Interaction method, apparatus and system for intelligent devices based on GPT model

Abstract

To provide a GPT(R) model-based intelligent device interaction method, apparatus, and system.SOLUTION: A method includes causing an intelligent device side to obtain user interaction information by masking initial interaction information, transmit it to a proxy layer together with user preference information, process the user interaction information through a query proxy, and transmit the information to a GPT model. The method further includes causing the model to perform intent recognition and task subdivision based on the processed interaction information, and return task processing information of the processed interaction information, and also causes a response proxy to accept corresponding information, obtain interaction response information based on the user preference information and the task processing information, and transmit it to the intelligent device side. The method further includes causing the intelligent device side to modify the interaction response information based on user privacy data, and respond based on the modified information.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to the technical field of human-computer interaction, and in particular to an intelligent device interaction method, apparatus and system based on the GPT model. [Background technology]

[0002] In modern society, intelligent devices have become deeply ingrained in people's daily lives. Human-computer interaction functions, especially for devices like smart speakers and in-car devices, are becoming increasingly important. However, current intelligent device interaction systems still face many challenges. First, traditional interaction methods cannot meet people's needs for more natural and intelligent interaction. Second, due to device hardware limitations, these devices are unable to produce high-quality digital humans (i.e., virtual humans). Most importantly, large-scale language models based on the Generative Pre-Trained Transformer (GPT) model deployed on the cloud can provide personalized digital human services, but they also pose data security and privacy issues. Therefore, how to provide more accurate interaction effects while ensuring data security and protecting user privacy during the real-time upload and analysis of user data is a key challenge that must be addressed. Summary of the Invention [Problem to be solved by the invention]

[0003] The present invention provides a method, apparatus and system for intelligent device interaction based on the GPT model to solve the shortcomings in the prior art, such as the contradiction between the interaction effect and the system complexity of intelligent devices, and the problem of user privacy leakage when providing personalized services. [Means for solving the problem]

[0004] The present invention provides an intelligent device interaction method based on the GPT model, which includes:

[0005] That is, the intelligent device receives user interaction information and user preference information transmitted from the intelligent device. The user interaction information is information obtained after the intelligent device has masked the initial interaction information input by the user.

[0006] After processing the user interaction information to obtain post-processing interaction information, the post-processing interaction information is sent to a GPT model, which returns task processing information of the post-processing interaction information.

[0007] After receiving the task processing information returned from the GPT model, the interaction response information is obtained by fusing the user preference information and the task processing information, and the interaction response information is sent to the intelligent device, whereby the intelligent device modifies the interaction response information based on the user privacy data related to the user interaction information, and responds to the user based on the modified information.

[0008] According to the intelligent device interaction method based on the GPT model provided in the present invention, when obtaining interaction response information by integrating the above-mentioned user preference information and the task processing information, the following steps are specifically included:

[0009] That is, by performing matching based on each noun in the task processing information and each preference label in the user preference information, preference labels that match each noun in the task processing information are obtained.

[0010] The intelligent device acquires interaction response information by replacing the corresponding noun in the task processing information based on the label value of the preference label matched to each noun in the task processing information. After receiving the interaction response information, the intelligent device acquires correction processing information by correcting the interaction response information based on the user privacy data related to the user interaction information, and executes the correction processing information to realize a response to the user.

[0011] According to the intelligent device interaction method based on the GPT model provided in the present invention, when sending the above-mentioned post-processing interaction information to the GPT model, it specifically includes:

[0012] That is, the post-processing interaction information is vectorized to obtain a post-processing interaction vector.

[0013] A search is performed in the knowledge base of each field based on the post-processing interaction vector, thereby identifying the field knowledge and the field knowledge vector that match the post-processing interaction vector in the knowledge base of each field.

[0014] The processed interaction vector is combined with the domain knowledge vector of the matched domain knowledge to obtain a fused query vector, and the fused query vector is sent to the GPT model.

[0015] According to the intelligent device interaction method based on the GPT model provided in the present invention, when sending the above-mentioned fused query vector to the GPT model, it specifically includes:

[0016] That is, by sending the fused query vector to multiple GPT models, task processing information of the post-processing interaction information returned from the multiple GPT models is obtained.

[0017] The post-processing interaction information returned from the plurality of GPT models is fused or selected based on task processing information to obtain optimized processing information, which is then fused with the user preference information to obtain the interaction response information.

[0018] According to the intelligent device interaction method based on the GPT model provided in the present invention, the proxy layer is constructed according to the following steps:

[0019] That is, an initial model is constructed based on a pre-trained text generation model.

[0020] Sample input information is input to the initial model, and test output information is output from the initial model.

[0021] The proxy layer is obtained by fine-tuning the initial model based on an evaluation of the test output information or based on differences between the test output information and sample output information.

[0022] The present invention further provides an intelligent device interaction method based on the GPT model, which includes:

[0023] That is, the initial interaction information input by the user is accepted, and the initial interaction information is subjected to a masking process to obtain user interaction information.

[0024] The user interaction information and the user preference information of the user are sent to a proxy layer, whereby the proxy layer processes the user interaction information to obtain post-processing interaction information, and then sends the post-processing interaction information to a GPT model, so that the GPT model returns task processing information of the post-processing interaction information.

[0025] The interaction response information sent by the proxy layer is received, the interaction response information is modified based on the user privacy data related to the user interaction information, and a response is made to the user based on the modified information. The interaction response information is obtained by fusing the user preference information and the task processing information after the proxy layer receives the task processing information of the processed interaction information returned from the GPT model.

[0026] According to the intelligent device interaction method based on the GPT model provided in the present invention, the interaction response information is modified according to the user privacy data related to the above-mentioned user interaction information, and when responding to the user based on the modified information, the following specifically is included:

[0027] That is, based on the user privacy data related to the user interaction information, the interaction response information is corrected to obtain correction process information, and the correction process information is executed to realize a response to the user.

[0028] The interaction response information is obtained by the proxy layer obtaining preference labels that match each noun in the task processing information by matching each noun in the task processing information with each preference label in the user preference information, and then replacing the corresponding noun in the task processing information based on the label value of the preference label matched to each noun in the task processing information.

[0029] According to the intelligent device interaction method based on the GPT model provided in the present invention, the method further includes:

[0030] That is, the initial interaction information and the corresponding revised information of multiple users are collected, and the initial interaction information corresponding to the revised information that does not meet the user's expectations is recorded as a directional optimization sample.

[0031] Intent recognition is performed on the directional optimization sample to obtain a user intent corresponding to the directional optimization sample, and a type label is set for relevant text in the directional optimization sample that is related to the user intent based on the user intent.

[0032] The related text is segmented into words to obtain keywords of the related text, and a search is performed within a knowledge base of each field based on the keywords of the related text and the type label of the related text to obtain field knowledge that matches the related text.

[0033] A relevance relationship is established between the related text and the domain knowledge matched with the related text, so that when the same user again inputs the same initial interaction information as the directional optimization sample and a search is performed in the knowledge base of each domain based on the post-processing interaction vector of the post-processing interaction information corresponding to the corresponding initial interaction information, the domain knowledge matched with the related text is identified as the domain knowledge matching the corresponding post-processing interaction vector, or a response is made to the initial interaction information based on the domain knowledge matched with the related text.

[0034] According to the intelligent device interaction method based on the GPT model provided in the present invention, the method further includes:

[0035] That is, the feedback information input by the user after the response is obtained, and an image of the user inputting the feedback information is collected.

[0036] Emotion recognition is performed on the feedback information to obtain an emotion type of the feedback information.

[0037] The facial expression type when the user inputs the feedback information is acquired by performing facial expression recognition on an image when the user inputs the feedback information.

[0038] If the feedback information contains a negative keyword, if the emotional type of the feedback information is a negative emotional type, or if the facial expression type when the user inputs the feedback information is a negative facial expression, it is determined that an interaction barrier exists at that time.

[0039] If an interaction barrier exists at that time, each piece of question and answer history information corresponding to the user is acquired, and question and answer history information that matches the initial interaction information is searched for from each piece of question and answer history information, and a response is made based on the question and answer history information that matches the initial interaction information.

[0040] According to the intelligent device interaction method based on the GPT model provided in the present invention, the method further includes:

[0041] That is, the type of network condition at that time is checked, and when the type of network condition is a network condition deterioration or a network disconnection state, the user interaction information and the user preference information of the user are sent to a local GPT model deployed locally, thereby obtaining local response information returned from the local GPT model, and responding to the user based on the local response information.

[0042] According to the GPT model-based intelligent device interaction method provided in the present invention, the local GPT model is obtained by performing model quantization compression on the sample GPT model.

[0043] The present invention further provides an intelligent device interaction method based on the GPT model, which includes:

[0044] That is, the intelligent device side receives the initial interaction information input by the user and obtains the user interaction information by masking the initial interaction information.

[0045] The intelligent device side sends the user interaction information and the user preference information of the user to a proxy layer.

[0046] The proxy layer receives user interaction information and user preference information sent from the intelligent device side, processes the user interaction information to obtain post-processing interaction information, and then sends the post-processing interaction information to the GPT model, which then returns task processing information of the post-processing interaction information.

[0047] The proxy layer receives the task processing information of the post-processing interaction information returned from the GPT model, and then obtains interaction response information by fusing it based on the user preference information and the task processing information, and transmits the interaction response information to the intelligent device side.

[0048] The intelligent device side receives the interaction response information sent by the proxy layer, modifies the interaction response information based on the user privacy data related to the user interaction information, and responds to the user based on the modified information.

[0049] The present invention further provides an intelligent device interaction device based on the GPT model, which includes:

[0050] The information receiving unit receives user interaction information and user preference information sent from the intelligent device. The user interaction information is the initial interaction information input by the user that has been masked by the intelligent device.

[0051] Inquiry Proxy: After processing the user interaction information to obtain post-processing interaction information, it sends the post-processing interaction information to the GPT model, which then returns the task processing information of the post-processing interaction information.

[0052] Response Proxy: After receiving the task processing information returned from the GPT model, it obtains interaction response information by fusing it based on the user preference information and the task processing information, and sends the interaction response information to the intelligent device side, so that the intelligent device side modifies the interaction response information based on the user privacy data related to the user interaction information, and responds to the user based on the modified information.

[0053] The present invention further provides an intelligent device interaction device based on the GPT model, which includes:

[0054] An input processing unit receives initial interaction information input by a user and performs a masking process on the initial interaction information to obtain user interaction information.

[0055] An information forwarding unit sends the user interaction information and the user preference information of the user to a proxy layer, so that the proxy layer processes the user interaction information to obtain post-processing interaction information, and then sends the post-processing interaction information to a GPT model, so that the GPT model returns task processing information of the post-processing interaction information.

[0056] An information response unit receives the interaction response information sent by the proxy layer, modifies the interaction response information based on the user privacy data related to the user interaction information, and responds to the user based on the modified information. The interaction response information is obtained by fusing the user preference information and the task processing information after the proxy layer receives the task processing information of the processed interaction information returned from the GPT model.

[0057] The present invention further provides an intelligent device interaction system based on the GPT model, which includes a proxy layer and an intelligent device side.

[0058] In the proxy layer, an interaction device of an intelligent device based on the GPT model corresponding to any of the above-mentioned proxy layers is deployed.

[0059] On the intelligent device side, an intelligent device interaction device based on the GPT model corresponding to any of the above-mentioned intelligent device sides is deployed. [Effects of the Invention]

[0060] In the intelligent device interaction method, apparatus, and system based on the GPT model provided by the present invention, the intelligent device masks initial interaction information input by the user to obtain user interaction information, which is then sent to the proxy layer together with user preference information. The query proxy in the proxy layer processes the user interaction information to obtain post-processing interaction information, which is then sent to the GPT model. The GPT model then performs intent recognition and task segmentation based on the post-processing interaction information and returns task processing information in the post-processing interaction information. The response proxy receives the task processing information in the post-processing interaction information returned from the GPT model, combines it with the user preference information and the task processing information to obtain interaction response information for the user interaction information, and sends the interaction response information to the intelligent device. The intelligent device then modifies the interaction response information based on user privacy data related to the user interaction information and responds to the user based on the modified information. The above provides a solution for realizing high-quality, personalized intelligent interactions while protecting user privacy, thereby improving the accuracy of responses and ensuring that user privacy data remains only on the intelligent device, thereby protecting the safety of user privacy.

[0061] In order to more clearly describe the technical solutions of the present invention or the prior art, the following briefly describes the drawings that need to be used in describing the embodiments or the prior art, it goes without saying that the drawings described below are only some of the embodiments of the present invention, and those skilled in the art can obtain other drawings from these drawings without any creative work. [Brief explanation of the drawings]

[0062] [Figure 1]FIG. 1 is a schematic flowchart 1 of the intelligent device interaction method based on the GPT model provided in the present invention. [Figure 2] FIG. 2 is a schematic flowchart 2 of the intelligent device interaction method based on the GPT model provided in the present invention. [Figure 3] FIG. 3 is a schematic flowchart 3 of the intelligent device interaction method based on the GPT model provided in the present invention. [Figure 4] FIG. 4 is a schematic structural diagram 1 of an intelligent device interaction device based on the GPT model provided in the present invention. [Figure 5] FIG. 5 is a schematic structural diagram 2 of an intelligent device interaction device based on the GPT model provided in the present invention. [Figure 6] FIG. 6 is a schematic structural diagram of an intelligent device interaction system based on the GPT model provided in the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0063] In order to make the objectives, technical solutions and advantages of the present invention clearer, the following will clearly and concisely describe the technical solutions of the present invention in combination with the drawings in the present invention. It goes without saying that the embodiments described here are only some of the embodiments of the present invention, and not all of the embodiments. All other embodiments that can be obtained by those skilled in the art based on the embodiments of the present invention without requiring creative work are all within the scope of protection of the present invention.

[0064] FIG. 1 is a schematic flowchart 1 of an intelligent device interaction method based on the GPT model provided by the present invention. The method is applied to a proxy layer deployed in the cloud, which is interposed between the intelligent device side used for interaction and the GPT model deployed in the cloud. The intelligent device side includes, but is not limited to, a smart speaker, an in-vehicle device, etc. As shown in FIG. 1, the method includes:

[0065] Step 110: Accept user interaction information and user preference information sent from the intelligent device side. The user interaction information is information obtained after the intelligent device side has masked the initial interaction information input by the user.

[0066] Step 120: After processing the user interaction information to obtain post-processing interaction information, send the post-processing interaction information to a GPT model, so that the GPT model returns task processing information of the post-processing interaction information.

[0067] Step 130: After receiving the task processing information returned from the GPT model, obtain interaction response information by fusing the user preference information and the task processing information, and send the interaction response information to the intelligent device side, so that the intelligent device side modifies the interaction response information based on the user privacy data related to the user interaction information, and responds to the user based on the modified information.

[0068] Specifically, after a user inputs initial interaction information through a human-computer interface (e.g., a voice input interface or a text input interface) provided by the intelligent device, the intelligent device masks the initial interaction information and replaces data related to the user's privacy (e.g., name, contact information, specific address, etc.) with a predetermined symbol (e.g., "*") to obtain user interaction information corresponding to the initial interaction information. The intelligent device then obtains user preference information of the user. The user preference information is obtained by performing a user profile analysis on the user's various interaction behaviors on the intelligent device (which may be local interaction behaviors on the intelligent device or interaction behaviors synchronized in a mobile application installed on the intelligent device and linked to the user's mobile phone). The intelligent device then sends the user preference information and the user interaction information together to a proxy layer in the cloud. The proxy layer in the cloud includes at least a query proxy and a response proxy.

[0069] The query proxy can receive user interaction information and user preference information from the intelligent device, process the user interaction information, and obtain post-processing interaction information. The query proxy can predefine processing patterns, and organize the user interaction information and user preference information according to the processing patterns to obtain post-processing interaction information. This allows the post-processing interaction information to be sent to the GPT model in the cloud, where the GPT model can perform intent recognition and task segmentation based on the post-processing interaction information, rather than simply feeding back the user interaction information. The processing pattern can be obtained by testing specific types of GPT models. For example, different prompts can be manually configured and then tested one by one to obtain prompts that can achieve the corresponding interaction needs (i.e., obtain a task processing method), but this is not a specific limitation in the embodiments of the present invention. For example, if the user interaction information is command control-related, the GPT model may perform intent recognition and task segmentation based on the post-processing interaction information, and then return task processing information capable of instructing the execution steps of the corresponding command. For example, if the command represented by the user interaction information is to order takeout lunch, the GPT model may return task processing information instructing processing steps such as selecting a takeout platform, selecting food to order, and determining contact information and an address, rather than simply executing the command "order takeout." Furthermore, if the user interaction information is information suggestion-related, the GPT model may perform intent recognition and task segmentation based on the post-processing interaction information, and then return task processing information capable of instructing the acquisition steps of the corresponding information. For example, if the user interaction information indicates that the waiting information for suggestions is restaurants near the user's home, the GPT model may return task processing information instructing processing steps such as selecting a dining area and a restaurant style based on the user's home address, rather than directly returning the selected restaurants.As is clear from the above, the task processing information returned by the GPT model indicates how to process the user interaction information.

[0070] After receiving the task processing information of the processed interaction information returned by the GPT model, the response proxy obtains interaction response information that better meets the user's needs and preferences through fusion based on the user preference information and the task processing information, so as to improve the user's interaction experience, and sends the interaction response information to the intelligent device side. The intelligent device side modifies the interaction response information based on the user privacy data related to the user interaction information, and responds to the user based on the modified information. This improves the accuracy of the response while ensuring that the user privacy data remains only on the intelligent device side, thereby protecting the user's privacy.

[0071] In some embodiments, when the user interaction information is command-controlled and the response proxy acquires interaction response information for the user interaction information by combining the user preference information and the task process information, a preference label matching each noun in the task process information may be acquired by matching each noun in the task process information with each preference label in the user preference information. Here, if the similarity between a noun and all preference labels is lower than a predetermined threshold, no preference label matches the noun. Furthermore, if the similarity between a noun and a preference label is not only higher than the similarity between all other preference labels but also higher than a predetermined threshold, the noun is matched to the preference label. Then, the interaction response information for the user interaction information is acquired by replacing the corresponding noun in the task process information based on the label value of the preference label matched to each noun in the task process information. After receiving the interaction response information, the intelligent device side can supplement and complete the corresponding part of the interaction response information based on the user privacy data related to the user interaction information (e.g., address information, contact information, etc.), obtain correction processing information, and call a local execution interpreter to execute the correction processing information, thereby realizing a response to the user. Note that the local GPT on the intelligent device side can be used to analyze the interaction response information and the user privacy data to identify the part in the interaction response information to be supplemented, and the user privacy data can be used to supplement it.

[0072] In some other embodiments, when the user interaction information is an information suggestion system and the response proxy acquires interaction response information for the user interaction information by fusing the user preference information and task process information, the response proxy may acquire preference labels matching each noun in the task process information by matching each noun in the task process information with each preference label in the user preference information. Then, the complete task process information is acquired by replacing the corresponding noun in the task process information based on the label value of the preference label matched to each noun in the task process information. Next, the interaction response information for the user interaction information is acquired by directly executing the steps indicated in the complete task process information. After receiving the interaction response information, the intelligent device may select the interaction response information based on user privacy data related to the user interaction information (e.g., select restaurants within a narrower area or range based on the user's home address), acquire the selected response information, and return the selected response information to the user, thereby responding to the user.

[0073] According to any of the above embodiments, when the query proxy sends post-processing interaction information to the GPT model, the query proxy may vectorize the post-processing interaction information to obtain a post-processing interaction vector, and then perform a search within each domain's knowledge base based on the post-processing interaction vector. Each domain's knowledge base stores domain knowledge for each domain and domain knowledge vectors for the domain knowledge. Through vector search, it is possible to identify domain knowledge and its domain knowledge vector that match the post-processing interaction vector in each domain's knowledge base. The post-processing interaction vector is then combined with the domain knowledge vector of the matched domain knowledge to obtain a fused query vector, and the fused query vector is sent to the GPT model. Combining and fusing the post-processing interaction vector of the post-processing interaction information with the domain knowledge vector of the matched domain knowledge and then sending it to the GPT model helps the GPT model understand the true meaning of the post-processing interaction information, thereby improving the accuracy of interaction responses.

[0074] Here, by transmitting the fusion query vector to multiple GPT models, task processing information in the post-processing interaction information returned from the multiple GPT models can be obtained, and optimal task processing information can be selected from the task processing information in the post-processing interaction information returned from the multiple GPT models based on the task processing information in the post-processing interaction information returned from the multiple GPT models. Alternatively, the task processing information returned from the multiple GPT models can be fused to obtain optimized processing information. The optimized processing information is fused with user preference information to obtain interaction response information in the user interaction information. That is, when obtaining interaction response information by fusion based on user preference information and task processing information, the response proxy can select from the task processing information returned from the multiple GPT models, or can obtain interaction response information by fusing the optimized processing information obtained by fusion with the user preference information.

[0075] Based on any of the above embodiments, to improve the data processing capabilities of the query proxies and response proxies of the proxy layer, reinforcement learning may be performed on the proxy layer (query proxies and response proxies) based on the following steps.

[0076] First, to utilize the text understanding and text generation capabilities of a conventional text generation model (e.g., a GPT model), initial models corresponding to the query proxy and response proxy of the proxy layer are constructed based on a pre-trained text generation model. Note that the same or different text generation models may be used for the query proxy and response proxy, and this is not specifically limited in the embodiments of the present invention. Next, the collected sample input information is input into the initial model to obtain test output information output from the initial model. It should be noted that the sample input information corresponding to the query proxy and the response proxy is different. For the query proxy, the corresponding sample input information may be user interaction information and processing patterns. For the response proxy, the corresponding sample input information may be user preference information and task processing information. The test output information output from the initial model is evaluated, and then a model loss is calculated based on the evaluation of the test output information. The model parameters of the initial model are then reverse-adjusted based on the model loss to fine-tune the initial model, thereby obtaining a trained query proxy or response proxy. Alternatively, accurate sample output information may be set for the sample input information, and the difference between the sample output information and the test output information may be calculated based on this difference. Then, the model parameters of the initial model are back-tuned based on the model loss to achieve fine-tuning of the initial model, thereby obtaining a trained query proxy or response proxy.

[0077] In a method provided in an embodiment of the present invention, the intelligent device masks initial interaction information input by a user to obtain user interaction information, which is then sent to the proxy layer together with user preference information. The query proxy in the proxy layer then processes the user interaction information to obtain post-processing interaction information, which is then sent to the GPT model. The GPT model then performs intent recognition and task segmentation based on the post-processing interaction information and returns task processing information in the post-processing interaction information. The response proxy receives the task processing information in the post-processing interaction information returned from the GPT model, combines it with the user preference information and the task processing information to obtain interaction response information for the user interaction information, and sends the interaction response information to the intelligent device. The intelligent device then modifies the interaction response information based on user privacy data related to the user interaction information and responds to the user based on the modified information. This improves the accuracy of responses while ensuring that user privacy data remains only on the intelligent device, thereby protecting the user's privacy.

[0078] 2 is a schematic flowchart 2 of the intelligent device interaction method based on the GPT model provided in the present invention, which is applied to the intelligent device side. As shown in FIG. 2, the method includes:

[0079] Step 210: Accept initial interaction information input by a user, and obtain user interaction information by performing a masking process on the initial interaction information.

[0080] Step 220: Send the user interaction information and the user preference information of the user to a proxy layer, so that the proxy layer processes the user interaction information to obtain post-processing interaction information, and then sends the post-processing interaction information to a GPT model, so that the GPT model returns task processing information of the post-processing interaction information.

[0081] Step 230: Accept the interaction response information sent by the proxy layer, modify the interaction response information based on the user privacy data related to the user interaction information, and respond to the user based on the modified information. The interaction response information is obtained by combining the user preference information and the task processing information after the proxy layer accepts the task processing information of the processed interaction information returned from the GPT model.

[0082] In a method provided in an embodiment of the present invention, the intelligent device masks initial interaction information input by a user to obtain user interaction information, which is then sent to the proxy layer together with user preference information. The query proxy in the proxy layer then processes the user interaction information to obtain post-processing interaction information, which is then sent to the GPT model. The GPT model then performs intent recognition and task segmentation based on the post-processing interaction information and returns task processing information in the post-processing interaction information. The response proxy receives the task processing information in the post-processing interaction information returned from the GPT model, combines it with the user preference information and the task processing information to obtain interaction response information for the user interaction information, and sends the interaction response information to the intelligent device. The intelligent device then modifies the interaction response information based on user privacy data related to the user interaction information and responds to the user based on the modified information. This improves the accuracy of responses while ensuring that user privacy data remains only on the intelligent device, thereby protecting the user's privacy.

[0083] Based on any of the above embodiments, if the user interaction information is a command control system, modifying the interaction response information based on the user privacy data related to the above user interaction information, and responding to the user based on the modified information specifically includes:

[0084] That is, based on the user privacy data related to the user interaction information, the interaction response information is corrected to obtain correction process information, and the correction process information is executed to realize a response to the user.

[0085] The interaction response information is obtained by the proxy layer obtaining preference labels that match each noun in the task processing information by matching each noun in the task processing information with each preference label in the user preference information, and then replacing the corresponding noun in the task processing information based on the label value of the preference label matched to each noun in the task processing information.

[0086] Based on any of the above embodiments, if the user interaction information is an information suggestion type, modifying the interaction response information based on the user privacy data related to the above user interaction information, and responding to the user based on the modified information specifically includes:

[0087] That is, the interaction response information is selected based on the user privacy data related to the user interaction information, selected response information is acquired, and the selected response information is returned to the user, thereby realizing a response to the user.

[0088] The interaction response information is obtained by the proxy layer obtaining preference labels that match each noun in the task processing information by matching each noun in the task processing information with each preference label in the user preference information, then obtaining complete task processing information by replacing the corresponding noun in the task processing information based on the label value of the preference label matched to each noun in the task processing information, and then executing the complete task processing information.

[0089] It should be mentioned that the interaction method presented in the above embodiment on the intelligent device side is similar to that presented in the embodiment corresponding to the proxy layer, so it will not be further detailed here.

[0090] In any of the above embodiments, if a user inputs speech but the speech is misrecognized, for example, if the Kongo River is misrecognized as Mount Kongo, it may be difficult to provide an accurate response. In particular, when the query proxy in the proxy layer searches within each domain's knowledge base based on the post-processing interaction vector of the post-processing interaction information, it becomes difficult to find truly relevant domain knowledge, making it difficult to guarantee the accuracy of the response. To further improve the accuracy of the interaction response, the intelligent device may collect multiple users' initial interaction information and corresponding revised information (i.e., the result of the intelligent device revising the interaction response information according to the corresponding initial interaction information based on user privacy data related to the user interaction information), and record the initial interaction information corresponding to the revised information that does not meet the user's expectations as a directional optimization sample. Here, the intelligent device may determine whether the revised information meets the user's expectations based on the corresponding user's reaction (e.g., the user's facial expression or feedback information input by the user) after responding based on the revised information.

[0091] Then, intent recognition is performed on each directional optimization sample to obtain the user intent corresponding to each directional optimization sample. Then, for each directional optimization sample, a type label is set for related text in the directional optimization sample that is related to the user intent based on the user intent corresponding to the directional optimization sample. For example, if the user intent is to watch a movie, the type label of the related text can be set to "movie." Then, the related text is segmented into words to obtain keywords for the related text. Furthermore, a search is performed within a knowledge base for each field based on the keywords and type label of the related text to obtain domain knowledge matching the related text. Here, domain knowledge that has the highest relevance to the keywords in the related text and whose type matches the type label of the related text can be selected as the domain knowledge matching the related text. Next, a relevance relationship is established between the related text and the domain knowledge matched with the related text.

[0092] After the relevance relationship is established, when the same user again inputs the same initial interaction information as the directional optimization sample, the query proxy in the proxy layer performs a search within the knowledge base for each domain based on the post-processing interaction vector of the post-processing interaction information corresponding to the corresponding initial interaction information. Because the above-mentioned relevance relationship has been established between the related text in the directional optimization sample and the domain knowledge matched to the related text, the domain knowledge can be immediately detected and identified as the domain knowledge matching the corresponding post-processing interaction vector. This enables the GPT model to more accurately understand the user's intent and the true meaning of each part of the text in the post-processing interaction information (even if the text is entered incorrectly), thereby improving the accuracy of responses. In addition, after the same user again inputs the same initial interaction information as the directional optimization sample, the intelligent device may respond directly based on the domain knowledge matched to the relevant text described above, thereby realizing quick understanding of the user's intent and quick response while avoiding excessive reliance on an external GPT model.

[0093] In any of the above embodiments, to further enhance the degree of user privacy protection, when the intelligent device sends user interaction information and user preference information to the query proxy, the user preference information may be obtained after obfuscating the true user preferences. In this case, based on the user interaction information, a true user profile label associated with the user interaction information in the true user preferences may be identified, and the associated true user profile label may be superordinated (e.g., mala hotpot may be superordinated as spicy food), to obtain user preference information including an obfuscated profile label corresponding to the associated true user profile label. After receiving the interaction response information sent from the response proxy, the intelligent device may first modify the interaction response information based on the true user preferences, and then secondarily modify the interaction response information after the first modification using user privacy data related to the user interaction information, thereby obtaining modification processing information.

[0094] If the user interaction information is information suggestion-based, the interaction response information may be first corrected using a local interference cancellation model on the intelligent device. The interference cancellation model includes four network layers. The first layer is a true preference extraction layer, which can be constructed based on a long short-term memory (LSTM) network or its modified results. It is used to extract hidden layer features of the associated true user profile labels based on the interaction behavior corresponding to the associated true user profile labels. The second layer is an obfuscated preference extraction layer, similar in structure to the true preference extraction layer, which is used to extract hidden layer features of the obfuscated profile labels based on the interaction behavior corresponding to the obfuscated profile labels corresponding to the associated true user profile labels. The third layer is a response vector conversion layer, which can be constructed based on a fully connected network. It is used to convert the type distribution of each response item (e.g., the description information of each restaurant) in the interaction response information into a response vector. Alternatively, the type distribution may be obtained using statistical methods by first classifying the description information of each response item using a classifier to obtain the type of each response item. The fourth layer is a correction layer, which is a fully connected layer that converts the combined results of the hidden layer features of the relevant true user profile labels, the hidden layer features of the obfuscated profile labels, and the response vectors into a new corrected type distribution, and uses the corrected type distribution to select response items corresponding to types with high occupancy rates and combine them as the first-stage corrected interaction response information.

[0095] When the user interaction information is of a command control system, by matching each noun in the interaction response information with the above-mentioned related true user profile label in the true user preferences, nouns in the interaction response information that match with the above-mentioned related true user profile label are obtained, and by replacing the matched nouns with these based on the above-mentioned related true user profile label, it is possible to obtain interaction response information after primary correction.

[0096] Based on any of the above embodiments, the intelligent device may further improve the user's interaction experience by appropriately processing interaction barrier situations. In this case, the intelligent device may acquire feedback information input by the user after a response, collect images of the user inputting the feedback information, and perform emotion recognition on the feedback information to acquire the emotion type of the feedback information, or perform facial expression recognition on the images of the user inputting the feedback information to acquire the facial expression type of the user inputting the feedback information. If the feedback information contains a negative keyword, if the emotion type of the feedback information is a negative emotion, or if the facial expression type of the user inputting the feedback information is a negative expression, it may determine that an interaction barrier exists at that time. If an interaction barrier exists at that time, it may acquire question-and-answer history information corresponding to the user, search for and acquire question-and-answer history information matching the initial interaction information from the question-and-answer history information, and respond based on the question-and-answer history information matching the initial interaction information.

[0097] In any of the above embodiments, a user may encounter abnormal situations during interaction, such as slow connection speeds or even network disconnections, for example, when passing through a section outside of network coverage, such as a tunnel, or a section with a weak signal. In such cases, it is necessary to support the user in using some basic interaction functions even when offline. Therefore, a local GPT model capable of performing inference and completing some simple answers using local computing power may be deployed on the intelligent device. In this case, the type of network condition at that time may be checked, and if the type of network condition is a deteriorating network condition or a network disconnection, user interaction information and user preference information of the user may be sent to the locally deployed local GPT model to obtain local response information returned from the local GPT model, and a response may be provided to the user based on the local response information. Here, since the computing power of the intelligent device is limited, a locally deployable local GPT model may be obtained by performing model quantization compression based on a large-scale sample GPT model (e.g., LLaMA-6B). Specifically, the following quantization compression method may be used to achieve model compression optimization.

[0098] 1. Optimization of the model parameter amount: By compressing floating-point constants from float32 to int8 and using the area mapping method, the model parameter amount can be reduced. This method allows the model to be converted from an fp32 state requiring 24GB of memory to an int8 state requiring only 6GB of memory, resulting in a savings of approximately four times the amount of calculation.

[0099] 2. Parameter Gradient Optimization: Next, the parameters of the model may be further optimized, which can be achieved by parameter gradient optimization.

[0100] 3. Use AdamW optimizer: You can use AdamW optimizer. This optimizer requires 6B x 2 bytes = 12GB of memory.

[0101] 4. Use CUDA kernel: You can use CUDA kernel. This tool occupies 1.3GB of memory.

[0102] 5. Mixed-precision optimization: Mixed-precision optimization methods can further improve computational efficiency and reduce memory occupation.

[0103] 6. High-order matrix transformation based on Lora: Using the high-order matrix transformation technique based on Lora, a high-order matrix can be transformed into the dot product of two low-order matrices, thereby reducing memory occupation.

[0104] In summary, by using the above optimization techniques, it is possible to control the memory requirements for a portion of the LLaMA-6B model with int8 precision to approximately 25.3 GB, and by quantizing the weights to 4 bits and using a common scale and zeros, it is possible to minimize the loss of precision.

[0105] It should be noted that the quantization method for model compression described above is only one possible method. For example, model compression methods such as model pruning may also be possible methods for building a local GPT model, and embodiments of the present invention are not limited in this respect.

[0106] 3 is a schematic flowchart 3 of the intelligent device interaction method based on the GPT model provided in the present invention. As shown in FIG. 3, the method includes:

[0107] Step 310: The intelligent device side receives the initial interaction information input by the user, and performs a masking process on the initial interaction information to obtain user interaction information.

[0108] Step 320: The intelligent device side sends the user interaction information and the user preference information of the user to the proxy layer.

[0109] Step 330: The proxy layer receives user interaction information and user preference information sent from the intelligent device side, processes the user interaction information to obtain post-processing interaction information, and then sends the post-processing interaction information to the GPT model, which then returns task processing information of the post-processing interaction information.

[0110] Step 340: After receiving the task processing information of the post-processing interaction information returned from the GPT model, the proxy layer obtains interaction response information by fusing it based on the user preference information and the task processing information, and sends the interaction response information to the intelligent device side.

[0111] Step 350: The intelligent device side receives the interaction response information sent by the proxy layer, modifies the interaction response information based on the user privacy data related to the user interaction information, and responds to the user based on the modified information.

[0112] It should be noted that the interaction method presented in this embodiment is similar to the interaction method presented in the embodiment corresponding to the proxy layer and the interaction method presented in the embodiment corresponding to the intelligent device side, and therefore will not be described in further detail here.

[0113] In a method provided in an embodiment of the present invention, the intelligent device masks initial interaction information input by a user to obtain user interaction information, which is then sent to the proxy layer together with user preference information. The query proxy in the proxy layer then processes the user interaction information to obtain post-processing interaction information, which is then sent to the GPT model. The GPT model then performs intent recognition and task segmentation based on the post-processing interaction information and returns task processing information in the post-processing interaction information. The response proxy receives the task processing information in the post-processing interaction information returned from the GPT model, combines it with the user preference information and the task processing information to obtain interaction response information for the user interaction information, and sends the interaction response information to the intelligent device. The intelligent device then modifies the interaction response information based on user privacy data related to the user interaction information and responds to the user based on the modified information. This improves the accuracy of responses while ensuring that user privacy data remains only on the intelligent device, thereby protecting the user's privacy.

[0114] Next, the interaction device of an intelligent device based on the GPT model provided by the present invention will be described. Note that the interaction device of an intelligent device based on the GPT model described below and the interaction method of an intelligent device based on the GPT model described above can be referred to correspondingly.

[0115] Based on any of the above embodiments, Figure 4 shows a schematic structural diagram 1 of an intelligent device interaction device based on the GPT model provided by the present invention. The device is deployed in the proxy layer. As shown in Figure 4, the device includes:

[0116] The information receiving unit 410 receives user interaction information and user preference information sent from the intelligent device side. The user interaction information is the initial interaction information input by the user that has been masked by the intelligent device side.

[0117] Inquiry proxy 420: After processing the user interaction information to obtain post-processing interaction information, it sends the post-processing interaction information to the GPT model, which then returns the task processing information of the post-processing interaction information.

[0118] Response proxy 430: After receiving the task processing information of the processed interaction information returned from the GPT model, it obtains interaction response information by fusing it based on the user preference information and the task processing information, and sends the interaction response information to the intelligent device side, so that the intelligent device side modifies the interaction response information based on the user privacy data related to the user interaction information, and responds to the user based on the modified information.

[0119] Based on any of the above embodiments, when the user interaction information is a command control system, obtaining interaction response information by combining the above user preference information and the task processing information specifically includes the following:

[0120] That is, by performing matching based on each noun in the task processing information and each preference label in the user preference information, preference labels that match each noun in the task processing information are obtained.

[0121] The intelligent device acquires interaction response information by replacing corresponding nouns in the task processing information based on the label values ​​of preference labels matched to each noun in the task processing information. After receiving the interaction response information, the intelligent device acquires correction processing information by correcting the interaction response information based on user privacy data related to the user interaction information, and executes the correction processing information to realize a response to the user.

[0122] Based on any of the above embodiments, when the user interaction information is an information suggestion type, obtaining interaction response information of the user interaction information by replacing corresponding nouns in the task processing information based on the label values ​​of preference labels matched to each noun in the task processing information specifically includes the following:

[0123] That is, by performing matching based on each noun in the task processing information and each preference label in the user preference information, preference labels that match each noun in the task processing information are obtained.

[0124] Based on the label value of the preference label matched to each noun in the task processing information, the corresponding noun in the task processing information is replaced to obtain complete task processing information.

[0125] The intelligent device side receives the interaction response information, selects the interaction response information based on the user privacy data related to the user interaction information, and returns the selected response information to the user to realize a response to the user.

[0126] According to any of the above embodiments, when sending the above-mentioned post-processing interaction information to the GPT model, it specifically includes:

[0127] That is, the post-processing interaction information is vectorized to obtain a post-processing interaction vector.

[0128] Based on the post-processing interaction vector, a search is performed within the knowledge base of each field to identify the field knowledge and the field knowledge vector that match the post-processing interaction vector within the knowledge base of each field.

[0129] The processed interaction vector is combined with the domain knowledge vector of the matched domain knowledge to obtain a fused query vector, and the fused query vector is sent to the GPT model.

[0130] According to any of the above embodiments, when sending the fused query vector to the GPT model, the following steps are specifically included:

[0131] That is, by sending the fused query vector to multiple GPT models, task processing information of the post-processing interaction information returned from the multiple GPT models is obtained.

[0132] The post-processing interaction information returned from the plurality of GPT models is fused or selected based on task processing information to obtain optimized processing information, which is then fused with the user preference information to obtain the interaction response information.

[0133] According to any of the above embodiments, the proxy layer is constructed according to the following steps.

[0134] That is, an initial model is constructed based on a pre-trained text generation model.

[0135] Sample input information is input to the initial model, and test output information is output from the initial model.

[0136] The proxy layer is obtained by fine-tuning the initial model based on an evaluation of the test output information or based on differences between the test output information and sample output information.

[0137] Based on any of the above embodiments, Figure 5 shows a schematic structural diagram 2 of an intelligent device interaction device based on the GPT model provided by the present invention. The device is deployed on the intelligent device side. As shown in Figure 5, the device includes:

[0138] Input processing unit 510: Accepts initial interaction information input by a user, and obtains user interaction information by performing masking processing on the initial interaction information.

[0139] An information forwarding unit 520 sends the user interaction information and the user preference information of the user to a proxy layer, so that the proxy layer processes the user interaction information to obtain post-processing interaction information, and then sends the post-processing interaction information to a GPT model, so that the GPT model returns task processing information of the post-processing interaction information.

[0140] An information response unit 530 receives the interaction response information sent by the proxy layer, modifies the interaction response information according to the user privacy data related to the user interaction information, and responds to the user based on the modified information. The interaction response information is obtained by fusing the user preference information and the task processing information after the proxy layer receives the task processing information of the processed interaction information returned from the GPT model.

[0141] Based on any of the above embodiments, if the user interaction information is a command control system, modifying the interaction response information based on the user privacy data related to the above user interaction information, and responding to the user based on the modified information specifically includes:

[0142] That is, based on the user privacy data related to the user interaction information, the interaction response information is corrected to obtain correction process information, and the correction process information is executed to realize a response to the user.

[0143] The interaction response information is obtained by the proxy layer obtaining preference labels that match each noun in the task processing information by matching each noun in the task processing information with each preference label in the user preference information, and then replacing the corresponding noun in the task processing information based on the label value of the preference label matched to each noun in the task processing information.

[0144] Based on any of the above embodiments, if the user interaction information is an information suggestion type, modifying the interaction response information based on the user privacy data related to the above user interaction information, and responding to the user based on the modified information specifically includes:

[0145] That is, the interaction response information is selected based on the user privacy data related to the user interaction information, selected response information is acquired, and the selected response information is returned to the user, thereby realizing a response to the user.

[0146] The interaction response information is obtained by the proxy layer obtaining preference labels that match each noun in the task processing information by matching each noun in the task processing information with each preference label in the user preference information, then obtaining complete task processing information by replacing the corresponding noun in the task processing information based on the label value of the preference label matched to each noun in the task processing information, and then executing the complete task processing information.

[0147] According to any of the above embodiments, the apparatus further comprises an optimization unit, which is used for:

[0148] That is, the initial interaction information and the corresponding revised information of multiple users are collected, and the initial interaction information corresponding to the revised information that does not meet the user's expectations is recorded as a directional optimization sample.

[0149] Intent recognition is performed on the directional optimization sample to obtain a user intent corresponding to the directional optimization sample, and a type label is set for relevant text in the directional optimization sample that is related to the user intent based on the user intent.

[0150] The related text is segmented into words to obtain keywords of the related text, and a search is performed within a knowledge base of each field based on the keywords of the related text and the type label of the related text to obtain field knowledge that matches the related text.

[0151] A relevance relationship is established between the related text and the domain knowledge matched with the related text, so that when the same user again inputs the same initial interaction information as the directional optimization sample and a search is performed in the knowledge base of each domain based on the post-processing interaction vector of the post-processing interaction information corresponding to the corresponding initial interaction information, the domain knowledge matched with the related text is identified as the domain knowledge matching the corresponding post-processing interaction vector, or a response is made to the initial interaction information based on the domain knowledge matched with the related text.

[0152] According to any of the above embodiments, the device further comprises an interaction barrier processing unit for use in:

[0153] That is, the feedback information input by the user after the response is obtained, and an image of the user inputting the feedback information is collected.

[0154] Emotion recognition is performed on the feedback information to obtain an emotion type of the feedback information.

[0155] The facial expression type when the user inputs the feedback information is acquired by performing facial expression recognition on an image when the user inputs the feedback information.

[0156] If the feedback information contains a negative keyword, if the emotional type of the feedback information is a negative emotional type, or if the facial expression type when the user inputs the feedback information is a negative facial expression, it is determined that an interaction barrier exists at that time.

[0157] If an interaction barrier exists at that time, each piece of question and answer history information corresponding to the user is acquired, and question and answer history information that matches the initial interaction information is searched for from each piece of question and answer history information, and a response is made based on the question and answer history information that matches the initial interaction information.

[0158] According to any of the above embodiments, the apparatus further comprises a local processing unit for:

[0159] That is, the type of network condition at that time is checked, and when the type of network condition is a network condition deterioration or a network disconnection state, the user interaction information and the user preference information of the user are sent to a local GPT model deployed locally, thereby obtaining local response information returned from the local GPT model, and responding to the user based on the local response information.

[0160] According to any of the above embodiments, the local GPT model is obtained by performing model quantization compression on the sample GPT model.

[0161] Based on any of the above embodiments, Figure 6 is a schematic structural diagram of an intelligent device interaction system based on the GPT model provided by the present invention. As shown in Figure 6, the system includes a proxy layer 610 and an intelligent device side 620.

[0162] The proxy layer 610 is deployed with an intelligent device interaction device based on the GPT model provided in the corresponding embodiment above, and the intelligent device side 620 is deployed with an intelligent device interaction device based on the GPT model provided in the corresponding embodiment above.

[0163] The above-described device embodiments are merely schematic, and the above units described as separate components may or may not be physically separated. Furthermore, the components shown as units may or may not be physical units, i.e., they may be located in one place or distributed across multiple network units. The objectives of the solutions of the present embodiments can be realized by selecting some or all of the modules according to actual needs. Furthermore, those skilled in the art can understand and implement the present embodiments without any creative effort.

[0164] From the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be realized by combining software with a required general-purpose hardware platform, or of course by hardware. Based on this understanding, the above technical solutions can be embodied essentially or in part contributing to the prior art in the form of a software product. This computer software product can be stored in a computer-readable storage medium such as a ROM / RAM, a magnetic disk, or an optical disk, and includes several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to execute the method described in each embodiment or any part of the embodiment.

[0165] Finally, it should be noted that the above embodiments are merely for illustrating the technical solutions of the present invention, and are not intended to be limiting. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments may be modified or some of the technical features may be replaced with equivalents, and such modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An intelligent device interaction method based on the GPT model, applied to a proxy layer, receiving user interaction information and user preference information transmitted from an intelligent device side, the user interaction information being information obtained after the intelligent device side has subjected initial interaction information input by a user to a masking process; After processing the user interaction information to obtain post-processing interaction information, the post-processing interaction information is sent to the GPT model, so that the GPT model returns task processing information of the post-processing interaction information; After receiving the task processing information returned from the GPT model, obtaining interaction response information by fusing based on the user preference information and the task processing information, and sending the interaction response information to the intelligent device side, so that the intelligent device side modifies the interaction response information based on user privacy data related to the user interaction information, and responds to the user based on the modified information; When acquiring interaction response information by fusing based on the user preference information and the task processing information, specifically, by performing matching based on each noun in the task processing information and each preference label in the user preference information, obtaining preference labels that match each noun in the task processing information; and acquiring the interaction response information by replacing the corresponding noun in the task processing information based on the label value of the preference label matched to each noun in the task processing information, and the intelligent device side, after receiving the interaction response information, acquiring correction process information by correcting the interaction response information based on user privacy data related to the user interaction information, and executing the correction process information to realize a response to the user. A method characterized by:

2. When transmitting the processed interaction information to the GPT model, specifically: vectorizing the processed interaction information to obtain a processed interaction vector; searching a knowledge base for each field based on the post-processing interaction vector to identify domain knowledge and its domain knowledge vector that match the post-processing interaction vector in the knowledge base for each field; Combining the processed interaction vector with the domain knowledge vector of the matched domain knowledge to obtain a fused query vector, and sending the fused query vector to the GPT model. The intelligent device interaction method based on the GPT model as claimed in claim 1.

3. When transmitting the fused query vector to the GPT model, specifically: Sending the fused query vector to a plurality of the GPT models to obtain task processing information of the post-processing interaction information returned from the plurality of the GPT models; and fusing or selecting the post-processing interaction information returned from the plurality of GPT models based on the task processing information to obtain optimized processing information, and fusing the optimized processing information with the user preference information to obtain the interaction response information. The intelligent device interaction method based on the GPT model as claimed in claim 2.

4. The proxy layer Build an initial model based on a pre-trained text generation model, inputting sample input information into the initial model to obtain test output information output from the initial model; obtaining the proxy layer by fine-tuning the initial model based on an evaluation of the test output information or based on a difference between the test output information and sample output information; It is built on the steps of The intelligent device interaction method based on the GPT model as claimed in claim 1.

5. An intelligent device interaction method based on the GPT model, which is applied to an intelligent device side, receiving initial interaction information input by a user and performing a masking process on the initial interaction information to obtain user interaction information; Send the user interaction information and the user preference information of the user to a proxy layer, so that the proxy layer processes the user interaction information to obtain post-processing interaction information, and then sends the post-processing interaction information to the GPT model, so that the GPT model returns task processing information of the post-processing interaction information; receiving interaction response information sent by the proxy layer, modifying the interaction response information based on user privacy data related to the user interaction information, and responding to the user based on the modified information, wherein the interaction response information is obtained by combining the task processing information of the processed interaction information returned from the GPT model by the proxy layer, based on the user preference information and the task processing information; The interaction response information is corrected based on the user privacy data related to the user interaction information, and when responding to the user based on the corrected information, specifically, correcting the interaction response information based on user privacy data related to the user interaction information to obtain correction processing information, and executing the correction processing information to realize a response to the user; The interaction response information is obtained by the proxy layer acquiring preference labels that match each noun in the task processing information by performing matching based on each noun in the task processing information and each preference label in the user preference information, and then replacing the corresponding noun in the task processing information based on the label value of the preference label that matches each noun in the task processing information. A method characterized by:

6. The method further comprises: Collecting initial interaction information and corresponding modified information of a plurality of users, and recording the initial interaction information corresponding to the modified information that does not meet the user's expectations as a directional optimization sample; performing intent recognition on the directional optimization sample to obtain a user intent corresponding to the directional optimization sample, and setting a type label for associated text in the directional optimization sample that is related to the user intent based on the user intent; segmenting the related text into words, acquiring keywords of the related text, and searching a knowledge base of each field based on the keywords of the related text and the type label of the related text to acquire domain knowledge matching the related text; establishing a relationship of relevance between the related text and the domain knowledge matched to the related text, so that when the same user again inputs the same initial interaction information as the directional optimization sample and then searches in the knowledge base of each domain based on the post-processing interaction vector of the post-processing interaction information corresponding to the corresponding initial interaction information, the domain knowledge matched to the related text is identified as domain knowledge matching the corresponding post-processing interaction vector, or a response is made to the initial interaction information based on the domain knowledge matched to the related text. The intelligent device interaction method based on the GPT model as claimed in claim 5.

7. The method further comprises: obtaining feedback information input by the user after the response, and collecting an image of the user inputting the feedback information; performing emotion recognition on the feedback information to obtain an emotion type of the feedback information; acquiring a facial expression type when the user inputs the feedback information by performing facial expression recognition on an image when the user inputs the feedback information; If the feedback information includes a negative keyword, if the emotion type of the feedback information is a negative emotion, or if the facial expression type when the user inputs the feedback information is a negative facial expression, it is determined that an interaction barrier exists at that time; If an interaction barrier exists at that time, acquiring each piece of question and answer history information corresponding to the user, searching for question and answer history information that matches the initial interaction information from each piece of question and answer history information, and responding based on the question and answer history information that matches the initial interaction information. The intelligent device interaction method based on the GPT model as claimed in claim 5.

8. The method further comprises: Checking a type of network condition at that time, and when the type of network condition is a network condition deterioration or a network disconnection state, sending the user interaction information and the user preference information of the user to a local GPT model deployed locally, thereby obtaining local response information returned from the local GPT model, and responding to the user based on the local response information; The intelligent device interaction method based on the GPT model as claimed in claim 5.

9. The local GPT model is obtained by performing model quantization compression on the sample GPT model. The intelligent device interaction method based on the GPT model as claimed in claim 8.

10. An intelligent device interaction method based on a GPT model, comprising: the intelligent device side receives initial interaction information input by the user and performs a masking process on the initial interaction information to obtain user interaction information; The intelligent device side sends the user interaction information and the user preference information of the user to a proxy layer; the proxy layer receives the user interaction information and the user preference information sent from the intelligent device side, processes the user interaction information to obtain post-processing interaction information, and then sends the post-processing interaction information to the GPT model, so that the GPT model returns task processing information of the post-processing interaction information; the proxy layer receives the task processing information of the post-processing interaction information returned from the GPT model, and then combines the user preference information and the task processing information to obtain interaction response information, and sends the interaction response information to the intelligent device side; the intelligent device side receives the interaction response information sent by the proxy layer, modifies the interaction response information based on user privacy data related to the user interaction information, and responds to the user based on the modified information; The intelligent device side modifies the interaction response information based on the user privacy data related to the user interaction information, and responds to the user based on the modified information, specifically, correcting the interaction response information based on user privacy data related to the user interaction information to obtain correction processing information, and executing the correction processing information to realize a response to the user; the interaction response information is obtained by the proxy layer performing matching based on each noun in the task processing information and each preference label in the user preference information to obtain preference labels that match with each noun in the task processing information, and then replacing the corresponding noun in the task processing information based on the label value of the preference label matched to each noun in the task processing information; A method characterized by:

11. An interaction apparatus for intelligent devices based on the GPT model, deployed in a proxy layer, an information receiving unit for receiving user interaction information and user preference information transmitted from an intelligent device, the user interaction information being information obtained after the intelligent device has subjected initial interaction information input by a user to a masking process; a query proxy that processes the user interaction information to obtain post-processing interaction information, and then sends the post-processing interaction information to the GPT model, so that the GPT model returns task processing information of the post-processing interaction information; a response proxy that receives the task processing information returned from the GPT model, and then combines the user preference information and the task processing information to obtain interaction response information, and sends the interaction response information to the intelligent device side, so that the intelligent device side modifies the interaction response information based on user privacy data related to the user interaction information, and responds to the user based on the modified information; When acquiring interaction response information by fusing based on the user preference information and the task processing information, specifically, by performing matching based on each noun in the task processing information and each preference label in the user preference information, obtaining preference labels that match each noun in the task processing information; The interaction response information is acquired by replacing the corresponding noun in the task processing information based on the label value of the preference label matched to each noun in the task processing information, and the intelligent device side, after receiving the interaction response information, acquires correction processing information by correcting the interaction response information based on user privacy data related to the user interaction information, and executes the correction processing information to realize a response to the user. An apparatus characterized in that

12. An intelligent device interaction device based on the GPT model, which is deployed on the intelligent device side, an input processing unit that receives initial interaction information input by a user and performs a masking process on the initial interaction information to obtain user interaction information; an information forwarding unit that sends the user interaction information and the user preference information of the user to a proxy layer, so that the proxy layer processes the user interaction information to obtain post-processing interaction information, and then sends the post-processing interaction information to a GPT model, so that the GPT model returns task processing information of the post-processing interaction information; an information response unit that receives interaction response information sent by the proxy layer, modifies the interaction response information based on user privacy data related to the user interaction information, and responds to the user based on the modified information, wherein the interaction response information is obtained by fusing the task processing information of the processed interaction information returned from the GPT model by the proxy layer based on the user preference information and the task processing information; The interaction response information is corrected based on the user privacy data related to the user interaction information, and when responding to the user based on the corrected information, specifically, based on user privacy data related to the user interaction information, correcting the interaction response information to obtain corrective processing information, and executing the corrective processing information to realize a response to the user; the interaction response information is obtained by the proxy layer performing matching based on each noun in the task processing information and each preference label in the user preference information to obtain preference labels that match with each noun in the task processing information, and then replacing the corresponding noun in the task processing information based on the label value of the preference label matched to each noun in the task processing information; An apparatus characterized in that

13. An intelligent device interaction system based on a GPT model, comprising: Including a proxy layer and an intelligent device side, The proxy layer is deployed with an interaction device for the intelligent device based on the GPT model as claimed in claim 11; The system is characterized in that an interaction device for an intelligent device based on the GPT model according to claim 12 is deployed on the intelligent device side.

Images