guide based on specific needs, including the best tour route, recommended attractions, special restaurants, best travel time and other contents; the above-mentioned user evaluation summary task requires the above-mentioned initial cultural and tourism model to generate an evaluation summary based on the real user evaluation data collected from the Internet to help users make better decisions.

[0032] In the embodiment of the present application, the above-mentioned initial cultural tourism model is a pre-trained large-scale language model, which can be used to perform the above-mentioned cultural tourism tasks. Among them, the large-scale language model (LLM) is an artificial intelligence model based on deep learning technology, which has powerful language understanding, generation and reasoning capabilities.

[0033] In the embodiment of the present application, the prompt information (Prompt) refers to the text content input by the user when interacting with the initial cultural tourism model, which is used to guide the initial cultural tourism model to generate specific output.

[0034] In the embodiment of the present application, the second cultural and tourism data is the cultural and tourism data generated based on the initial cultural and tourism model. Exemplarily, the second cultural and tourism data can be obtained in the following ways: select the initial cultural and tourism model based on actual needs, and guide the initial cultural and tourism model to generate text content such as scenic spot descriptions and travel guides corresponding to the cultural and tourism tasks based on the prompt words by customizing specific cultural and tourism tasks and designing tourism-related prompt words; screen the generated text content based on evaluation indicators such as bilingual evaluation understudy (BLEU) and recall-oriented automatic summary evaluation indicators (ROUGE), and combine manual review to ensure quality, and obtain the second cultural and tourism data; merge the obtained second cultural and tourism data with the first cultural and tourism data to form a rich and consistent data set to provide support for subsequent model training.

[0035] Specifically, the above-mentioned use of the initial cultural tourism model to generate the second cultural tourism data based on the prompt information can be implemented as follows: S1: Determine the initial cultural tourism model and customize multiple cultural tourism tasks.

[0036] In the embodiment of the present application, this step is used to select a pre-trained large-scale language model as the above-mentioned initial cultural tourism model based on the actual needs of the cultural tourism industry, and customize multiple cultural tourism tasks. For example, the Tongyi Qianwen-72B (Qwen-72B) model can be used as the above-mentioned initial cultural tourism model.

[0037] S2: Determine the prompt words of the initial cultural tourism model, so as to guide the initial cultural tourism model to generate text content corresponding to the above cultural tourism tasks through the prompt words.

[0038] In the embodiment of the present application, this step is used to set prompts for the above-mentioned initial cultural tourism model, and guide the initial cultural tourism model to generate text content corresponding to the above-mentioned cultural tourism task based on the prompts. The design of the prompts needs to be optimized according to the task objectives to ensure that the generated content meets the actual needs. For example, the prompts for the above-mentioned scenic spot description generation task can include the name of the scenic spot, geographical location, historical background, etc.

[0039] S3: Evaluate the text content generated by the initial cultural tourism model to obtain the second cultural tourism data.

[0040] In the embodiment of the present application, this step is used to evaluate the text content generated by the initial cultural tourism model. For example, the similarity between the model-generated text and the reference text can be quantified by automated evaluation indicators (such as BLEU, ROUGE, etc.). At the same time, for scenarios with high content quality requirements, it is also necessary to combine manual review to further screen the generated text and obtain the second cultural tourism data to ensure the accuracy, fluency and practicality of tourism information.

[0041] S4: Integrate the second cultural and tourism data with the above-mentioned first cultural and tourism data.

[0042] In the embodiment of the present application, this step is used to integrate the first cultural tourism data and the second cultural tourism data. For example, after the second cultural tourism data is obtained through screening and evaluation in step S3, the obtained second cultural tourism data is integrated with the first cultural tourism data collected through the network.

[0043] In the embodiment of the present application, the second cultural tourism data generated by the initial cultural tourism model is used as a supplement to the first cultural tourism data, which can make up for the deficiency of network collected data and provide richer and more comprehensive information. In the integration process, it is necessary to ensure the consistency of the format and content of data from different sources.

[0044] In step S130, a pre-trained sequence generative adversarial network is used to generate third cultural and tourism data based on the first cultural and tourism data and the second cultural and tourism data.

[0045] In an embodiment of the present application, the above-mentioned sequence generative adversarial network (Sequence Generative AdversarialNets, SeqGAN) is a variant of a generative adversarial network (GAN) for sequence data generation. Specifically, a traditional SeqGAN network includes a generator and a discriminator. Among them, the generator is usually composed of a network based on a sequence model (such as RNN, LSTM or GRU) for generating discrete sequence data. During the generation process, the generator gradually generates each element of the sequence and outputs discrete symbols (such as words or characters) in turn. Due to the discreteness of the sequence data, the gradient cannot be obtained directly from the back propagation of GAN, so the generator is updated by the policy gradient (Policy Gradient) in reinforcement learning (RL) to deal with the gradient propagation problem. The discriminator is usually composed of a convolutional neural network (CNN) or a fully connected network, which is used to determine whether the sequence generated by the generator matches the true sequence, and evaluates the entire sequence to output a binary classification result (true or false).

[0046] The above-mentioned traditional sequence model network-based generator is difficult to capture the global dependencies in a long sequence when gradually generating a sequence, which will cause the generated sequence to be weak in long-range dependencies and global coherence. Therefore, preferably, the embodiment of the present application can introduce the following into the generator: Figure 2 The multi-head self-attention mechanism shown in Figure 1 is used to improve this problem, so that the generator can dynamically pay attention to the global information of the sequence at each generation moment, effectively enhance the modeling ability of the long-range dependency relationship between sequences, and significantly improve the quality and coherence of the generated sequence. For example, the network structure diagram of the improved SeqGAN network can be shown as follows: Figure 3 shown.

[0047] Among them, the above-mentioned multi-head self-attention mechanism allows the generator to dynamically pay attention to elements at other moments in the sequence when generating sequence elements at a certain moment, thereby capturing the long-range dependencies between sequences. The self-attention mechanism provides more global sequence information for each generation step. Through multiple parallel attention heads, the generator can simultaneously pay attention to features at different positions in the sequence and integrate this information for subsequent generation. This mechanism is particularly suitable for processing long sequences and significantly improves the performance of the generator in complex scenarios (such as natural language generation).

[0048] In an embodiment of the present application, the third cultural and tourism data is the cultural and tourism data generated by the improved SeqGAN network based on the first cultural and tourism data and the second cultural and tourism data. Exemplarily, the generation of the third cultural and tourism data based on the first cultural and tourism data and the second cultural and tourism data using the pre-trained sequence generation adversarial network can be implemented as follows: the target text sequence consisting of the first cultural and tourism data and the second cultural and tourism data is input into the sequence generation adversarial network; the context information of the current time step of the target text sequence is captured by the generator of the sequence generation adversarial network using the multi-head attention mechanism, and the output sequence probability distribution of the current time step is generated based on the context information of the current time step, and the output sequence of the current time step is determined based on the sequence probability distribution; the binary classification probability of the output sequence is determined by the discriminator of the sequence generation adversarial network, and the third cultural and tourism data is determined based on the binary classification probability.

[0049] Specifically, the process of generating the third cultural and tourism data based on the first cultural and tourism data and the second cultural and tourism data by using the improved SeqGAN network may include the following steps: S1: The text sequence Text (that is, the target text sequence) obtained by integrating the first cultural and tourism data and the second cultural and tourism data is input into the improved SeqGAN network model, and converted into a word vector through the Embedding layer as follows:

[0050] in, is the transformed text sequence word vector.

[0051] S2: The converted Input to the long short-term memory network (LSTM) structure for input sequence processing, and input the current word vector at each time step and the hidden state at the previous moment , and a recurrent neural network (RNN) is used to capture the short-term dependencies in the sequence as follows:

[0052] in, is the hidden state output at the current moment.

[0053] S3: At each time step, the generator uses the hidden state of the current time step as the query Q based on the self-attention mechanism, and the hidden states of all historical moments as the key K and value V as follows:

[0054] Among them, H is the historical hidden state matrix.

[0055] S4: Calculate the attention weight of the input word vector and add the multi-head attention mechanism as follows:

[0056] in, is the dimension of the health vector.

[0057]

[0058] in, .

[0059] S5: Combining the output of the RNN and the context information of the self-attention layer, the generator generates the output sequence probability distribution of the current time step through the Softmax layer, and selects the symbol with the highest probability as the generated output of the current time step to obtain the third cultural and tourism data. Among them, the above output sequence probability distribution is as follows:

[0060] S6: The discriminator receives the generated text sequence and the target text sequence from the generator, and converts each symbol of the generated text sequence and the target text sequence into a word vector through the word embedding layer as follows:

[0061] in, is the word vector of the converted text sequence.

[0062] The discriminator uses a convolutional layer to extract the features F of the input text sequence as follows:

[0063] Among them, CNN is a convolutional neural network layer.

[0064] The extracted features are processed through the fully connected layer, and the discriminator outputs a binary classification probability (true or false) as follows:

[0065] S7: Through the policy gradient method, the generator selects the output symbol according to the current state at each time step, and adds a reward mechanism to use the classification probability of the discriminator as a reward signal. Through reinforcement learning, the strategy is continuously optimized, and finally the generated sequence is consistent with the real sequence.

[0066] In an embodiment of the present application, a large amount of third cultural and tourism data can be generated based on the first cultural and tourism data and the second cultural and tourism data through the above process, thereby effectively reducing the time for collecting real data and improving the efficiency of subsequent model training.

[0067] In step S140, the initial cultural tourism model is fine-tuned based on the first cultural tourism data, the second cultural tourism data and the third cultural tourism data to obtain a target cultural tourism model.

[0068] In the embodiment of the present application, the target cultural tourism model is a large language model obtained by fine-tuning the initial cultural tourism model using the first cultural tourism data, the second cultural tourism data and the third cultural tourism data as training data sets. Compared with the initial cultural tourism model, the target cultural tourism model has better applicability to current cultural tourism needs.

[0069] Exemplarily, the process of fine-tuning the initial cultural and tourism model based on the first cultural and tourism data, the second cultural and tourism data, and the third cultural and tourism data to obtain the target cultural and tourism model can be implemented as follows: determine the training data set based on the first cultural and tourism data, the second cultural and tourism data, and the third cultural and tourism data; determine the fine-tuning parameters, and start the fine-tuning command of the initial cultural and tourism model based on the fine-tuning parameters. By running the fine-tuning command, the initial cultural and tourism model is fine-tuned with the fine-tuning parameters based on the training data set to obtain the target cultural and tourism model.

[0070] Specifically, the embodiment of the present application can adopt the instruction supervision fine-tuning (SFT) strategy as the fine-tuning strategy, and select the llama_factory tool platform for fine-tuning. Among them, the above llama_factory is a mainstream language large model fine-tuning framework, and the specific fine-tuning parameters can be shown in the following Table 1: Table 1:

[0071] After determining the above fine-tuning parameters, you can start the following fine-tuning command for the initial cultural tourism model to achieve fine-tuning:

[0072] During the above-mentioned fine-tuning process, the embodiment of the present application can also generate a loss function graph of the model based on the model log file. The loss function graph is composed of training steps and loss values. In order to improve the deployability of the model, the large model can be quantified based on the above-mentioned loss function graph, and the large language model generated after fine-tuning and precision quantization can be used as the above-mentioned target cultural and tourism model.

[0073] In step S150, the travel needs input by the user are obtained through the cultural and travel assistant application and the target cultural and travel model is called. The cultural and travel information corresponding to the travel needs is generated through the target cultural and travel model and pushed to the user.

[0074] In the embodiment of the present application, the above-mentioned cultural and tourism assistant application is an application developed based on the above-mentioned target cultural and tourism model. The user can call the above-mentioned target cultural and tourism model through the cultural and tourism assistant application to generate the cultural and tourism information required by the user through the target cultural and tourism model, and display it to the user through the cultural and tourism assistant application.

[0075] For example, the interface of the above-mentioned travel assistant application can be as follows: Figure 4As shown in the figure, the interface of the cultural and tourism assistant application includes controls such as recommended attractions, guides, travel notes, and cultural and creative products. Users can obtain corresponding cultural and tourism data by clicking on the corresponding controls, such as Figure 4 As shown, when the selected control is the recommended scenic spot control, the relevant content for scenic spot recommendations displayed on the Cultural and Tourism Assistant application interface.

[0076] In addition, the process of the user calling the target cultural travel model through the cultural travel assistant application can also be implemented based on the interaction between the user and the interface of the cultural travel assistant application. Figure 5 As shown, the user can publish cultural and tourism tasks to the cultural and tourism assistant application by inputting language or text, such as inputting "What are the precautions for hiking in this season", "Make a three-day travel guide for city A", "Recommendations for food and accommodation near scenic spot A", "I want to know the customs of city A" and other cultural and tourism tasks. When the cultural and tourism assistant application receives the cultural and tourism tasks, it can call the above-mentioned target cultural and tourism model through the pre-configured calling interface, generate the information corresponding to the cultural and tourism tasks input by the user and display it on the interface for the user to view.

[0077] Specifically, the above-mentioned cultural and travel assistant application may include functional modules such as a user information management module, a cultural and travel information generation module, and a user order management module. Among them: The above user information management module is used for user information collection and processing and data verification and synchronization. Specifically: The above user information collection and processing can be implemented as follows: the front end of the cultural and tourism assistant application receives user information (mobile phone number, ID number, etc.) and uploads it to the back-end server through form data. The user information is added and processed using a symmetric encryption algorithm, and then the encrypted information is used to generate a unique identity identifier ID using a hash algorithm, and finally the user information is stored in the database.

[0078] The above data verification and synchronization can be achieved as follows: by regularly calling the third-party real-name authentication interface to verify the legality and accuracy of the ID card information submitted by the tourist, the authenticity and reliability of the tourist's identity is ensured, and the above identity identifier ID is passed as a globally unique identifier in different modules.

[0079] The above-mentioned cultural and tourism information generation module is used to call the target cultural and tourism model, generate cultural and tourism information corresponding to travel needs through the target cultural and tourism model, and push the cultural and tourism information to the user.

[0080] Specifically, the above process can be implemented as follows: according to the travel needs input by the user, the travel needs are clarified based on the keywords input by the user and the travel date and destination selected by the user; the above target cultural tourism model is called, and the travel needs are input into the target cultural tourism model, so as to analyze the user's input and user's historical behavior through the target cultural tourism model, and generate a personalized travel plan in combination with external data sources (such as real-time traffic, weather, etc.). Among them, the contents of the plan include recommended scenic spots, tour route planning, dining recommendations, etc.; according to the user's real-time feedback (such as modifying the destination, adjusting the itinerary, etc.), the generated plan is dynamically adjusted through the feedback mechanism of reinforcement learning to ensure that the personalized needs of the user are met.

[0081] The above user order management module is used for information integration and display as well as online booking and payment. Among them: The above information integration and display can be achieved as follows: when generating personalized travel plans, key information such as scenic spot tickets, hotel accommodation, and transportation reservations are integrated into card-style displays. Each card contains key information (such as hotel name, price, etc.) and a booking link. Users can click on the card to jump to the corresponding third-party booking platform to complete the booking operation. By connecting to the API interface of third-party travel service providers, the system can obtain real-time price and inventory information of tickets, hotels, transportation, etc., and intuitively display the latest information to users.

[0082] The above online booking and payment can be implemented as follows: After the user clicks on the card, the system will carry the tourist's identity ID and booking information, and jump to the booking page of the third-party platform through a GET / POST request to ensure that the user does not need to enter the information again, thereby optimizing the booking experience. The payment link can integrate a third-party payment interface. After the payment is successful, the system obtains the payment result through an API callback and simultaneously updates the user's order status. At the same time, an order management module is set up in the system background to track the user's booking progress in real time, and store the order information after the payment is successful in the database. The user can view and manage all order information in the personal center. Among them, the above GET / POST requests are the two most commonly used request methods in the HTTP protocol. GET requests are usually used to obtain resources from the server, and POST requests are mainly used to submit data to the server to create or update resources.

[0083] In an embodiment of the present application, after the above-mentioned cultural and tourism assistant application is developed based on the target cultural and tourism model, the above-mentioned travel needs input by the user can be obtained through the cultural and tourism assistant application and the target cultural and tourism model can be called based on the cultural and tourism assistant application. The cultural and tourism information corresponding to the travel needs can be generated through the target cultural and tourism model, and the cultural and tourism information can be pushed to the user.

[0084] Exemplarily, the above-mentioned obtaining the travel needs input by the user through the cultural and travel assistant application and calling the target cultural and travel model, and generating the cultural and travel information corresponding to the travel needs through the target cultural and travel model can be implemented as follows: obtaining the travel needs input by the user through the cultural and travel assistant application; calling the target cultural and travel model through the cultural and travel assistant application, inputting the travel needs into the target cultural and travel model, so that the target cultural and travel model generates travel information based on the travel needs, the user's historical travel data and external data; wherein the external data includes traffic data and weather data. It should be noted that the specific implementation of this process has been described in detail in the corresponding position of the above-mentioned cultural and travel information generation module, so it will not be repeated here.

[0085] Exemplarily, the above-mentioned pushing of cultural and travel information to users can be achieved as follows: generating information cards corresponding to the cultural and travel information through the cultural and travel assistant application, and displaying the information cards in the application page of the cultural and travel assistant application, so that users can obtain cultural and travel information through interactive operations with the information cards. It should be noted that this process has been described in detail in the information integration and display part of the above-mentioned user order management module, so it will not be repeated here.

[0086] In addition, after the information card corresponding to the cultural and travel information is generated by the cultural and travel assistant application and the information card is displayed on the application page of the cultural and travel assistant application, the embodiment of the present application can also receive the user's selection operation on the information card through the cultural and travel assistant application, and jump to the order page corresponding to the selection operation, so that the user can complete the reservation operation corresponding to the cultural and travel information through the interactive operation with the order page. It should be noted that this process has been described in detail in the online booking and payment part of the above-mentioned user order management module, so it will not be repeated here.

[0087] The embodiment of the present application uses the first cultural and tourism data obtained from the target web page, the second cultural and tourism data generated based on the initial cultural and tourism model, and the third cultural and tourism data generated by the sequence generative adversarial network as a training data set for fine-tuning the above initial cultural and tourism model, to obtain a target cultural and tourism model, and develops a cultural and tourism assistant application based on the target cultural and tourism model; the above cultural and tourism assistant application accepts the user's travel needs and calls the target cultural and tourism model, so that the target cultural and tourism model generates a customized travel guide according to the user's travel needs, and returns it to the user through the cultural and tourism assistant application. Based on the above process, the cultural and tourism information push method provided by the embodiment of the present application has the following beneficial effects: (1) Improving tourists’ experience: Through the AI ​​travel assistant based on the target cultural tourism model (i.e. the cultural tourism assistant application mentioned above), tourists can be provided with personalized travel plans, real-time scenic spot recommendations, tour route planning, special dining recommendations and other services. The plans can be dynamically adjusted according to tourists’ needs to meet personalized travel needs, greatly improving tourists’ travel experience; (2) Efficient data collection and processing: Using web crawler technology and large model generation methods, it can automatically collect cultural and tourism related information from multiple tourism platforms and other data sources. Through data cleaning, standardization and deduplication, the accuracy and consistency of the data are ensured, providing a high-quality data source for model training, reducing the time of manual data collection and improving system efficiency; (3) Improved performance of sequence generative adversarial network generator: By improving the traditional SeqGAN network and introducing a multi-head self-attention mechanism, the generator’s ability to model long sequences is enhanced, solving the problem of insufficient global coherence in the long sequence generation process, and improving the quality of the target cultural tourism model when generating tourism-related texts, making the generated travel guides, scenic spot descriptions, and other content more accurate and comprehensive; (4) Intelligent cultural and tourism data generation: Based on network data, a pre-trained model is used to generate cultural and tourism data. Combined with automated evaluation indicators and manual review, this achieves efficient supplementation and expansion of tourism information, making up for the lack of incomplete information coverage when relying solely on network data, ensuring that the system can provide users with more diverse travel advice and guidance; (5) The Cultural and Tourism Assistant application has a high degree of system integration and convenient booking operations: The Cultural and Tourism Assistant application integrates functions such as scenic spot tickets, hotels, and transportation reservations, and optimizes the user's operation process through card-style information display. Users can complete the booking of travel products in one stop without jumping to multiple platforms. By connecting with the API of the third-party platform, the system updates the price and inventory information in real time, ensuring the timeliness and accuracy of the information; (6) Security and privacy protection: The Cultural and Tourism Assistant application generates a unique visitor identity identifier using encryption and hash algorithms through the user information management module, and encrypts and stores visitor information, effectively ensuring user privacy. At the same time, by connecting to a third-party identity authentication interface, the authenticity and legitimacy of the visitor's identity are ensured, thereby improving security; (7) Improved order management function: The Cultural Travel Assistant application tracks the user's booking status in real time through the user order management module. After successful payment, the order information is automatically updated and synchronized to the user's personal center, making it convenient for users to view and manage orders. It also integrates a variety of mainstream payment methods, simplifies the payment process, and improves the convenience of user booking; (8) Improved model fine-tuning efficiency: Through the SFT fine-tuning strategy, the initial cultural and tourism model is quantified and fine-tuned. Combined with high-quality cultural and tourism data sets, the generation effect of the model can be quickly improved, effectively shortening the fine-tuning cycle of the large model and enhancing the practical application ability of the large model. To sum up, the present invention has significant advantages in enhancing tourists' travel experience, improving system efficiency, security and convenience through technological innovation and process optimization, and realizes the efficient application of the target cultural and tourism model in the cultural and tourism assistant application.

[0088] The above mainly introduces the solution provided by the embodiment of the present invention from the perspective of the method. In order to achieve the above functions, it includes hardware structures and / or software modules corresponding to the execution of each function. Those skilled in the art should easily realize that, in combination with the units and algorithm steps of each example described in the embodiments disclosed herein, the present invention can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0089] The embodiment of the present invention can divide the functional modules of the cultural and tourism information push device according to the above method example. For example, each functional module can be divided corresponding to each function, or two or more functions can be integrated into one processing module. The above integrated modules can be implemented in the form of hardware or in the form of software functional modules. It should be noted that the division of modules in the embodiment of the present invention is schematic and is only a logical functional division. There may be other division methods in actual implementation.

[0090] Accordingly, the embodiment of the present disclosure also provides a device for pushing cultural and tourism information. Figure 6 As shown, the cultural tourism information push device 600 may include a data acquisition module 610, a model training module 620 and an information push module 630, wherein: The data acquisition module 610 is used to acquire the first cultural and travel data in the target webpage; The data acquisition module 610 is further used to determine the prompt information of the cultural tourism task, and generate the second cultural tourism data based on the prompt information by using the initial cultural tourism model; wherein the initial cultural tourism model is a pre-trained large language model; The data acquisition module 610 is further used to generate third cultural and tourism data based on the first cultural and tourism data and the second cultural and tourism data using a pre-trained sequence generation adversarial network; A model training module 620 is used to fine-tune the initial cultural tourism model based on the first cultural tourism data, the second cultural tourism data, and the third cultural tourism data to obtain a target cultural tourism model; The information push module 630 is used to obtain the travel needs input by the user through the cultural and travel assistant application and call the target cultural and travel model, generate cultural and travel information corresponding to the travel needs through the target cultural and travel model, and push the cultural and travel information to the user.

[0091] In an embodiment of the present application, the above-mentioned data acquisition module is specifically used to: input the target text sequence consisting of the first cultural and tourism data and the second cultural and tourism data into a sequence generation adversarial network; through the generator of the sequence generation adversarial network, use the multi-head attention mechanism to capture the context information of the current time step of the target text sequence, generate the output sequence probability distribution of the current time step based on the context information of the current time step, and determine the output sequence of the current time step based on the sequence probability distribution; determine the binary classification probability of the output sequence through the discriminator of the sequence generation adversarial network, and determine the third cultural and tourism data based on the binary classification probability.

[0092] In an embodiment of the present application, the above-mentioned model training module is specifically used to: determine a training data set based on first cultural and tourism data, second cultural and tourism data, and third cultural and tourism data; determine fine-tuning parameters, and start the fine-tuning command of the initial cultural and tourism model based on the fine-tuning parameters, and by running the fine-tuning command, fine-tune the initial cultural and tourism model with the fine-tuning parameters based on the training data set to obtain the target cultural and tourism model.

[0093] In an embodiment of the present application, the above-mentioned information push module is specifically used to: obtain travel needs input by the user through the cultural and travel assistant application; call the target cultural and travel model through the cultural and travel assistant application, and input the travel needs into the target cultural and travel model, so that the target cultural and travel model generates travel information based on travel needs, the user's historical travel data and external data; wherein the external data includes traffic data and weather data.

[0094] In an embodiment of the present application, the above-mentioned information push module is specifically used to: generate information cards corresponding to cultural and tourism information through a cultural and tourism assistant application, and display the information cards on the application page of the cultural and tourism assistant application, so that users can obtain cultural and tourism information through interactive operations with the information cards.

[0095] In an embodiment of the present application, the above-mentioned information push module is also used to: receive the user's selection operation on the information card through the cultural and tourism assistant application, and jump to the order page corresponding to the selection operation, so that the user can complete the reservation operation corresponding to the cultural and tourism information through interactive operations with the order page.

[0096] In the embodiment of the present application, the above-mentioned cultural and travel assistant application includes a user information management module, a cultural and travel information generation module and a user order management module. Among them: the above-mentioned user information management module is used to receive user information input by the user, and register, encrypt and verify the user information; the above-mentioned cultural and travel information generation module is used to call the target cultural and travel model, generate cultural and travel information corresponding to travel needs through the target cultural and travel model, and push the cultural and travel information to the user; the above-mentioned user order management module is used to respond to the user's selection operation on the cultural and travel information, jump to the order page corresponding to the selection operation, so as to complete the corresponding user order based on the interaction between the user and the order page, and provide storage and management operations for user orders.

[0097] The specific implementation details of the above-mentioned cultural and tourism information push device have been described in detail in the corresponding position of the cultural and tourism information push method, so they will not be repeated here.

[0098] In the embodiment of the present application, the above-mentioned cultural and tourism information push device greatly enhances the process customization and automation processing capabilities while ensuring the basic task management function. Not only is the standard task control process preset, but it also allows users to customize the task process, automatic test process and automatic calibration process according to specific needs, realizing modular construction from the minimum control instruction to the realization of complex functions. By carefully combining control commands, setting preconditions and decision branches, it is possible to create a command sequence that adapts to various measurement and control scenarios to meet the execution, testing and calibration + requirements of different tasks. In addition, through the intelligent execution mechanism driven by the above-mentioned process scheduler, it is ensured that all customized processes and measurement and control tasks can be accurately started according to the time sequence plan, and the relevant subsystems and equipment resources are automatically coordinated. Based on the above settings, the cultural and tourism information push device provided in the embodiment of the present application not only effectively tracks and receives satellite signals, completes data transmission tasks, but also realizes real-time task status monitoring, provides a powerful and stable platform for satellite remote monitoring and control, and relies on its excellent scalability, flexible process scheduling capabilities and automation level, significantly improves the efficiency and reliability of satellite measurement and control, and reduces the complexity and cost of operation.

[0099] Figure 7 Schematic diagram of the structure of an electronic device in the embodiment of the present disclosure. Figure 7 , which shows a structural schematic diagram of an electronic device 700 suitable for implementing the embodiments of the present disclosure. Figure 7 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0100] like Figure 7As shown, the electronic device 700 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 701, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 702 or a program loaded from a storage device 708 to a random access memory (RAM) 703 to implement the cultural travel information push method of the embodiment described in the present disclosure. In RAM703, various programs and data required for the operation of the electronic device 700 are also stored. The processing device 701, ROM702, and RAM703 are connected to each other via a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0101] Typically, the following devices may be connected to the I / O interface 705: an input device 706 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 707 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 708 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 709. The communication device 709 may allow the electronic device 700 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 7 The electronic device 700 is shown with various devices, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0102] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program contains a program code for executing the method shown in the flowchart, thereby implementing the cultural and travel information push method as described above. In such an embodiment, the computer program can be downloaded and installed from the network through the communication device 709, or installed from the storage device 708, or installed from the ROM 702. When the computer program is executed by the processing device 701, the above-mentioned functions defined in the method of the embodiment of the present disclosure are executed.

[0103] It should be noted that the computer-readable medium mentioned above in the present disclosure may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. Computer readable signal media may also be any computer readable medium other than computer readable storage media, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0104] In some embodiments, the client and the server may communicate using any currently known or future developed network protocol such as HTTP (Hyper Text Transfer Protocol), and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), an internet (e.g., the Internet), and a peer-to-peer network (e.g., an ad hoc peer-to-peer network), as well as any currently known or future developed network.

[0105] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0106] The computer-readable medium carries one or more programs. When the one or more programs are executed by the electronic device, the electronic device: Get the First Culture and Tourism data from the target web page; Determine prompt information for the cultural and tourism task, and use the initial cultural and tourism model to generate second cultural and tourism data based on the prompt information; wherein the initial cultural and tourism model is a pre-trained large language model; Generate third cultural and tourism data based on the first cultural and tourism data and the second cultural and tourism data using a pre-trained sequence generation adversarial network; Fine-tune the initial cultural tourism model based on the first cultural tourism data, the second cultural tourism data, and the third cultural tourism data to obtain a target cultural tourism model; The travel needs input by the user are obtained through the cultural and tourism assistant application and the target cultural and tourism model is called. The cultural and tourism information corresponding to the travel needs is generated through the target cultural and tourism model and pushed to the user.

[0107] Optionally, when the above one or more programs are executed by the electronic device, the electronic device may also execute other steps described in the above embodiments.

[0108] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages ​​or a combination thereof, including, but not limited to, object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0109] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0110] The units involved in the embodiments described in the present disclosure may be implemented by software or hardware, wherein the name of a unit does not, in some cases, limit the unit itself.

[0111] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0112] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0113] The above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present disclosure is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present disclosure (but not limited to) by each other to form a technical solution.

[0114] In addition, although each operation is described in a specific order, this should not be understood as requiring these operations to be performed in the specific order shown or in a sequential order. Under certain circumstances, multitasking and parallel processing may be advantageous. Similarly, although some specific implementation details are included in the above discussion, these should not be interpreted as limiting the scope of the present disclosure. Some features described in the context of a separate embodiment can also be implemented in a single embodiment in combination. On the contrary, the various features described in the context of a single embodiment can also be implemented in multiple embodiments individually or in any suitable sub-combination mode.

[0115] Although the subject matter has been described in language specific to structural features and / or methodological logical actions, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. On the contrary, the specific features and actions described above are merely example forms of implementing the claims.

Claims

1. A method for pushing cultural and tourism information, characterized in that: The method comprises: Get the First Culture and Tourism data from the target web page; Determine prompt information of the cultural and tourism task, and generate second cultural and tourism data based on the prompt information using an initial cultural and tourism model; wherein the initial cultural and tourism model is a pre-trained large language model; Generate third cultural and tourism data based on the first cultural and tourism data and the second cultural and tourism data using a pre-trained sequence generation adversarial network; Fine-tune the initial cultural tourism model based on the first cultural tourism data, the second cultural tourism data, and the third cultural tourism data to obtain a target cultural tourism model; The travel needs input by the user are obtained through the cultural and travel assistant application and the target cultural and travel model is called. The cultural and travel information corresponding to the travel needs is generated through the target cultural and travel model, and the cultural and travel information is pushed to the user.

2. The method for pushing cultural and tourism information according to claim 1, characterized in that: The method of using a pre-trained sequence generation adversarial network to generate third cultural and tourism data based on the first cultural and tourism data and the second cultural and tourism data includes: Inputting a target text sequence consisting of the first cultural and tourism data and the second cultural and tourism data into the sequence generation adversarial network; Capturing context information of the current time step of the target text sequence by using a multi-head attention mechanism through a generator of the sequence generative adversarial network, generating an output sequence probability distribution of the current time step based on the context information of the current time step, and determining an output sequence of the current time step based on the sequence probability distribution; The binary classification probability of the output sequence is determined by the discriminator of the sequence generative adversarial network, and the third cultural and tourism data is determined based on the binary classification probability.

3. The method for pushing cultural and tourism information according to claim 1, characterized in that: The fine-tuning of the initial cultural tourism model based on the first cultural tourism data, the second cultural tourism data, and the third cultural tourism data to obtain a target cultural tourism model includes: Determine a training data set based on the first cultural and tourism data, the second cultural and tourism data, and the third cultural and tourism data; Determine fine-tuning parameters, start the fine-tuning command of the initial cultural tourism model, and fine-tune the initial cultural tourism model with the fine-tuning parameters based on the training data set by running the fine-tuning command to obtain the target cultural tourism model.

4. The method for pushing cultural and tourism information according to claim 1, characterized in that: The step of obtaining the travel demand input by the user through the cultural and travel assistant application and calling the target cultural and travel model, and generating cultural and travel information corresponding to the travel demand through the target cultural and travel model, includes: Acquire the travel demand input by the user through the cultural and travel assistant application; The target cultural and tourism model is called through the cultural and tourism assistant application, and the travel demand is input into the target cultural and tourism model, so that the target cultural and tourism model generates the cultural and tourism information based on the travel demand, the user's historical travel data and external data; wherein the external data includes traffic data and weather data.

5. The method for pushing cultural and tourism information according to claim 1, characterized in that: The pushing the cultural travel information to the user includes: An information card corresponding to the cultural and travel information is generated through the cultural and travel assistant application, and the information card is displayed on the application page of the cultural and travel assistant application, so that the user can obtain the cultural and travel information through interactive operations with the information card.

6. The method for pushing cultural and tourism information according to claim 5, characterized in that: After generating an information card corresponding to the cultural and travel information by the cultural and travel assistant application and displaying the information card on an application page of the cultural and travel assistant application, the method further includes: The cultural and travel assistant application receives the user's selection operation on the information card and jumps to the order page corresponding to the selection operation, so that the user can complete the reservation operation corresponding to the cultural and travel information through interactive operations with the order page.

7. A device for pushing cultural and tourism information, characterized in that: The device comprises: A data acquisition module, used to acquire the first cultural and tourism data in the target web page; The data acquisition module is further used to determine prompt information of the cultural and tourism task, and generate second cultural and tourism data based on the prompt information using the initial cultural and tourism model; wherein the initial cultural and tourism model is a pre-trained large language model; The data acquisition module is further used to generate third cultural and tourism data based on the first cultural and tourism data and the second cultural and tourism data using a pre-trained sequence generation adversarial network; A model training module, used for fine-tuning the initial cultural tourism model based on the first cultural tourism data, the second cultural tourism data and the third cultural tourism data to obtain a target cultural tourism model; The information push module is used to obtain the travel needs input by the user through the cultural and travel assistant application and call the target cultural and travel model, generate cultural and travel information corresponding to the travel needs through the target cultural and travel model, and push the cultural and travel information to the user.

8. The cultural and tourism information push device according to claim 7, characterized in that: The cultural and travel assistant application includes a user information management module, a cultural and travel information generation module and a user order management module; wherein: The user information management module is used to receive user information and register, encrypt and verify the user information; The cultural tourism information generation module is used to call the target cultural tourism model, generate cultural tourism information corresponding to the travel demand through the target cultural tourism model, and push the cultural tourism information to the user; The user order management module is used to respond to the user's selection operation on the cultural and travel information, jump to the order page corresponding to the selection operation, complete the corresponding user order based on the user's interactive operation with the order page, and provide storage and management operations for the user order.

9. An electronic device, comprising: processor; and a memory storing a program, wherein the program comprises instructions, and the instructions are used to cause the processor to execute the method according to any one of claims 1-6.

10. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-6.

Images