AI interaction system and storage medium thereof

Abstract

The invention discloses an AI interaction system which comprises an AI interaction function module, a power management module, a signal management module and a communication module. The AI interaction function module is deeply fused with hardware equipment, convenient, fast and intelligent AI interaction experience is achieved, user experience and equipment functionality are remarkably improved, the connection state of Host equipment is automatically detected through a built-in recognition unit and a built-in software management unit, AI software is automatically downloaded and installed in the first-time connection process, and therefore the user experience is improved. The software is automatically started during non-first-time connection, a user does not need to manually install or update the software and does not need to frequently open a browser to access a specific page, and the operation process is greatly simplified; through deep fusion of the AI interaction function module and the hardware equipment, the limitation of the existing AI interaction mode is solved, the interaction convenience, the intelligent level and the user experience are remarkably improved, and meanwhile, the system performance and the resource utilization efficiency are optimized through cooperative work of the local and the cloud.

Description

Technical Field

[0001] The present invention relates to the field of intelligent interaction technologies, and particularly to an AI interaction system and its storage medium. Background Art

[0002] With the rapid development of artificial intelligence technologies, especially the wide application of large AI models, the combination of AI and hardware devices has become an important direction for promoting industrial upgrading. However, the common interaction methods of large AI models in the market mainly include two forms: web version and local application. However, both of these interaction methods have certain limitations: web version interaction requires users to frequently open browsers and access specific pages, with cumbersome operations; although local applications can be used offline, they require users to manually install and update, and occupy device storage space.

[0003] In recent years, the field of AI hardware has gradually shown a development trend of specialization, integration, and customization. For example, multimodal interaction technologies provide a more natural and user-friendly experience by integrating multiple input methods such as sound, images, and gestures. Some enterprises have begun to explore integrating large AI models into hardware devices. This technology not only enhances the functionality of hardware devices but also provides users with a more convenient and intelligent interaction experience to achieve more efficient interaction and intelligent experience.

[0004] Therefore, how to effectively combine such devices with large AI model technologies to achieve a convenient and fast AI interaction experience through hardware devices is a problem of great practical significance. Summary of the Invention

[0005] The present invention provides an AI interaction system and its storage medium to solve the above-mentioned existing technical problems.

[0006] The technical solution of the present invention is implemented as follows:

[0007] An AI interaction system includes an AI interaction function module, a power management module, a signal management module, and a communication module;

[0008] The AI interaction function module automatically identifies the host device, installs or opens AI software on the host device, receives interaction instructions input by the user in the AI software, and performs preliminary processing on the instructions. It calls the cloud large model interface through a network connection, sends the processed user instructions to the cloud for calculation, receives the calculation results returned by the cloud large model, and displays them in the AI software in real time to complete the closed-loop of user interaction;

[0009] The signal management module processes and manages signal transmission within the storage medium of the AI interaction system and between external devices, converts signals of different interfaces into formats suitable for internal processing, distributes signals to corresponding interfaces or devices according to user requirements and device connection status, and performs enhancement and optimization processing on the transmitted signals to reduce signal attenuation and interference.

[0010] The communication module is responsible for network connection and data transmission between the storage medium of the AI interaction system and the cloud large model, automatically establishes and maintains a network connection with the cloud server, securely transmits the instructions input by the user to the cloud large model interface, and encrypts the transmitted data.

[0011] Furthermore, the AI interaction function module includes an identification unit, a software management unit, an interaction information processing unit, a cloud communication unit, a feedback unit, and a multimodal interaction unit.

[0012] The identification unit detects the connection status between the host device and the storage medium of the AI interaction system and determines whether it is the first connection.

[0013] The software management unit manages the installation, update, and startup of AI software.

[0014] The interaction information processing unit receives and processes the interaction instructions input by the user.

[0015] The cloud communication unit conducts network connection and data transmission with the cloud large model.

[0016] The feedback unit displays the results returned by the cloud large model in the AI software in real time and collects user feedback.

[0017] The multimodal interaction unit supports and expands multimodal interaction capabilities.

[0018] Furthermore, the interaction information processing unit supports multiple input methods, including speech recognition, button input, touch input, and gesture recognition. The interaction information processing unit performs preprocessing operations on the instructions input by the user to ensure that the instructions can be correctly understood by the cloud large model.

[0019] The interaction information processing unit fuses multiple input signals according to the instructions input by the user through multiple input methods simultaneously to generate a unified interaction instruction.

[0020] Caches the instructions input by the user and supports the user to view or modify historical instructions at any time.

[0021] Further, the process of the interaction information processing unit preprocessing the instructions input by the user is specifically as follows: receiving the instructions issued by the user through various input methods, converting the collected input signals into a unified format, where the voice signals are converted into text and background noise is removed, and natural language processing technology is used to preliminarily analyze the text to extract key information;

[0022] converting the gesture actions into corresponding instructions and removing interference signals;

[0023] formatting the input text, removing extra spaces and punctuation marks, and performing preliminary grammar checks to ensure the logic of the text.

[0024] Further, the multi-modal interaction unit reserves interfaces and modules to support newly added input methods in the future, supports seamless switching between different input methods by the user, and automatically selects the optimal input modality according to the scenario. For multi-modal input, the interaction logic and algorithms are optimized, and the supported input modalities are automatically adapted according to the hardware capabilities of different devices.

[0025] Further, the process of the multi-modal interaction unit optimizing the interaction logic and algorithms for multi-modal input is specifically as follows: perceiving and processing the signals from different input modalities, converting the multi-modal input signals into a unified feature vector, and integrating them into a unified multi-modal context model;

[0026] Based on the multi-modal input fusion, the interaction unit judges the user's intention through intention recognition technology;

[0027] The multi-modal interaction unit generates the optimal behavior sequence according to the result of intention recognition;

[0028] Integrating emotion recognition technology to adjust the interaction method according to the user's emotional state;

[0029] Through cross-modal fusion technology, the input signals of different modalities are comprehensively processed, and the interaction performance is improved through continuous learning and optimization algorithms. According to the optimized interaction logic, multi-modal output signals are generated and fed back to the user.

[0030] Further, the AI interaction function module also has a local AI large model built in. When the AI software is started, the local AI large model is automatically loaded. Combining the natural language processing ability of the local AI large model, the user's intention is quickly recognized. Combining the user's local data and preferences, the local AI large model provides personalized interaction results.

[0031] Furthermore, the AI interaction function module communicates with the cloud large model, regularly synchronizes the data and parameters of the local AI large model and the cloud large model. When more powerful computing capabilities or access to cloud resources are needed, the local AI large model collaborates with the cloud large model through the cloud communication unit, sending complex tasks to the cloud for processing;

[0032] The AI interaction function module optimizes the operating efficiency of the local AI large model according to the hardware capabilities of the device;

[0033] Dynamically adjusts the parameters and strategies of the local AI large model according to the user's real-time feedback and interaction history;

[0034] The local AI large model optimizes its own performance by continuously learning the user's input and feedback.

[0035] Furthermore, the power management module provides stable power support for the AI interaction system storage medium and the devices connected thereto, manages the power distribution and conversion, dynamically allocates power according to the types of connected devices, converts the input power into voltages and currents suitable for the internal components of the AI interaction system storage medium and the connected devices, monitors the power status in real time, and automatically enters the low-power mode when the device is idle.

[0036] An AI interaction system storage medium includes an expander. The expander is provided with a number of interfaces for connecting to host devices, and a number of buttons for interacting with AI software. The expander is built-in with a central processing unit for carrying the AI interaction system and realizes the operation of the AI interaction system;

[0037] The expander is also provided with interaction hardware that supports multiple instruction input methods.

[0038] Compared with the prior art, the present invention has the following beneficial effects:

[0039] 1. By deeply integrating the AI interaction function module with the hardware device, the present invention realizes a convenient, fast and intelligent AI interaction experience, significantly improving the user experience and device functionality. Through the built-in recognition unit and software management unit, it automatically detects the connection status of the Host device, automatically downloads and installs the AI software at the first connection, and automatically starts the software at non-first connection. Users do not need to manually install or update the software, nor do they need to frequently open the browser to access specific pages, greatly simplifying the operation process. In addition, the interaction information processing unit supports multiple input methods such as voice, button, touch and gesture, and can integrate multiple input signals to generate unified interaction instructions. Users can choose the most convenient interaction method according to their own needs and scenarios, further improving the flexibility of interaction;

[0040] 2. Through the fast processing ability of the local AI large model and combined with the optimized communication module, the present invention provides real-time interaction feedback, reduces latency, and significantly improves the user experience. The local AI large model combines the user's local data and preferences to provide personalized interaction results. For example, it automatically adjusts the interaction logic according to the user's historical records and preferences to provide more accurate services. In addition, the multi-modal interaction unit integrates emotion recognition technology and can adjust the interaction method according to the user's emotional state to make the interaction more natural and user-friendly.

[0041] 3. By integrating the AI interaction function module into the expander, the present invention not only retains the interface expansion function of the traditional expander but also endows it with intelligent interaction capabilities. The microphone, camera, and touch screen on the expander support voice, image, and gesture input, enhancing the versatility of the device. The multi-modal interaction unit can automatically adapt to the supported input modalities according to the hardware capabilities of the device and dynamically adjust the interaction logic based on the user's real-time feedback and interaction history, further enhancing the intelligent level of the hardware device.

[0042] By deeply integrating the AI interaction function module with the hardware device, the present invention not only solves the limitations of existing AI interaction methods but also significantly improves the convenience, intelligent level, and user experience of the interaction. At the same time, through the collaborative work of the local and the cloud, the system performance and resource utilization efficiency are optimized. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a system framework diagram of an AI interaction system for Embodiment 1;

[0044] Figure 2 It is a module framework diagram of the AI interaction function module in Embodiment 1;

[0045] Figure 3 It is a functional flowchart of an AI interaction system for Embodiment 1. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0046] To make the objectives, features, and advantages of the present invention more obvious and understandable, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the embodiments described below are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0047] Embodiment 1

[0048] As Figures 1 - 3 shown, an AI interaction system includes an AI interaction function module, a power management module, a signal management module, and a communication module;

[0049] The AI interaction function module automatically identifies the host device, installs or opens the AI software on the host device, receives the interaction instructions input by the user in the AI software, and performs preliminary processing on the instructions. It calls the cloud large model interface through a network connection, sends the processed user instructions to the cloud for calculation, receives the calculation results returned by the cloud large model, and displays them in real time in the AI software to complete the closed-loop of user interaction;

[0050] The signal management module processes and manages the signal transmission inside the storage medium of the AI interaction system and between external devices, converts the signals of different interfaces into a format suitable for internal processing, and distributes the signals to the corresponding interfaces or devices according to user requirements and device connection status. It performs enhancement and optimization processing on the transmitted signals to reduce signal attenuation and interference;

[0051] The communication module is responsible for the network connection and data transmission between the storage medium of the AI interaction system and the cloud large model. It automatically establishes and maintains a network connection with the cloud server, securely transmits the instructions input by the user to the cloud large model interface, and encrypts the transmitted data.

[0052] In one embodiment, the AI interaction function module is the core of the system. It can automatically identify the connection status of the Host device and automatically install or start the AI software at the first connection. This module receives the interaction instructions of the user, preprocesses them, and then calls the cloud large model interface through the network to complete the calculation and provide real-time feedback. In addition, the module has a built-in local AI large model that can quickly process simple tasks locally, reducing the dependence on the cloud. At the same time, it provides personalized services in combination with user preferences, significantly improving the convenience and intelligence level of interaction.

[0053] The signal management module is responsible for processing the signal transmission between the system internal and external devices, converting the signals of different interfaces into a format adapted to internal processing, and distributing the signals according to the device connection status and user requirements. It enhances and optimizes the signals to reduce attenuation and interference during transmission, ensuring the stability and compatibility of the signals, and solving the problem of unstable signal transmission in existing devices;

[0054] The communication module automatically establishes and maintains a stable network connection, securely transmits the user instructions to the cloud, and encrypts the data to protect user privacy and data security, solving the problems of unstable network connection and insufficient data transmission security in existing interaction methods;

[0055] The power management module provides stable power support for the entire system and the devices connected thereto, dynamically allocates power according to the device type, converts the input power into voltages and currents suitable for the system and devices, monitors the power status in real time, and automatically enters the low-power mode when the device is idle to save energy, solving the problems of low power management efficiency and high energy consumption in existing devices.

[0056] Furthermore, the AI interaction function module includes an identification unit, a software management unit, an interaction information processing unit, a cloud communication unit, a feedback unit, and a multimodal interaction unit;

[0057] The identification unit detects the connection status between the host device and the storage medium of the AI interaction system and determines whether it is the first connection;

[0058] The software management unit manages the installation, update, and startup of AI software;

[0059] The interaction information processing unit receives and processes the interaction instructions input by the user;

[0060] The cloud communication unit conducts network connection and data transmission with the cloud large model;

[0061] The feedback unit displays the results returned by the cloud large model in the AI software in real time and collects the feedback from the user;

[0062] The multimodal interaction unit supports and expands multimodal interaction capabilities.

[0063] The interaction information processing unit supports multiple input methods such as voice, button, touch, and gesture recognition, preprocesses the instructions input by the user, fuses multiple input signals to generate unified interaction instructions, and caches the user instructions for viewing or modification at any time, improving the diversity and flexibility of interaction and solving the problems of single input method and low instruction processing efficiency in existing interaction methods;

[0064] The multimodal interaction unit reserves interfaces and modules, supports newly added input methods in the future, allows users to seamlessly switch between different input methods, automatically selects the optimal input modality according to the scenario, optimizes the interaction logic and algorithms through cross-modal fusion technology, integrates emotion recognition technology, adjusts the interaction method according to the user's emotion, and optimizes the interaction performance through continuous learning, solving the problems of insufficient multimodal fusion and unnatural interaction in existing interaction methods.

[0065] Furthermore, the interaction information processing unit supports multiple input methods including voice recognition, button input, touch input, and gesture recognition. The interaction information processing unit performs preprocessing operations on the instructions input by the user to ensure that the instructions can be correctly understood by the cloud large model;

[0066] In some embodiments, the interaction information processing unit supports multiple or one input method, including but not limited to speech recognition, button input, touch input, and gesture recognition.

[0067] The interaction information processing unit fuses multiple input signals according to the instructions simultaneously input by the user through multiple input methods, and generates a unified interaction instruction.

[0068] Cache the instructions input by the user, and support the user to view or modify historical instructions at any time.

[0069] Further, the process of the interaction information processing unit performing preprocessing operations on the instructions input by the user is specifically as follows: receive the instructions issued by the user through various input methods, convert the collected input signals into a unified format, where convert the voice signal into text, remove background noise, and use natural language processing technology to perform a preliminary analysis on the text to extract key information.

[0070] Convert the gesture actions into corresponding instructions and remove interference signals.

[0071] Format the input text, remove extra spaces and punctuation marks, and perform a preliminary grammar check to ensure the logic of the text.

[0072] Further, the multimodal interaction unit reserves interfaces and modules to support future newly added input methods, supports seamless switching between different input methods by the user, and automatically selects the optimal input modality according to the scenario. For multimodal input, optimize the interaction logic and algorithms, and automatically adapt to the supported input modalities according to the hardware capabilities of different devices.

[0073] Further, the process of the multimodal interaction unit optimizing the interaction logic and algorithms for multimodal input is specifically as follows: sense and process the signals from different input modalities, convert the multimodal input signals into a unified feature vector, and fuse them into a unified multimodal context model.

[0074] Based on the multimodal input fusion, the interaction unit judges the user's intention through intention recognition technology.

[0075] The multimodal interaction unit generates an optimal behavior sequence according to the result of intention recognition.

[0076] Integrate emotion recognition technology to adjust the interaction method according to the user's emotional state.

[0077] Through cross-modal fusion technology, comprehensively process the input signals of different modalities, improve the interaction performance through continuous learning and optimization algorithms, and generate multimodal output signals according to the optimized interaction logic and feedback them to the user.

[0078] In one embodiment, the AI interaction function module also has a local AI large model built in. When the AI software is started, the local AI large model is automatically loaded. Combining the natural language processing ability of the local AI large model, the user intention is quickly recognized. Combining the local data and preferences of the user, the local AI large model provides personalized interaction results.

[0079] Furthermore, the AI interaction function module communicates with the cloud large model, regularly synchronizing the data and parameters of the local AI large model and the cloud large model. When more powerful computing power or access to cloud resources is required, the local AI large model collaborates with the cloud large model through the cloud communication unit, sending complex tasks to the cloud for processing;

[0080] The AI interaction function module optimizes the operation efficiency of the local AI large model according to the hardware capabilities of the device;

[0081] According to the user's real-time feedback and interaction history, the parameters and strategies of the local AI large model are dynamically adjusted;

[0082] The local AI large model optimizes its own performance by continuously learning the user's input and feedback.

[0083] Furthermore, the power management module provides stable power support for the AI interaction system storage medium and the devices connected thereto, and manages the distribution and conversion of power, dynamically allocating power according to the types of connected devices, converting the input power into voltages and currents suitable for the internal components of the AI interaction system storage medium and the connected devices, and real-time monitoring the power status, automatically entering the low-power mode when the device is idle.

[0084] In one embodiment, when the user connects the AI interaction system (such as an expander) to the Host device, the recognition unit automatically detects the connection status and determines whether it is the first connection. If it is the first connection, the software management unit will automatically jump to the download page of the AI software to complete the automatic download and installation of the software; if it is not the first connection, the installed AI software will be directly started, and this process does not require manual operation by the user, significantly improving the convenience of interaction;

[0085] The user inputs interaction instructions through the AI software. The input methods can be voice, keys, touch, gestures, etc. After receiving these instructions, the interaction information processing unit preprocesses the input signals. For example, voice signals will be converted into text and background noise will be removed; gesture actions will be converted into corresponding instructions and interference signals will be removed; text input will be formatted and grammar-checked. The preprocessed instructions are converted into a unified format to ensure that they can be correctly understood by the cloud large model;

[0086] If the user issues commands through multiple input methods (such as voice and gestures) simultaneously, the multimodal interaction unit will fuse these signals to generate a unified interaction command. It also supports seamless switching between different input methods and automatically adapts to the supported input modalities according to the hardware capabilities of the device. In addition, through continuous learning and optimization algorithms, the multimodal interaction unit continuously improves the interaction performance, making the interaction more natural and intelligent;

[0087] The preprocessed user command is first sent to the cloud communication unit, which is responsible for establishing a network connection with the cloud large model and securely transmitting the command to the cloud for calculation. At the same time, the local AI large model will also participate in the processing, combining the user's local data and preferences to quickly generate a preliminary result. For complex tasks, the local model will send the task to the cloud large model through the cloud communication unit for further processing. The collaborative work between the local and the cloud ensures the efficiency and accuracy of the interaction;

[0088] After the cloud large model finishes processing, it returns the calculation result to the AI interaction system. The feedback unit displays the result in the AI software in real time, and the user can immediately see the interaction result. At the same time, the feedback unit will collect the user's feedback information (such as confirmation, modification, or re-request) and send this information back to the cloud large model for optimizing the subsequent interaction logic. The local AI large model will also dynamically adjust its own parameters and strategies according to the user's real-time feedback and interaction history, continuously optimizing the performance;

[0089] During the entire interaction process, the power management module provides stable power support for the AI interaction system and its connected devices, dynamically allocates power according to the device type, and automatically enters the low-power mode when the device is idle to save energy. The signal management module is responsible for processing the signal transmission between the internal system and external devices, converting the signals of different interfaces into a format suitable for internal processing, and enhancing and optimizing the signals to ensure stable signal transmission;

[0090] The local AI large model continuously optimizes its own performance by continuously learning the user's input and feedback. At the same time, the system will regularly synchronize the data and parameters of the local model with the cloud large model to ensure that the local model can be updated in a timely manner, improving its accuracy and intelligence level. This dynamic optimization mechanism enables the system to continuously evolve according to the user's usage habits and preferences, providing a more personalized and efficient interaction experience.

[0091] Embodiment 2

[0092] This embodiment provides a storage medium for an AI interaction system, including an expander. A number of interfaces for connecting to a host device are provided on the expander. A number of buttons for interacting with AI software are also provided on the expander. A central processing unit for carrying the AI interaction system is built into the expander to implement the operation of the AI interaction system;

[0093] An interaction hardware is also provided on the expander, which may include but is not limited to a microphone, a camera, and a touch screen, supporting multiple instruction input methods.

[0094] In some embodiments, the expander can be a docking station. The expander is connected to the Host device by setting multiple interfaces, and these interfaces can also be connected to other external devices, playing a role in connecting external devices to the host. At the same time, it is equipped with interaction buttons, a microphone, a camera, and a touch screen, supporting multiple input methods such as voice, image, and gesture. The built-in central processing unit carries a complete AI interaction system, which can automatically identify devices, install or start AI software, process user input instructions, and interact with the cloud large model. This design not only improves the convenience and intelligence level of interaction but also provides users with an efficient, flexible, and personalized AI interaction experience through multi-modal input methods and the collaborative work of local and cloud, while reducing the threshold for users to use AI technology and promoting the popularization and application of AI technology in more scenarios.

[0095] The specific embodiments of the invention have been described in detail above, but they are only examples. The invention is not limited to the specific embodiments described above. Those skilled in the art should understand that the above embodiments and the descriptions in the specification only illustrate the principles of the invention. Without departing from the spirit and scope of the invention, the invention will have various changes and improvements, and these changes and improvements fall within the scope of the claimed invention. The scope of the invention claimed is defined by the appended claims and their equivalents.

Claims

1. An AI interactive system, characterized in that: Including AI interactive function module, power management module, signal management module and communication module; The AI ​​interactive function module automatically identifies the host device, installs or opens the AI ​​software on the host device, receives the interactive instructions input by the user in the AI ​​software, performs preliminary processing on the instructions, calls the cloud large model interface through the network connection, sends the processed user instructions to the cloud for calculation, receives the calculation results returned by the cloud large model, and displays them in real time in the AI ​​software, thus completing the closed loop of user interaction; The signal management module processes and manages the signal transmission within the AI ​​interactive system storage medium and between the AI ​​interactive system and external devices; The communication module is responsible for the network connection and data transmission between the storage medium of the AI ​​interactive system and the cloud-based large model, automatically establishes and maintains the network connection with the cloud server, and securely transmits the user input instructions to the cloud-based large model interface.

2. An AI interactive system according to claim 1, characterized in that: The AI ​​interactive function module includes an identification unit, a software management unit, an interactive information processing unit, a cloud communication unit, a feedback unit and a multimodal interaction unit; The identification unit detects the connection status between the host device and the storage medium of the AI ​​interaction system, and determines whether it is the first connection; The software management unit manages the installation, update and startup of the AI ​​software; The interactive information processing unit receives and processes the interactive instructions input by the user; The cloud communication unit is connected to the cloud large model through a network and performs data transmission; The feedback unit displays the results returned by the cloud-based large model in real time in the AI ​​software and collects user feedback; The multimodal interaction unit supports and expands multimodal interaction capabilities.

3. An AI interactive system according to claim 2, characterized in that: The interactive information processing unit supports multiple input methods, and performs preprocessing operations on the instructions input by the user to ensure that the instructions can be correctly understood by the cloud-based large model; The interactive information processing unit generates a unified interactive instruction by fusing multiple input signals according to the instructions inputted simultaneously by the user through multiple input methods; Cache the commands entered by the user, allowing the user to view or modify historical commands at any time.

4. The AI ​​interactive system according to claim 3, characterized in that: The interactive information processing unit performs preprocessing operations on the instructions input by the user, specifically, receiving instructions issued by the user through various input methods, converting the collected input signals into a unified format, wherein the voice signal is converted into text, and background noise is removed, and the text is preliminarily analyzed using natural language processing technology to extract key information; Convert gestures into corresponding commands and remove interference signals; Format the input text, remove extra spaces and punctuation, and perform a preliminary grammar check to ensure the logic of the text.

5. The AI ​​interactive system according to claim 2, characterized in that: The multimodal interaction unit reserves interfaces and modules to support input methods that will be added in the future, supports users to switch seamlessly between different input methods, and automatically selects the optimal input mode according to the scenario. It optimizes the interaction logic and algorithm for multimodal input, and automatically adapts the supported input modes according to the hardware capabilities of different devices.

6. The AI ​​interactive system according to claim 5, characterized in that: The multimodal interaction unit optimizes the interaction logic and algorithm for the multimodal input by sensing and processing the signals from different input modes, converting the multimodal input signal into a unified feature vector, and fusing it into a unified multimodal context model; Based on the fusion of multimodal input, the interaction unit uses intention recognition technology to determine the user's intention; The multimodal interaction unit generates the optimal behavior sequence based on the results of intention recognition; Integrate emotion recognition technology to adjust the interaction method according to the user's emotional state; Through cross-modal fusion technology, input signals of different modalities are comprehensively processed, and the interactive performance is improved through continuous learning and optimization algorithms. Based on the optimized interaction logic, multimodal output signals are generated and fed back to users.

7. The AI ​​interactive system according to claim 1, characterized in that: The AI ​​interaction function module also has a built-in local AI big model. When the AI ​​software is started, the local AI big model is automatically loaded. Combined with the natural language processing capabilities of the local AI big model, the user's intentions can be quickly identified. Combined with the user's local data and preferences, the local AI big model provides personalized interaction results.

8. An AI interactive system according to claim 7, characterized in that: The AI ​​interaction function module communicates with the cloud big model, regularly synchronizes the data and parameters of the local AI big model and the cloud big model, and when more powerful computing power or access to cloud resources is needed, the local AI big model works with the cloud big model through the cloud communication unit to send complex tasks to the cloud for processing; The AI ​​interaction function module optimizes the operating efficiency of the local AI large model according to the hardware capabilities of the device; Dynamically adjust the parameters and strategies of the local AI model based on real-time feedback and interaction history from users; The local AI large model optimizes its own performance by continuously learning from user input and feedback.

9. The AI ​​interactive system according to claim 1, characterized in that: The power management module provides stable power support for the AI ​​interactive system storage medium and its connected devices, manages the distribution and conversion of power, dynamically allocates power according to the type of connected device, converts input power into voltage and current suitable for the internal components of the AI ​​interactive system storage medium and connected devices, monitors the power status in real time, and automatically enters low power mode when the device is idle.

10. An AI interactive system storage medium, characterized in that: The device comprises an expander, the expander is provided with a plurality of interfaces for connecting to a host device, the expander is also provided with a plurality of buttons for interacting with AI software, the expander is built with a central processor for carrying any one of the AI ​​interactive systems described in claims 1-8, and realizes the operation of any one of the AI ​​interactive systems described in claims 1-8; The expander is also provided with interactive hardware to support multiple command input methods.

Images