Personalized upgrading method and system for smart television

Abstract

The application discloses a personalized upgrading method and system of an intelligent television, and relates to the technical field of intelligent television, comprising a cloud platform and an intelligent television. The cloud platform constructs a user portrait based on user behavior data reported by the intelligent television. When there is a component to be upgraded, the function feature information of the component is matched with the user portrait to determine a target user. The value index of the target user to the upgrade and the risk index based on the device state data of the target user are calculated respectively. After joint decision, if the preset upgrading condition is met, a personalized upgrading package is generated and pushed to the intelligent television of the target user. The intelligent television is responsible for receiving the personalized upgrading package pushed by the cloud platform and performing personalized upgrading. The application can realize safe and accurate personalized system upgrading.

Description

Technical Field

[0001] This application relates to the field of smart TV technology, and in particular to a personalized upgrade method and system for smart TVs. Background Technology

[0002] Current OTA (Over-the-Air) upgrade solutions for smart TVs, whether pre-set unified pushes by device manufacturers, intelligent decisions based on device (e.g., smart TV) operating status, or simple adaptations based on device hardware attributes, all suffer from relatively singular decision-making dimensions. These upgrade solutions either focus solely on the device's own stability or compatibility, or only consider static attributes such as hardware configuration, lacking consideration for users' personalized usage needs. More importantly, when deciding whether to push a system component upgrade to a specific user's device, existing solutions struggle to balance creating value for the user (i.e., providing new features) with controlling systemic risks (i.e., the risk of upgrades causing anomalies in specific device environments).

[0003] Therefore, existing upgrade solutions cannot achieve secure and accurate personalized system upgrades. Summary of the Invention

[0004] The main purpose of this application is to provide a personalized upgrade method and system for smart TVs, aiming to solve the technical problem that existing upgrade solutions cannot achieve safe and accurate personalized system upgrades.

[0005] To achieve the above objectives, this application proposes a personalized upgrade method for smart TVs, applied to a cloud platform, the method comprising:

[0006] In response to the component to be upgraded, the functional feature information of the component to be upgraded is obtained. Based on the matching degree between the functional feature information and each user profile, users whose matching degree reaches a preset matching degree threshold are identified as target users. The user profile is constructed based on the user behavior data of each user. The value indicators of the component to be upgraded to the target user are determined, and based on the device status data corresponding to the target user, the risk indicators of upgrading the component to be upgraded on the target user's smart TV are determined, wherein the user behavior data and the device status data are obtained through reporting by the smart TV; Based on the value indicators and the risk indicators, the upgrade decision is obtained; When the upgrade decision result meets the preset upgrade conditions, a personalized upgrade package containing the components to be upgraded is generated, and the personalized upgrade package is pushed to the target user's smart TV so that the smart TV can perform a personalized upgrade based on the personalized upgrade package.

[0007] In one embodiment, the step of obtaining the functional characteristic information of the component to be upgraded in response to the component to be upgraded includes: Feature extraction is performed on the user behavior data of each user to obtain a preference feature vector representing the user's usage preferences; Cluster analysis is performed on the preference feature vectors, and each user is divided into multiple user groups based on the clustering results of the cluster analysis; Based on the common behavioral characteristics of users within each user group, group profile tags reflecting these common behavioral characteristics are generated, and these group profile tags are used as user profiles.

[0008] In one embodiment, the step of determining the value index of the component to be upgraded for the target user includes: Obtain the functional feature vector corresponding to the functional feature information; Extract the target user's preference feature vector from the user profile; Calculate the similarity between the functional feature vector and the preference feature vector, and use the similarity result as the value indicator of the component to be upgraded for the target user.

[0009] In one embodiment, the step of determining the risk indicators for upgrading the component to be upgraded on the target user's smart TV based on the device status data corresponding to the target user includes: Extract the stability parameters of the smart TV of the target user and the success rate parameters of the component to be upgraded in historical upgrades from the device status data corresponding to the target user; The stability parameter and the success rate parameter are used as input features and input into a pre-trained risk prediction model. The success rate of the upgrade output by the risk prediction model is used as the risk indicator.

[0010] In one embodiment, the step of obtaining the upgrade decision result based on the value indicator and the risk indicator includes: Obtain a preset decision function, wherein the decision function is used to perform a weighted summation of the value indicator and the risk indicator, and the parameters of the decision function are adaptively optimized based on historical upgrade feedback data; The value index and the risk index are input into the decision function to calculate the comprehensive decision score, and the comprehensive decision score is used as the upgrade decision result.

[0011] In one embodiment, the step of generating a personalized upgrade package containing the component to be upgraded includes: Obtain a preset component dependency graph, wherein the component dependency graph is used to represent the dependency relationships and version constraints between various system components in the smart TV operating system; Based on the component dependency graph, the set of components that the component to be upgraded depends on is obtained by parsing. The components to be upgraded and their dependent component sets are packaged together to generate a personalized upgrade package.

[0012] Furthermore, to achieve the above objectives, this application also proposes a personalized upgrade method for smart TVs, applied to smart TVs, the method comprising: The system receives a personalized upgrade package pushed by the cloud platform and performs a personalized upgrade based on the personalized upgrade package. The personalized upgrade package is generated by the cloud platform after joint decision-making based on value indicators and risk indicators. The value indicators are determined based on user profiles constructed from user behavior data collected by the smart TV, and the risk indicators are determined based on device status data collected by the smart TV.

[0013] In one embodiment, the step of performing a personalized upgrade based on the personalized upgrade package includes: Install the personalized upgrade package to the inactive partition of the smart TV to update the inactive partition to the new version partition; Restart the smart TV and switch to the new version partition, and monitor the preset key operating indicators within the preset monitoring window period using the health probe; If the key operating indicators are detected to exceed the preset threshold during the monitoring window period, an automatic rollback will be triggered, the smart TV will be restarted and restored to the active partition before the personalized upgrade, and the rollback log will be reported to the cloud platform. If the key operational indicators are not detected to exceed the preset threshold during the monitoring window period, the personalized upgrade is determined to be successful, an upgrade success log is generated, and the upgrade success log is reported to the cloud platform.

[0014] In one embodiment, prior to the step of receiving the personalized upgrade package pushed by the cloud platform, the method includes: Collect user behavior data and device status data; The user behavior data and the device status data are reported to the cloud platform.

[0015] Furthermore, to achieve the above objectives, this application also proposes a personalized upgrade system for smart TVs, the system comprising a smart TV and a cloud platform: The smart TV includes: a data acquisition and reporting module and an upgrade execution module; The cloud platform includes: a data processing and modeling platform and decision-making and packaging services.

[0016] In addition, to achieve the above objectives, this application also proposes a smart TV, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the personalized upgrade method for the smart TV as described above.

[0017] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the personalized upgrade method for smart TVs as described above.

[0018] In addition, to achieve the above objectives, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the personalized upgrade method for a smart TV as described above.

[0019] One or more technical solutions proposed in this application have at least the following technical effects: By collecting and reporting user behavior data from smart TVs on a cloud platform, user profiles reflecting personalized usage preferences are constructed. When a component needs to be upgraded, the cloud platform obtains the functional characteristics of that component and calculates its matching degree with each user profile. Users whose matching degree reaches a preset matching degree threshold are accurately selected as target users, thus realizing the transformation of upgrade push from unified push to precise push based on user needs.

[0020] Building on this foundation, the cloud platform doesn't immediately push the upgrade after matching target users. Instead, it further considers the value and risk for each target user and their specific device. Specifically, the cloud platform determines the value metrics of the upgrade for the target user and the risk metrics of the target user's device status data. Based on these value and risk metrics, it makes a joint decision on whether to push the upgrade. This process transforms the previously difficult-to-balance user benefits and system risks into quantifiable value and risk metrics, thus enabling a trade-off between creating new functional value for users and controlling the potential system risks of the upgrade—two goals that were previously difficult to reconcile.

[0021] When the decision meets the preset upgrade conditions, the cloud platform generates and pushes a personalized upgrade package containing the component to be upgraded to the target user's smart TV, ensuring the accuracy of the upgrade. The smart TV receives this personalized upgrade package and executes the upgrade, completing the personalized system update for the target device.

[0022] Through the above approach, the technical solution of this application considers both the personalized needs of users reflected by user behavior data and the actual state of the device reflected by device status data when deciding whether to push a system component upgrade to a specific user's device. This allows the upgrade decision to effectively balance providing new functions to users with avoiding system anomalies caused by the upgrade, solving the problem of the existing technology having a single upgrade strategy and being unable to effectively control system risks while meeting personalized needs. Ultimately, it can achieve safe and accurate personalized system upgrades. Attached Figure Description

[0023] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0024] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 A first flowchart illustrating an embodiment of the personalized upgrade method for smart TVs according to this application; Figure 2 A second flowchart illustrating an embodiment of the personalized upgrade method for smart TVs according to this application; Figure 3 A schematic diagram of the third process provided for an embodiment of the personalized upgrade method for smart TVs in this application; Figure 4 A schematic diagram of the fourth process provided in an embodiment of the personalized upgrade method for smart TVs according to this application; Figure 5 A fifth flowchart illustrating an embodiment of the personalized upgrade method for smart TVs according to this application; Figure 6 This is a schematic diagram of the personalized upgrade system for the smart TV in this application.

[0026] Label 100 represents the smart TV, label 110 represents the data collection and reporting module, label 120 represents the upgrade execution module; label 200 represents the cloud platform, label 210 represents the data processing and modeling platform, and label 220 represents the decision-making and packaging service.

[0027] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0028] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of this application and are not intended to limit this application.

[0029] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0030] With the widespread adoption of smart TVs and other smart devices, online upgrades have become a core technology for device manufacturers to fix system vulnerabilities, optimize performance, and release new features. OTA upgrades primarily include full-device upgrades (downloading the complete firmware package) and component upgrades (updating only certain modules). How to efficiently, safely, and intelligently select and execute upgrades, balancing upgrade risks while meeting users' personalized needs, is key to improving user experience and operational efficiency, and is also the core demand of current smart TV OTA technology development. Currently, mainstream OTA upgrade solutions can be broadly categorized into three types: The first type: Device manufacturers push the update package uniformly according to a preset strategy. Based on the device model and firmware version, the device manufacturer pushes the same update package to all eligible users. This is currently the most widely used basic update solution, such as the "full push" or "batch fixed-time push" mode adopted by most smart TV manufacturers in the early days.

[0031] The second category involves intelligent decision-making based on the operating status of devices (such as smart TVs). For example, CN120255924A discloses a method that calculates a "mirror component quality score" and a "system software quality score" by monitoring the operating status (such as the number and level of anomalies) of various components in a smart device. Based on these two quality scores and preset thresholds, a decision is made to upgrade either the entire system (if the overall system quality score is too low, the entire system is upgraded) or individual components (if the overall system quality is acceptable but the quality scores of individual components are too low, the individual components are upgraded). Alternatively, CN115002546A discloses a method that analyzes the compatibility between the device's hardware configuration and the upgrade package to select suitable devices for upgrade delivery.

[0032] The third category: simple adaptation and upgrade based on device hardware attributes. For example, CN103747315A discloses the method of pushing out corresponding version upgrade packages based on the basic attributes of the TV (such as size and resolution).

[0033] However, the above-mentioned upgrade solutions have the following limitations: 1. Adopting a uniform upgrade strategy means that all users, regardless of their usage habits, application preferences, or device operating status, receive the exact same firmware upgrade package. This ignores the huge differences in user needs, and the upgrade strategy is too simplistic to provide targeted optimization for specific user groups, resulting in poor accuracy.

[0034] 2. Current upgrades primarily aim to passively fix known system crashes, application unresponsiveness, and other stability issues. This means that upgrades are only pushed out when stability problems are detected in components or the system, based on monitoring device operation (such as the number of anomalies and crash logs). This passive, fixative approach lacks the ability to proactively optimize functionality and improve user experience based on potential user needs.

[0035] 3. When pushing out existing upgrade packages, the decision-making is singular. They are usually pushed out in full or in batches based on a single dimension such as device model and version number. When there are compatibility or stability issues in the upgrade package that have not been found in the test (for example, when the upgrade package conflicts with a peripheral that is not popular but has a large user base), it will have a negative impact on a large number of users.

[0036] 4. For certain innovative functions or performance optimizations at the system's underlying level, due to the lack of a secure and controllable gray-scale verification environment, equipment manufacturers are very cautious when pushing out such updates in order to avoid the risk of upgrade failure due to unknown problems. This results in long verification cycles and high deployment risks for innovative functions, making it difficult to iterate quickly.

[0037] This application identifies target users by building profiles based on user behavior data on a cloud platform, and jointly quantifies and decides on upgrade value indicators and risk indicators respectively, thereby generating and pushing accurately matched personalized upgrade packages. This overcomes the technical problems of existing technologies, such as single upgrade strategies lacking personalization, passive decision-making processes, and inability to effectively balance value creation and risk control. As a result, it achieves the technical effect of proactive, accurate, and risk-controllable personalized system upgrades.

[0038] It should be noted that the executing entity in this embodiment can be a personalized upgrade system, such as a collaborative system composed of a smart TV and a cloud platform, or a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, mobile phone, server, etc., or an electronic device, IoT device, chip module, etc., capable of realizing the above functions. The following description uses a personalized upgrade system as an example to illustrate this embodiment and the subsequent embodiments.

[0039] Based on this, this application provides a personalized upgrade method for smart TVs, applied to a cloud platform, as described above. Figure 1 , Figure 1 This is a flowchart illustrating an embodiment of the personalized upgrade method for a smart TV according to this application. In this embodiment, the personalized upgrade method for the smart TV includes steps S1 to S4: Step S1: In response to the component to be upgraded, obtain the functional feature information of the component to be upgraded. Based on the matching degree between the functional feature information and each user profile, determine the users whose matching degree reaches the preset matching degree threshold as target users. The user profile is constructed based on the user behavior data of each user. Optionally, the components to be upgraded refer to the atomic modules that constitute the smart TV operating system that need to be updated or optimized, such as dynamic link libraries for implementing Dolby Vision decoding or GPU driver libraries for graphics processing.

[0040] Optionally, functional characteristic information refers to a quantitative description of the capabilities provided by the component to be upgraded. It can be a component function vector pre-annotated by the engineer according to the function of the component. For example, for the Dolby Vision decoding library, its function vector can be represented as picture quality 0.9, audio 0.6, and game 0.1.

[0041] Optionally, building a user profile refers to the process of extracting, transforming, and loading massive amounts of user behavior data in a batch-stream integrated manner on a cloud-based data processing and modeling platform, and then abstracting individual users into structured feature vectors through machine learning algorithms.

[0042] Optionally, as a preferred implementation method for constructing user profiles, the K-Means clustering algorithm can be used to aggregate users with similar behavioral patterns into multiple user clusters. This algorithm takes a multi-dimensional vector composed of behavioral data such as "usage time of the main application types in the past 30 days" of each user as input and divides all users into K clusters. The number of clusters K is not a fixed value. It can be found at the optimal inflection point between the accuracy and universality of the model by means of techniques such as elbow method and silhouette coefficient method, so as to ensure that each user cluster can accurately reflect the core common needs of a type of user and has a sufficient user scale to support subsequent personalized decision-making.

[0043] Optionally, based on the completion of user segmentation, a user preference vector is further generated for each user or each user cluster to quantify their interests. Each dimension of the user preference vector represents the degree of preference for a certain type of function or content, such as image quality, audio, or games. The specific values ​​can be 0.95 for image quality, 0.8 for audio, and 0.05 for games.

[0044] Optionally, matching degree refers to the process of identifying which users have potential value for the component to be upgraded by calculating the similarity between the functional feature information of the component to be upgraded and the user profile. Specifically, it can be achieved by calculating the cosine similarity between the component's functional vector and each user's user preference vector. When the matching degree is higher than a preset matching degree threshold, such as 80%, it can be said that the corresponding user is the target user for this upgrade.

[0045] Optionally, the target user refers to the user or user group that the personalized upgrade system determines to be highly relevant to the component to be upgraded after matching.

[0046] Step S2: Determine the value indicators of the component to be upgraded for the target user, and based on the device status data corresponding to the target user, determine the risk indicators of upgrading the component to be upgraded on the target user's smart TV. The user behavior data and device status data are obtained through reporting from the smart TV. In one feasible embodiment, the personalized upgrade system determines the value indicators and risk indicators of the component to be upgraded for the target user.

[0047] Optionally, the cloud platform can continuously monitor and acquire data sent from the data collection and reporting module on the smart TV side through a secure network transmission protocol.

[0048] Optionally, user behavior data refers to time-series data that records users' operating habits and content preferences on smart TVs. User behavior data can be structured information in JSON format collected based on the terminal event dictionary, such as application foreground running records and playback start records containing parameters such as timestamps, event identifiers, application package names, usage duration, resolution, and high dynamic range.

[0049] Optionally, device status data refers to quantitative indicators that reflect the health status of smart TV hardware and software. Device status data can be structured information in JSON format, such as native crash records, application unresponsive records, and hot frequency limiting records that include timestamps, event identifiers, component names, log hash values, and the number of occurrences.

[0050] Optionally, the value index is a value prediction score between 0 and 1, calculated by the value quantification model. It is used to predict and quantify the potential value or relevance of an upgrade component to a specific user. Specifically, the value quantification model can use the cosine similarity algorithm, and its output is the cosine similarity value between the user preference vector and the functional vector of the upgrade component. The closer the score is to 1, the better the functional characteristics of the component match the user's interests and preferences.

[0051] Optionally, the risk indicator is a stability success probability between 0 and 1, calculated by a risk quantification model. It is used to predict and quantify the success probability of performing this upgrade operation on a specific device. The risk quantification model preferably uses a logistic regression classifier. Its input feature vector includes, but is not limited to, the stability parameters of the smart TV and the success rate parameters of the component to be upgraded in historical upgrades, such as the current stability factor of the device, the historical rollback rate of the component to be upgraded, the device system-on-a-chip model code, the device random access memory size, the current system major version number code, and the device's time on the market. The output of the risk quantification model represents the degree of confidence that the personalized upgrade system can predict that this upgrade will be completed smoothly without causing new stability problems. The closer the probability value is to 1, the lower the risk.

[0052] Step S3: Based on value indicators and risk indicators, obtain the upgrade decision result; In one feasible embodiment, the personalized upgrade system obtains the upgrade decision result based on value indicators and risk indicators.

[0053] Optionally, the upgrade decision result is a comprehensive score calculated by a decision function. The role of this decision function is to weigh the two key dimensions of value indicators and risk indicators, avoiding two extreme situations: one-sidedly pursuing high value while ignoring high risk, or missing high-value opportunities due to excessive risk avoidance.

[0054] Alternatively, the decision function can take the form of a weighted summation model, with the expression Score = w1. P_match + w2 S_stability, where P_match represents the value index, S_stability represents the risk index, and w1 and w2 are the value weight and risk weight, respectively, which can also be called the parameters of the decision function; the personalized upgrade system will preset a decision threshold. Only when the final calculated comprehensive score is higher than the decision threshold will the upgrade be determined to be an operation with more benefits than harms and worth executing.

[0055] Step S4: When the upgrade decision result meets the preset upgrade conditions, a personalized upgrade package containing the components to be upgraded is generated, and the personalized upgrade package is pushed to the target user's smart TV so that the smart TV can perform a personalized upgrade based on the personalized upgrade package.

[0056] In one feasible embodiment, when the upgrade decision result meets the preset upgrade conditions, the personalized upgrade system generates a personalized upgrade package containing the components to be upgraded, and pushes the personalized upgrade package to the target user's smart TV, so that the smart TV can perform a personalized upgrade based on the personalized upgrade package.

[0057] Optionally, the preset upgrade conditions refer to the upgrade decision result, i.e., the comprehensive score, being greater than the preset decision threshold of the personalized upgrade system, and the hard security check passing, for example, the success probability of stability in the risk indicators, needing to be higher than a certain minimum tolerance threshold.

[0058] Optionally, generating a personalized upgrade package refers to the cloud platform's decision-making and packaging service generating an upgrade package that only contains the components required by the target user and their minimum dependencies.

[0059] Optionally, the personalized upgrade package can include a list of SHA-256 hashes of all files and be signed by the original equipment manufacturer.

[0060] Optionally, push refers to the process of securely and efficiently distributing the generated personalized upgrade package to the target user's smart TV via a content delivery network.

[0061] Optionally, personalized upgrades for smart TVs based on personalized upgrade packages refer to the upgrade execution module on the device side performing upgrades based on the personalized upgrade package after downloading the upgrade package.

[0062] This embodiment provides a personalized upgrade method for smart TVs. It constructs user profiles reflecting personalized usage preferences based on user behavior data collected and reported by the smart TV on a cloud platform. When a component needs upgrading, the cloud platform obtains the component's functional characteristics and calculates its matching degree with each user profile. Users whose matching degree reaches a preset matching degree threshold are accurately selected as target users, thus realizing a shift from unified upgrade push to precise push based on user needs.

[0063] Building on this foundation, the cloud platform doesn't immediately push the upgrade after matching target users. Instead, it further considers the value and risk for each target user and their specific device. Specifically, the cloud platform determines the value metrics of the upgrade for the target user and the risk metrics of the target user's device status data. Based on these value and risk metrics, it makes a joint decision on whether to push the upgrade. This process transforms the previously difficult-to-balance user benefits and system risks into quantifiable value and risk metrics, thus enabling a trade-off between creating new functional value for users and controlling the potential system risks of the upgrade—two goals that were previously difficult to reconcile.

[0064] When the decision meets the preset upgrade conditions, the cloud platform generates and pushes a personalized upgrade package containing the component to be upgraded to the target user's smart TV, ensuring the accuracy of the upgrade. The smart TV receives this personalized upgrade package and executes the upgrade, completing the personalized system update for the target device.

[0065] Through the above approach, the technical solution of this application considers both the personalized needs of users reflected by user behavior data and the actual state of the device reflected by device status data when deciding whether to push a system component upgrade to a specific user's device. This allows the upgrade decision to effectively balance providing new functions to users with avoiding system anomalies caused by the upgrade, solving the problem of the existing technology having a single upgrade strategy and being unable to effectively control system risks while meeting personalized needs. Ultimately, it can achieve safe and accurate personalized system upgrades.

[0066] In one feasible implementation, before step S1, which involves obtaining the functional characteristic information of the component to be upgraded in response to the component to be upgraded, the following steps are included: Step S11: Extract features from the user behavior data of each user to obtain a preference feature vector representing the user's usage preferences; Optionally, feature extraction refers to the process by which a cloud-based data processing and modeling platform converts raw, unstructured user behavior data, such as application foreground running records and playback start records, into structured numerical vectors.

[0067] Specifically, the personalized upgrade system will collect statistics on each user's usage time and frequency of various applications or functions within a specific period, such as the past 30 days. These behaviors will be quantified and normalized to form a multi-dimensional preference feature vector. Each dimension of this preference feature vector represents the intensity of the user's behavior towards a specific function or content category, such as video playback, game running, or music listening.

[0068] Step S12: Perform cluster analysis on the preference feature vectors, and divide each user into multiple user groups based on the clustering results of the cluster analysis; Alternatively, cluster analysis refers to the process of using unsupervised machine learning algorithms to take massive amounts of user preference feature vectors as input and automatically group users with similar behavioral patterns into the same cluster based on the spatial distance between user preference feature vectors.

[0069] In this way, previously independent individual users are aggregated into several user groups with common characteristics, thus solving the technical challenges of huge computational resource consumption and difficulty in modeling sparse features when performing detailed analysis on a single individual. The number of clusters formed by clustering, i.e., the number of user groups, can be determined as a variable hyperparameter using methods such as the elbow method and the profile coefficient method. The aim is to find the optimal balance between the accuracy and universality of the number of clusters.

[0070] Step S13: Generate group profile tags that reflect the common behavioral characteristics of users in each user group, and use the group profile tags as user profiles.

[0071] Optionally, generating group profile tags refers to extracting common behavioral patterns or significant features of all users within each predefined user group through statistical analysis, and summarizing and labeling them with easily understandable semantic tags. For example, a user group consisting of a large number of users who frequently watch 4K high-definition movies and Dolby Vision content would be given the group profile tag "Cinephile_4K_HDR"; while another user group consisting of a large number of users who play high-performance games for extended periods would be given the tag "Hardcore_Gamer".

[0072] Optionally, these group profile tags that can accurately reflect the common needs of a type of user constitute the user profile used for subsequent personalized decision-making.

[0073] In one feasible implementation, step S2, determining the value metrics of the component to be upgraded for the target user, includes: Step S21: Obtain the functional feature vector corresponding to the functional feature information; Optionally, the functional feature vector is a specific numerical representation of the functional feature information (i.e., the functional feature vector is obtained through the functional feature information). It can be pre-labeled and defined by engineers according to the functions implemented by the component to be upgraded. It is a vector that quantifies the capabilities of the component. For example, the functional feature vector of a component that optimizes Dolby Vision decoding can be represented as picture quality 0.9, audio 0.6, and game 0.1.

[0074] Step S22: Extract the target user's preference feature vector from the user profile; Optionally, the preference feature vector is a structured vector generated by extracting features from user behavior data during the process of building a user profile. It is used to represent the usage preferences of individual users and can be stored in the data structure of the user profile. It can quantify the user's preference for various functions or content, such as picture quality, audio, and games. For example, for a user who loves audio-visual products, their preference feature vector can be represented as picture quality 0.95, audio 0.8, and games 0.05.

[0075] Step S23: Calculate the similarity between the functional feature vector and the preference feature vector, and use the similarity result as a value indicator of the component to be upgraded for the target user.

[0076] Optionally, similarity calculation can employ a cosine similarity algorithm, which assesses the degree of matching by measuring the cosine of the angle between two vectors in their directions. The closer the similarity value is to 1, the more consistent the component's functional characteristics are with the user's interests and preferences. The calculated similarity result is a value prediction score between 0 and 1, quantifying the potential value of this upgrade for the target user. The higher the score, the greater the likelihood that the user will benefit from the component upgrade.

[0077] In one feasible implementation, step S2, which involves determining the risk indicators for upgrading the component to be upgraded on the target user's smart TV based on the device status data corresponding to the target user, includes: Step S24: Extract the stability parameters of the target user's smart TV and the success rate parameters of the components to be upgraded in historical upgrades from the device status data corresponding to the target user. Optionally, stability parameters refer to quantitative indicators that reflect the current health of the target device's hardware and software. Specifically, they may include the overall stability factor of the device, which is calculated by weighting the device status data reported by the end side (i.e., smart TV), such as various abnormal events, native crashes, application unresponsiveness, thermal throttling, etc. In addition, they may include static or dynamic characteristics related to stability, such as the device's system-on-a-chip model, random access memory size, system major version number, and device time on the market.

[0078] Optionally, the success rate parameter refers to a statistical indicator that measures the stability of the component to be upgraded during historical push processes. Its specific manifestations may include the rollback rate of the component to be upgraded in historical upgrades, that is, the proportion of the number of times that the probe detected abnormalities after the upgrade to the total number of pushes, as well as the average crash rate of the component to be upgraded on similar devices or globally.

[0079] Step S25: Input the stability parameter and success rate parameter as input features into the pre-trained risk prediction model; Optionally, the input feature refers to the multidimensional feature vector formed by combining all the extracted parameters, which is used as the basis for model prediction.

[0080] Optionally, the risk prediction model is a machine learning model that has been pre-trained through supervised learning. Its training process uses a massive amount of historical upgrade records. Each historical upgrade record contains the device status before the upgrade, the historical performance of the components, and the label of whether the upgrade was successful (i.e., no rollback was triggered and the system was stable). The model learns from these data to fit the mapping relationship between features and the probability of successful upgrade, thereby obtaining the risk prediction model, which is used to predict the probability of successful upgrade.

[0081] Step S26: Use the success probability of the upgrade output by the risk prediction model as a risk indicator.

[0082] Optionally, the success probability of the upgrade is a value between 0 and 1 output by the risk prediction model for the current input feature vector. This value represents the degree of confidence that the personalized upgrade system can successfully complete the component upgrade on the target user's smart TV without causing new stability problems. The closer the value is to 1, the lower the risk. The obtained probability value is the risk indicator used in the subsequent decision function.

[0083] In one feasible implementation, step S3, the step of obtaining the upgrade decision result based on value indicators and risk indicators, includes: Step S31: Obtain a preset decision function, wherein the decision function is used to perform a weighted summation of value indicators and risk indicators, and the parameters of the decision function are adaptively optimized based on feedback data from historical upgrades. Optionally, obtaining a preset decision function means calling a predefined mathematical model from the configuration center or model library of the personalized upgrade system. The function is to jointly evaluate the quantitative values ​​of two different dimensions: the value indicator representing the return and the risk indicator representing the risk.

[0084] Optionally, the parameters of the decision function are the value weight and the risk weight.

[0085] Optionally, the parameters of the decision function are not fixed values ​​set manually, but are determined by the reinforcement learning model based on long-term reward data from historical upgrades. This reinforcement learning model takes as input a state vector describing the health of the user group, such as the average user active time in the past 7 days, net promoter score, average app unresponsiveness rate, and average crash rate. It then uses a weighted combination of changes in comprehensive business indicators resulting from the adjustment of weight combinations, such as an increase in user retention rate, a decrease in app unresponsiveness rate, and a decrease in rollback rate, as a reward signal to continuously learn and find the optimal balance between creating value and controlling risk.

[0086] Step S32: Input the value index and risk index into the decision function, calculate the comprehensive decision score, and use the comprehensive decision score as the upgrade decision result.

[0087] Optionally, inputting value indicators and risk indicators into the decision function means substituting the calculated value prediction score and the calculated success probability of upgrading into the decision function as specific values, and obtaining the numerical result, i.e., the comprehensive decision score.

[0088] Optionally, this comprehensive decision score measures the expected benefits and potential risks of this upgrade. This comprehensive decision score is used as the upgrade decision result and compared with a preset upgrade threshold to determine whether to trigger the subsequent upgrade package generation and push process.

[0089] In one feasible implementation, step S4, generating a personalized upgrade package containing the components to be upgraded, includes: Step S41: Obtain a preset component dependency graph, wherein the component dependency graph is used to represent the dependency relationships and version constraints between various system components in the smart TV operating system; Optionally, the component dependency graph can be a knowledge graph managed using a directed acyclic graph data structure. It is predefined, where each node represents an independently upgradable system component, such as a dynamic link library file or a system service. The directed edges between nodes indicate the direction of dependencies between components; for example, the normal operation of component B depends on a specific version of component A. Simultaneously, the dependency graph also records the version evolution history of each component and the compatibility constraints between different versions.

[0090] Step S42: Based on the component dependency graph, parse out the set of components that the component to be upgraded depends on; Optionally, parsing refers to, once it's determined that an upgrade component needs to be pushed to a user, querying the component dependency graph. Starting from the node of the component to be upgraded, a depth-first or breadth-first traversal is performed along the directed edges to identify all other components that need to be upgraded or installed to ensure the correct operation of the current component. During the traversal, if version conflicts are encountered, such as two dependent components requiring different and incompatible versions of a common underlying component, preset priority rules are followed to automatically resolve the conflicts, such as security fixes taking precedence over the latest stable version, or the shortest dependency path taking precedence. The final output is a set of components that satisfy all dependencies without conflicts.

[0091] Step S43: Package the component to be upgraded and the set of components it depends on to generate a personalized upgrade package.

[0092] Optionally, packaging refers to the process of extracting and encapsulating the parsed set of components—the component to be upgraded and all its necessary dependencies—from the component repository. During packaging, the SHA-256 hash value of all files to be packaged is calculated, and a manifest file containing these hash values ​​is generated to ensure the integrity of the upgrade package during transmission and storage. Finally, the upgrade package containing all components and the manifest file is digitally signed using the original equipment manufacturer's root certificate to verify its authenticity and legitimacy. Upgrade packages generated in this way are highly personalized and secure because they contain only the target upgrade components for a specific user and their minimal necessary dependencies, avoiding the waste of network bandwidth and device storage resources associated with traditional full upgrade packages.

[0093] Based on the above embodiments of this application, in another embodiment of this application, the same or similar content as the above embodiments can be referred to the above description, and will not be repeated hereafter. This application provides a personalized upgrade method for a smart TV, applied to a smart TV. In this embodiment, the personalized upgrade method for a smart TV further includes: The system receives personalized upgrade packages pushed by the cloud platform and performs personalized upgrades based on these packages. These personalized upgrade packages are generated by the cloud platform based on a joint decision based on value indicators and risk indicators. The value indicators are determined based on user profiles built from user behavior data collected by the smart TV, and the risk indicators are determined based on device status data collected by the smart TV.

[0094] Optionally, receiving refers to the smart TV's upgrade execution module downloading a digitally signed upgrade package file generated by the cloud platform via a content delivery network.

[0095] Optionally, personalized upgrades based on personalized upgrade packages refer to the upgrade execution module following the Android A / B partitioning specification, securely installing the downloaded upgrade package into the currently unused inactive partition. After installation, only the boot flag needs to be modified to achieve a millisecond-level atomic switch to the new system upon the next reboot. Within a preset window period after switching to the new system, such as 15 minutes, a health probe background service is automatically started to continuously monitor key indicators such as startup time, core service liveness status, and whether consecutive severe crashes have occurred. Once any indicator exceeds a preset threshold that can be dynamically configured in the cloud, such as startup time exceeding 60 seconds, the boot flag will be automatically modified without user intervention, and a seamless rollback to the active partition before the upgrade will occur upon reboot, with a log containing detailed fault codes reported.

[0096] Optionally, the personalized upgrade package is generated by the cloud platform based on a joint decision based on value indicators and risk indicators. The value indicators are determined based on user profiles built from user behavior data, while the risk indicators are determined based on device status data. This ensures that the upgrade package pushed to the device is a set of personalized components that has been quantitatively evaluated and is beneficial to a specific user.

[0097] In one feasible implementation, the step of performing a personalized upgrade based on a personalized upgrade package includes: Step D11: Install the personalized upgrade package to the inactive partition of the smart TV to update the inactive partition to the new version partition; Optionally, a personalized upgrade package refers to a digitally signed file containing the components to be upgraded and their minimum dependency set, generated and pushed by the cloud platform after a joint decision based on value and risk indicators.

[0098] Optionally, an inactive partition refers to a separate, complete operating system partition in the storage of a smart TV that supports the A / B partitioning mechanism. This partition is independent of and does not interfere with the currently running operating system partition (i.e., the active partition). The installation process follows the Android A / B partitioning specification, securely writing and updating the components in the upgrade package to this inactive partition. During this process, the user can use the operating system in the currently active partition normally, achieving a seamless upgrade. After installation, the original inactive partition is updated to a new version partition containing the new component versions, but the operating system has not yet been switched.

[0099] Step D12: Restart the smart TV and switch to the new version partition, and monitor the preset key operating indicators through the health probe within the preset monitoring window period; Optionally, restarting and switching refers to modifying the flag of the operating system's boot partition, pointing the boot target to the updated new version partition, and then performing an operating system restart so that the smart TV can boot and run from the new version partition.

[0100] Optionally, the health probe is a background guardian service that runs automatically after the new system starts up. During a preset monitoring window, such as 15 minutes after the new system starts up, it continuously and frequently collects and calculates preset key operating indicators.

[0101] Optionally, key operational metrics are core parameters used to measure the basic health of the system, such as total operating system startup time, the survival status of core system services like system_server, and whether multiple consecutive severe crashes, such as more than three consecutive crashes, have occurred. The monitoring thresholds for these metrics, such as startup time exceeding 60 seconds, are preset and can be dynamically adjusted through cloud configuration. Their initial values ​​can be derived from statistical analysis of massive amounts of historical device data.

[0102] Step D13: If key operating indicators exceed preset thresholds during the monitoring window, an automatic rollback is triggered, the smart TV is restarted and restored to the activity partition before the personalized upgrade, and the rollback log is reported to the cloud platform. Optionally, triggering automatic rollback means that once the health probe detects any critical operational indicator, such as startup time or the number of consecutive crashes, exceeding a preset threshold issued by the cloud, it will immediately initiate a recovery process without user intervention. The overall process may be: modifying the boot loader flag to point the boot target back to the old active partition used before the upgrade, and then performing a system restart. After restarting, the smart TV will be restored to a stable and usable system state before the personalized upgrade, ensuring that users recover from potential system failures without noticing.

[0103] Optionally, reporting rollback logs means that when a rollback is triggered, the upgrade execution module will automatically generate a structured log containing detailed fault codes, abnormal indicator values, and the current system status, and report it to the cloud platform through a reliable network connection to provide negative sample data for subsequent problem analysis and model optimization.

[0104] Step D14: If the key operating indicators are not exceeded within the monitoring window period, the personalized upgrade is confirmed to be successful, an upgrade success log is generated, and the upgrade success log is reported to the cloud platform.

[0105] Optionally, if the health probe detects that all preset key operating indicators do not exceed the preset indicator thresholds within the preset monitoring window period, the smart TV's upgrade execution module will determine that the personalized upgrade has successfully passed the health check and is considered a successful upgrade.

[0106] Optionally, confirming a successful personalized upgrade means that the new version partition performs stably during the critical monitoring phase after startup, with all core indicators such as startup time, core service liveness, and crash frequency within the preset normal range, thus determining that the new system can run safely and stably continuously. After confirming a successful upgrade, the upgrade execution module automatically generates an upgrade success log, which records detailed information such as the upgraded component information, upgrade completion time, and measured values ​​of key indicators during the monitoring window. Finally, the smart TV will report the upgrade success log to the cloud platform via a reliable network connection, providing positive feedback samples for the cloud data processing and modeling platform. These samples will be used for the continuous iteration and optimization of subsequent value quantification models, risk quantification models, and reinforcement learning decision weights.

[0107] In one feasible implementation, prior to the step of receiving the personalized upgrade package pushed by the cloud platform, the following steps are included: Step D15: Collect user behavior data and device status data; Step D16: Report user behavior data and device status data to the cloud platform.

[0108] Optionally, data collection refers to the data collection and reporting module built into the smart TV, which monitors and records user actions and device operating status in real time.

[0109] Optionally, user behavior data can be structured JSON format information. For example, when a user opens a video application, the data collection and reporting module will record a piece of behavior data that includes a timestamp, an event identifier of APP_FOREGROUND, the application package name, and the resolution and high dynamic range parameters of the currently playing content.

[0110] Optionally, the device status data can be structured JSON format information. For example, when a system crash occurs at the underlying level, the data acquisition and reporting module will record status data that includes a timestamp, an event identifier of NATIVE_CRASH, the name of the component that crashed, the event hash value of the crash, and the cumulative number of crashes for that day.

[0111] To protect user privacy, the data collection and reporting module performs SHA-256 hashing and salting on sensitive fields such as application package names on the client side during collection to achieve irreversible anonymization, and performs local aggregation to remove identifiers for numerical data.

[0112] Optionally, after collecting user behavior data and device status data, the data collection and reporting module reports it to the cloud platform. Reporting refers to the data collection and reporting module transmitting the collected and preprocessed data to the cloud platform through a secure network connection, according to a certain strategy. To ensure the reliability of the reporting, the data collection and reporting module integrates a breakpoint resume function and adopts an exponential backoff strategy for failure retries. At the same time, to avoid data duplication, each reported data item contains an idempotent ID generated by a hash algorithm using a timestamp, event identifier, and device identifier.

[0113] For example, please refer to Figure 2 , Figure 2 This document provides a flowchart illustrating a personalized upgrade method for smart TVs. Specifically, the cloud platform first receives and processes user behavior data and device status data reported by the smart TV, constructing a user profile based on this data. When a component needs upgrading, the cloud platform obtains the component's functional characteristics and matches them with the user profile to identify potential beneficiary users. Subsequently, the cloud platform calculates the value index of the component to be upgraded for each target user, as well as the upgrade risk index based on the user's device status data. Next, the cloud platform performs a comprehensive evaluation based on the calculated value and risk indices to obtain the final upgrade decision. If the decision meets the preset upgrade conditions, the cloud platform generates a personalized upgrade package containing the component to be upgraded and accurately pushes it to the target user's smart TV to complete the personalized upgrade.

[0114] For example, please refer to Figure 3 , Figure 3This document provides a simplified flowchart of a personalized upgrade method for smart TVs. Specifically, after receiving a personalized upgrade package pushed by a cloud platform, the smart TV initiates the upgrade process. First, the smart TV begins receiving the personalized upgrade package. After receiving the package, a preset health probe is activated for continuous monitoring. The health probe checks the key operating indicators and status of the system in real time after the upgrade. If no abnormalities are found during monitoring, the upgrade is considered successful; if any abnormalities are detected, a rollback mechanism is immediately triggered, restoring the system to its stable state before the upgrade. Regardless of whether the final result is a successful upgrade or a rollback, each upgrade attempt is ensured to be completed within a safe and controllable closed loop.

[0115] For example, please refer to Figure 4 , Figure 4 This document provides a simplified flowchart of a personalized upgrade method for smart TVs, specifically executed collaboratively by a cloud platform and the smart TV. After receiving the upgrade package from the cloud platform and performing the upgrade, the smart TV enters the result verification phase: if the upgrade is successful, the smart TV reports a success log and behavioral data; if the upgrade fails or an anomaly occurs, a rollback mechanism is triggered to restore the system and report a rollback log. Regardless of success or rollback, the reported logs and data will serve as new feedback information, which will be received and processed again by the cloud platform to optimize subsequent decision-making models.

[0116] In the next upgrade, the cloud platform continues to receive and process data reported from smart TVs (user behavior data, device status data, etc.), and calculates the upgrade decision based on this. The cloud platform then determines whether the upgrade decision meets the upgrade conditions. If not, it enters a "waiting for the next decision" state and the upgrade is not executed; if it does meet the conditions, the cloud platform "generates and pushes the upgrade package" to the corresponding smart TV. This forms a complete technical closed loop from decision-making and execution to feedback optimization.

[0117] For example, refer to Figure 5 , Figure 5 A simplified schematic diagram of a personalized upgrade method for a smart TV is provided. Specifically, the smart TV 100 continuously collects local user behavior data and device status data through its data collection and reporting module 110, and reports it to the cloud platform 200, completing the data upload and aggregation (corresponding to...). Figure 5 1. Data Reporting). After receiving the data, the cloud platform 200 uses the data processing and modeling platform 210 to process and analyze the data, constructing and updating user profiles or decision models that reflect user preferences and device characteristics (corresponding to...). Figure 52. Providing model / profile data). When there are components to be upgraded, the decision-making and packaging service 220 calls upon this model and profile data to perform a quantitative assessment of value and risk and a joint decision-making process, and dynamically generates a personalized upgrade package for the target users whose decisions are approved. Subsequently, the cloud platform 200 pushes the personalized upgrade package to the corresponding smart TV 100 (corresponding to...). Figure 5 3. Pushing the upgrade package). The smart TV 100 receives and installs the upgrade package through its upgrade execution module 120, completes the personalized system upgrade, and reports the upgraded status or result data as closed-loop feedback information, thus forming a complete technical closed loop from data collection, cloud decision-making, precise push to end-side execution and feedback (corresponding to...). Figure 5 4. Closed-loop feedback.

[0118] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the personalized upgrade method of smart TVs in this application. Any simple changes based on this technical concept, such as the interaction and combination of various embodiments, are all within the protection scope of this application.

[0119] The following is for reference. Figure 6 It shows a structural schematic diagram of a personalized upgrade system (hereinafter referred to as the personalized upgrade system) suitable for implementing the embodiments of the present application for a smart TV. Figure 6 The personalized upgrade system shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0120] like Figure 6 As shown, the personalized upgrade system may include a smart TV 100 and a cloud platform 200. The smart TV 100 also includes a data acquisition and reporting module 110 and an upgrade execution module 120, wherein the data acquisition and reporting module 110 is used to: collect user behavior data and device status data; and report the user behavior data and device status data to the cloud platform; the upgrade execution module 120 is used to: receive a personalized upgrade package pushed by the cloud platform, and perform a personalized upgrade based on the personalized upgrade package.

[0121] The cloud platform 200 also includes a data processing and modeling platform 210 and a decision-making and packaging service 220. The data processing and modeling platform 210 is used to: receive user behavior data and device status data reported by the smart TV; construct a user profile based on the user behavior data; in response to a component to be upgraded, obtain the functional characteristic information of the component to be upgraded, match the user profile based on the functional characteristic information, and determine the target user; determine the value index of the component to be upgraded for the target user, and determine the risk index of upgrading the component to be upgraded on the target user's smart TV based on the device status data corresponding to the target user. The decision-making and packaging service 220 is used to: obtain an upgrade decision result based on the value index and the risk index; when the upgrade decision result meets preset upgrade conditions, generate a personalized upgrade package containing the component to be upgraded, and push the personalized upgrade package to the target user's smart TV, so that the smart TV can perform a personalized upgrade based on the personalized upgrade package.

[0122] While the diagram illustrates a customized upgrade system with various modules, it should be understood that implementing or having all of the modules shown is not required. Alternatively, more or fewer modules may be implemented.

[0123] The personalized upgrade system provided in this application, employing the personalized upgrade method for smart TVs described in the above embodiments, can solve the technical problem that existing upgrade schemes cannot achieve secure and accurate personalized system upgrades. Compared with the prior art, the beneficial effects of the personalized upgrade system provided in this application are the same as those of the personalized upgrade method for smart TVs provided in the above embodiments, and other technical features of this personalized upgrade system are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0124] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0125] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0126] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to execute the personalized upgrade method for a smart TV in the above embodiments.

[0127] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems or devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0128] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0129] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0130] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0131] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the personalized upgrade method for the smart TV described above. This solves the technical problem that existing upgrade schemes cannot achieve secure and accurate personalized system upgrades. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the personalized upgrade method for the smart TV provided in the above embodiments, and will not be repeated here.

[0132] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the personalized upgrade method for a smart TV as described above.

[0133] The computer program product provided in this application can solve the technical problem that existing upgrade solutions cannot achieve secure and accurate personalized system upgrades. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the personalized upgrade method for smart TVs provided in the above embodiments, and will not be repeated here.

[0134] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A method for personalized upgrades of a smart TV, characterized in that, Applied to a cloud platform, the method includes: In response to the component to be upgraded, the functional feature information of the component to be upgraded is obtained. Based on the matching degree between the functional feature information and each user profile, users whose matching degree reaches a preset matching degree threshold are identified as target users. The user profile is constructed based on the user behavior data of each user. The value indicators of the component to be upgraded to the target user are determined, and based on the device status data corresponding to the target user, the risk indicators of upgrading the component to be upgraded on the target user's smart TV are determined, wherein the user behavior data and the device status data are obtained through reporting by the smart TV; Based on the value indicators and the risk indicators, the upgrade decision is obtained; When the upgrade decision result meets the preset upgrade conditions, a personalized upgrade package containing the components to be upgraded is generated, and the personalized upgrade package is pushed to the target user's smart TV so that the smart TV can perform a personalized upgrade based on the personalized upgrade package.

2. The method as described in claim 1, characterized in that, Prior to the step of obtaining the functional characteristic information of the component to be upgraded in response to the component to be upgraded, the following is included: Feature extraction is performed on the user behavior data of each user to obtain a preference feature vector representing the user's usage preferences; Cluster analysis is performed on the preference feature vectors, and each user is divided into multiple user groups based on the clustering results of the cluster analysis; Based on the common behavioral characteristics of users within each user group, group profile tags reflecting these common behavioral characteristics are generated, and these group profile tags are used as user profiles.

3. The method as described in claim 1, characterized in that, The step of determining the value index of the component to be upgraded for the target user includes: Obtain the functional feature vector corresponding to the functional feature information; Extract the target user's preference feature vector from the user profile; Calculate the similarity between the functional feature vector and the preference feature vector, and use the similarity result as the value indicator of the component to be upgraded for the target user.

4. The method as described in claim 1, characterized in that, The step of determining the risk indicators for upgrading the component to be upgraded on the target user's smart TV based on the device status data corresponding to the target user includes: Extract the stability parameters of the smart TV of the target user and the success rate parameters of the component to be upgraded in historical upgrades from the device status data corresponding to the target user; The stability parameter and the success rate parameter are used as input features and input into a pre-trained risk prediction model. The success rate of the upgrade output by the risk prediction model is used as the risk indicator.

5. The method as described in claim 1, characterized in that, The steps for obtaining the upgrade decision result based on the value indicator and the risk indicator include: Obtain a preset decision function, wherein the decision function is used to perform a weighted summation of the value indicator and the risk indicator, and the parameters of the decision function are adaptively optimized based on historical upgrade feedback data; The value index and the risk index are input into the decision function to calculate the comprehensive decision score, and the comprehensive decision score is used as the upgrade decision result.

6. The method as described in claim 1, characterized in that, The step of generating a personalized upgrade package containing the components to be upgraded includes: Obtain a preset component dependency graph, wherein the component dependency graph is used to represent the dependency relationships and version constraints between various system components in the smart TV operating system; Based on the component dependency graph, the set of components that the component to be upgraded depends on is obtained by parsing. The components to be upgraded and their dependent component sets are packaged together to generate a personalized upgrade package.

7. A method for personalized upgrades of a smart TV, characterized in that, Applied to smart TVs, the method includes: The system receives a personalized upgrade package pushed by the cloud platform and performs a personalized upgrade based on the personalized upgrade package. The personalized upgrade package is generated by the cloud platform after joint decision-making based on value indicators and risk indicators. The value indicators are determined based on user profiles constructed from user behavior data collected by the smart TV, and the risk indicators are determined based on device status data collected by the smart TV.

8. The method as described in claim 7, characterized in that, The steps for performing a personalized upgrade based on the personalized upgrade package include: Install the personalized upgrade package to the inactive partition of the smart TV to update the inactive partition to the new version partition; Restart the smart TV and switch to the new version partition, and monitor the preset key operating indicators within the preset monitoring window period using the health probe; If the key operating indicators are detected to exceed the preset threshold during the monitoring window period, an automatic rollback will be triggered, the smart TV will be restarted and restored to the active partition before the personalized upgrade, and the rollback log will be reported to the cloud platform. If the key operational indicators are not detected to exceed the preset threshold during the monitoring window period, the personalized upgrade is determined to be successful, an upgrade success log is generated, and the upgrade success log is reported to the cloud platform.

9. The method as described in claim 7, characterized in that, Before the step of receiving the personalized upgrade package pushed by the cloud platform, the following steps are included: Collect user behavior data and device status data; The user behavior data and the device status data are reported to the cloud platform.

10. A personalized upgrade system for a smart TV, characterized in that, The system includes a smart TV and a cloud platform: The smart TV includes: a data acquisition and reporting module and an upgrade execution module; The cloud platform includes: a data processing and modeling platform and decision-making and packaging services.

Images