74 results about "User expectations" patented technology

Embodiments of the present application provide an operating parameter recommendation method and apparatus, used for in a big data processing scenario, recommending operating parameters for a computing device on the basis of the resource consumption under different operating parameters, without depending on historical data. The method comprises: first, on the basis of information obtained during the processing of historical tasks in a big data system, establishing a first relationship among the corresponding data information, parameters of a computing device processing the tasks, and time required to process the tasks, and a second relationship among the data information, the parameters, and resources of the computing device consumed for processing the tasks; inputting data information of a task to be processed, and the first relationship and the second relationship to a genetic algorithm, and outputting multiple sets of time values and resource values, i.e., modeling the process of the computing device executing data processing tasks, so as to fit the relationship between the time and resource consumption of the computing device and the operating parameters of the computing device; and determining operating parameters on the basis of actual available resources of the computing device or resource consumption that a user desires.

The invention relates to the technical field of artificial intelligence, in particular to a verbal skill confrontation training generation method and device based on a large model, equipment and a storage medium, and the method comprises the steps: generating a simulation inquiry based on a preset user portrait, then obtaining a response of a service large model, and comprehensively evaluating the response quality; and the specific defect type of the service model in logic understanding or verbal skill expression is diagnosed through the cognitive deviation between the response and the user expectation by another evaluation model, so that a targeted optimization strategy is generated, the service model is finally driven to carry out directional reinforcement learning, and the verbal skill ability of the large service model is iteratively improved. According to the method, an adversarial sample is dynamically generated, and deep diagnosis is carried out in combination with cognitive deviation, so that a service model can understand deep intentions and emotion demands of different user groups fundamentally. Therefore, the strain capacity and the communication accuracy of the service model in a real interaction scene are remarkably improved, and the user satisfaction is effectively improved.

The invention provides a method and device for making a diversified analysis chart based on a UE engine, and belongs to the technical field of charts.According to the embodiment of the invention, based on source data, a service index weight matrix conforming to a service scene can be obtained, and based on the source data and demand description information, the diversified analysis chart is made. The method comprises the following steps: obtaining a chart type weight matrix which accords with data characteristics and user expectations through a business index weight matrix and a business index weight matrix, further recommending an appropriate chart type to a user while realizing deep analysis of data by comprehensively considering the business index weight matrix and the chart type weight matrix, and further improving the user experience by converting a multi-dimensional matrix into executable parameters of a UE engine. According to the embodiment of the invention, the powerful computing power support of the cloud server and the powerful graphic rendering capability of the UE engine can be utilized to provide a more visual and comprehensive diversified analysis chart for the user, so that the display efficiency is improved, and the use experience of the user is remarkably improved.

The invention relates to the technical field of advertisement design, and discloses an AI virtual human interaction method and system, and the method comprises the steps: receiving original demand data inputted by a user, generating a brand feature vector, and generating first virtual human data, first line data and first product model data based on the brand feature vector; and performing cooperative adjustment based on the ith interaction instruction to generate a virtual human image. By implementing the method and the system, the functions of full-process intelligent design, data processing, collaborative generation, interactive response, scene adaptation and the like are improved, an efficient and complete virtual human image customization scheme is formed, and the problem that multi-source heterogeneous data are difficult to comprehensively analyze when user requirements are processed by a traditional method is solved. The problems of large deviation between generated content and user expectation and poor scene adaptability caused by lack of a collaborative adjustment mechanism of a virtual human image, lines and a product model, low processing efficiency of a modification instruction and incapability of ensuring cross-modal consistency in the prior art are solved.

The invention provides an adaptive test case generation method and system based on a large model, and the method comprises the steps: calling a pre-constructed cue word generation agent in response to an instruction for generating a test case, and generating a matched target cue word and a recommended target test case generation model based on data input by a user; generating a first test case based on the target cue word by using the target test case generation model; displaying the first test case, and obtaining a first label input by a user for the displayed first test case; the target test case generation model is called, a second test case is generated in combination with the user intention and the first test case, and the matching degree of the second test case and the test case expected by the user is higher than the matching degree of the first test case and the test case expected by the user.

The present disclosure relates generally to systems and methods for updating an input prompt for a generative AI model (e.g., an LLM) based on feedback that is provided in connection with an output from the generative AI model that is unsatisfactory. For example, where a user indicated that an output from the generative AI model is incorrect, inaccurate, or is an otherwise unsatisfactory response to an input prompt, this disclosure describes models to facilitate generation of feedback hints and / or additional information that can be included within an updated prompt that, when provided as an input to the generative AI model, has an improved likelihood to return an output that is in-line with user expectations. Indeed, features of the systems and methods described herein provide a framework for improving outputs of generative AI models that are more accurate or otherwise responsive to the input prompts.

The invention relates to the field of vehicles, and provides a vehicle-mounted man-machine interaction method, a vehicle-mounted man-machine interaction device and electronic equipment aiming at the problem of how to enable a vehicle-mounted terminal to accurately respond to the intention of a user and realize an application function expected by the user. The method comprises the steps that any one or a combination of more of vehicle data, driver data and environment and context data serves as a data source, a semantic text sequence is obtained according to the data source, and the semantic text sequence comprises one or more feature texts; the feature text is used for describing any one or a combination of more of a vehicle feature, a driver feature, an environment feature and a context feature, and the feature text is a text character string described by referring to a natural language; taking the semantic text sequence as the input of an intelligent model, and obtaining a prediction result output by the intelligent model; and calling the corresponding application service according to the prediction result. According to the method provided by the embodiment of the invention, the prediction accuracy of the intelligent model can be improved, and the user experience of man-machine interaction is improved.

The invention provides a perceptual engineering-based vehicle body selection design method. The perceptual engineering-based vehicle body selection design method comprises the following steps of: acquiring vehicle body data and comment information of a whole vehicle from data permitted to be accessed; analyzing and sorting the obtained comment text data by using a natural language processing technology to form a database; collecting user expectations, and performing semantic reasoning and analysis on the collected user expectations based on the formed database to obtain product performance semantic weights of the user expectations; and performing quantitative matching on the obtained product performance semantic weight expected by the user and the obtained vehicle body data by using the perceptual engineering model, and determining the vehicle body data corresponding to the vehicle model most conforming to the user expectation. And finally, the product with the highest customer expectation score is determined to correspondingly obtain the target product, the designed optimal product scheme is determined, and the vehicle body design efficiency and reliability are improved.

This application provides a model training method, training system, and related equipment. The method includes the following steps: acquiring positive sample data, wherein the positive sample data includes input samples and positive labels, the input samples include questions, and the positive labels include the answers corresponding to the questions; inputting the input samples into a first large model to be fine-tuned, obtaining the first negative label corresponding to the input sample; determining the first score of the first negative label based on the relevance of the first negative label to the question and the quality of the response to the first negative label based on a scoring model; inputting the input samples into the first large model; and fine-tuning the first large model based on a loss function to obtain a second large model. The loss function is used to guide the first large model to improve the prediction probability of positive labels and the prediction probability of negative labels with scores greater than a threshold, so that the model can not only learn standard answers that meet user expectations, but also generate more diverse answers, thereby improving the fine-tuning effect of the large model.

The invention discloses an LLM Agent task execution method and device and electronic equipment. The method comprises the following steps: S1, identifying a user role; s2, task planning and interaction point configuration; s3, performing active interaction execution; and S4, dynamically adjusting the execution path. According to the method, interaction can be actively initiated at key nodes, an execution strategy is dynamically optimized, personalized interaction experience is provided, task execution efficiency is effectively improved, the risk that an execution result deviates from user expectation is reduced, and system availability and user satisfaction are improved.

The invention relates to a man-machine interaction method and device, computer equipment, a storage medium and a computer program product. Relates to the technical field of cloud computing. The method comprises the steps of receiving a current interaction task input by a user, and obtaining historical interaction information of the user; detecting the similarity between the current interaction task and each historical interaction task of the user; if a target historical interaction task of which the similarity is greater than or equal to a preset threshold value is detected, obtaining a target tool calling list of the target historical interaction task; inputting the current interaction task, the target tool calling list and the user information into a large language model to obtain a first tool calling list for executing the current interaction task; according to the input parameters and the calling sequence of all the tools in the first tool calling list, tool calling operation is executed, and a response result of the current interaction task is obtained. By adopting the method, the targeted content matched with the user expectation can be generated.

The application discloses an intelligent AI marketing data management system based on a global member system, relates to the technical field of data management, and comprises a historical information collection module, a marketing group correlation module, a portrait deviation prediction module and a marketing real-time pushing module. The historical information collection module is used for generating corresponding marketing logs, collecting characteristic data generated by users and obtaining user portraits. The marketing group correlation module is used for analyzing the correlation between different user groups, analyzing the marketing data differences of different marketing logs and obtaining the correlation between different marketing types. The portrait deviation prediction module is used for analyzing the portrait deviation of users, making predictive adjustment and obtaining expected user portraits. The marketing real-time pushing module is used for generating corresponding real-time marketing logs in real time, making real-time prediction on the expected demands of users and providing corresponding personalized marketing recommendations for users.

This application relates to the field of vehicle-related technology and discloses a parking method, apparatus, device, and vehicle based on voice interaction. The method includes: responding to a user's voice parking command and extracting parking constraints; the parking constraints include at least one of parking area, parking location attribute features, and parking method, each parking constraint including positive and / or negative requirements; while the vehicle is driving in the parking lot, sensing environmental information around the vehicle and obtaining multi-dimensional environmental features of available parking spaces; the multi-dimensional environmental features include at least the area and location attribute features; filtering available parking spaces based on the multi-dimensional environmental features according to the priority of the parking constraints, and controlling the vehicle to park in the available parking space when a parking space that meets the parking constraints is selected. Applying the solution of this application can solve the problems of existing automatic parking systems failing to understand users' personalized parking needs and resulting in parking results that do not meet user expectations.

The present disclosure provides a method for optimizing large language model conversation, including: in response to entering a current conversation turn, acquiring configuration information corresponding to the current conversation turn based on a type of the current conversation turn; acquiring user input information corresponding to the current conversation turn; determining whether the current conversation turn needs to be switched based on a conversation turn switching determination strategy corresponding to the conversation turn switching strategy configuration information; when the current conversation turn needs to be switched, determining a next conversation turn corresponding to the current conversation turn, and entering the next conversation turn; when the current conversation turn does not need to be switched, generating a response task, and performing the response task calling a large language model to generate a response message; and outputting the response message in the current conversation turn, to meet user expectations.

This application provides a voice control method, apparatus, device, system, storage medium, and program product. On one hand, voice control based on user expectations offers a novel voice control mode. Users can express their personalized needs by conveying their expected physical or emotional feelings, thereby enabling automatic device control. This achieves intelligent and flexible voice control, embodies a technological feel, and satisfies users' personalized needs. On the other hand, after generating control commands based on voice data including information related to the user's expected feelings, the accuracy of the control commands is determined. If the commands are accurate, home appliances are controlled according to the commands. This significantly reduces the possibility of misoperation, ensures the alignment of executed control commands with the user's actual intentions, and further guarantees a superior user experience.

Embodiments of the present disclosure provide a method, apparatus and electronic device for training a recommendation model. The method can include, in response to receiving a data deletion request of a user, determining data requested to be deleted in original data. The method can also include obtaining remaining data for the user based on the data requested to be deleted. Furthermore, the method can further include training the recommendation model using the remaining data. Through the technical solutions of the embodiments of the present disclosure, the long model training time and large amount of computing overhead caused by retraining the recommendation model can be avoided, and the model updating time is shortened. In addition, since the training of the recommendation model forgets the data expected to be deleted by the user, the performance of the recommendation model is improved, the recommendation result is more likely to hit the user's preference, and thus the user experience is improved.

The invention relates to the field of terminals, in particular to a video generation method, electronic equipment and a computer readable storage medium. The method comprises the steps that when a user instruction is monitored, character information contained in the user instruction is recognized, and the character information comprises gender information; obtaining a target material from a preset image library according to the character information, wherein the character gender of a first character contained in the target material accords with gender information in the character information; wherein the figure gender of the first figure contained in the target material is determined according to the estimated gender of the first figure contained in each material in a material set of the first figure in the preset image library, and the estimated gender represents the gender obtained by performing gender detection on the figure in the material; and generating a video according to the target material. Through the method, the character information of the materials in the image library is more accurate, and the accuracy of obtaining the materials can be improved when the related materials are obtained according to the user instruction, so that the generated video is more in line with the user expectation.

The invention discloses a method and equipment for constructing a data warehouse model, a storage medium and a product, and relates to the technical field of artificial intelligence, automation of table-level design and field-level design is completed through a large model, a field design template is combined with the large model to ensure a structured design specification, field collaborative design is performed based on a table-level context, and the design efficiency is improved. And outputting the design core elements of the data warehouse model meeting user expectations and normative standards. By analyzing the demand document and automatically generating based on the core elements of the large model logarithmic bin model, the time consumption of manual lookup is remarkably reduced, by deeply embedding the structured design template into the generation process, it is ensured that an output result is forcibly aligned with an organization standard, and by combining table-level context with field-level design, the design efficiency is improved. Context logic fusion of field name design and table level design is ensured, logic consistency in the model is improved, and design efficiency and design quality of the number bin model are improved.

The invention provides a search method and system, a product application and a storage medium. The search method comprises the following steps: performing intention recognition on a user search dialogue to obtain a search intention; based on the search intention, determining a knowledge base associated with directional query of the search intention as a target knowledge base; and searching the search intention from the target knowledge base to obtain an intention result, and outputting the intention result. Therefore, the search intention is used, the associated knowledge base is directionally queried, learning is fast, the whole knowledge base does not need to be queried, query efficiency is high, and the effect is good. In addition, learning training or keyword configuration in advance is not needed, and the cost of repeated practice is reduced. Meanwhile, through the search intention, complex user search dialogues can be achieved, the query capacity is high, and the search result better meets the user expectation.

The application discloses a server performance analysis method and electronic equipment, and relates to the technical field of electric digital data processing, can perform performance analysis according to the running data of the server when the user has performance analysis demand, and in the case that the server is determined to be in a business state and the frequency of the central processing unit is less than a frequency threshold, the actual analysis intention in the user description information is used to perform a loss output ratio calculation, and then the performance bottleneck type of the server is determined, the analysis of the server performance is realized, the technical problem that it is difficult to determine the performance analysis direction from the user's subjective and fuzzy description sentence in the related art is solved, the performance analysis result does not match the target expected to be improved by the user, and the user experience is reduced, the analysis direction is consistent with the scene direction described by the user, and invalid analysis caused by disconnection of the intention is avoided.

The invention relates to an information feedback method and device, electronic equipment and a storage medium. The method comprises the following steps: in response to received input information of a user, obtaining a query result from a target database based on the input information; based on the query result, processing the input information into cue word data; based on the prompt word data, feedback information for the user is generated through a large model; wherein the target database comprises a database which is dynamically updated according to a new word determination result and / or a hot word determination result. According to the scheme provided by the invention, the target database can update the content in the target database in time by dynamically updating the database according to the new word determination result and / or the hot word determination result, so that the timeliness of the content in the target database is relatively high, the content lag in the target database is avoided, and the user experience is improved. Therefore, more accurate knowledge which is easier to meet user expectation is provided for the large model, accurate feedback information can be output to the user, and the fluency of communication with the user and the user experience are improved.

The invention relates to the technical field of data processing, and discloses an information processing method and device of a vehicle-mounted system, computer equipment and a program product, the method comprises the following steps: detecting an information request instruction of a target object, the information request instruction being an instruction sent by a user based on media content of the vehicle-mounted system; obtaining object information of the target object; and based on a pre-configured health database, filtering the object information, and generating a target material list of the target object. Therefore, the problem that information related to media content is inconvenient to operate, record or search when a user uses a vehicle-mounted system is effectively avoided, the information request instruction of the user is timely and accurately responded, the obtained object information is filtered according to the pre-configured health database, and the user experience is improved. The target material list of the target object conforming to user expectation is obtained, and the user experience is remarkably improved.

The application discloses a JPEG image compression method and system based on double Markov chain optimization, relates to the technical field of image compression, and comprises a receiving / acquiring unit, an activity level analysis unit, a state updating unit, a Markov prediction unit, a JPEG compression unit and an output processed image unit. The system can multi-dimensionally analyze user activity level, dynamically update a state transition matrix, adopt a multi-Markov chain model to predict a quality factor, so that the best balance between image quality and compression efficiency is found, resource waste or poor image quality caused by fixed parameter compression is avoided, appropriate compression decisions can be made by virtue of multi-source data, system state is dynamically updated, changes of system environment and user behavior are responded in time, and image quality meeting user expectation is provided according to user operation on images and activity level.

This application relates to a method and apparatus for controlling vehicle lighting, a vehicle, and a program product. The method includes: acquiring target lighting interaction habit data associated with the vehicle's current driving route from a pre-built memory module; acquiring initial lighting control parameters for the current driving route based on the target lighting interaction habit data; optimizing the initial lighting control parameters based on real-time driving environment information of the current driving route to obtain target lighting control parameters; and sending the target lighting control parameters to a lighting execution system; the lighting execution system is used to control the vehicle's lighting using the target lighting control parameters. This method enables more efficient lighting control that better meets user expectations for lighting performance, thereby improving the user's driving experience.

This disclosure relates generally to the field of user / device interactions. More particularly, it relates to techniques for detecting when a user's attention is directed at an electronic device, e.g., as determined based, at least in part, on analysis of images captured by one or more cameras integrated in the electronic device. Attention awareness can help to reduce the power and / or computing resources consumed by the electronic device, e.g., by only providing certain user experiences at the electronic device when they are actually likely to be desired by the user. In some embodiments, an attention awareness algorithm may be initiated by some triggering event or action. Once initiated, images captured by a camera of the electronic device may be fed to an attention detection algorithm to determine whether the user's head is in a pose where the algorithm believes that the user likely desires to interact with the device's user interface.

The application discloses a kind of federal feature forgetfulness method and system based on blockchain, it is related to computer technology field.The method comprises: when client receives the multi-feature forgetfulness request initiated by user to global model, client adds Gaussian noise to the feature to be forgotten in multi-feature forgetfulness request;Global model is trained based on the local dataset of all clients;Local dataset includes multiple features;The feature to be forgotten is the feature that user expects to remove from global model;Client calculates the sensitivity of the feature to be forgotten after adding Gaussian noise, and determines the mask weight of the feature to be forgotten based on sensitivity;Sensitivity is used to quantify the degree of influence of the feature to be forgotten on global model;Blockchain updates global model based on the mask weight of the feature to be forgotten, and obtains global model after feature forgetfulness.The method reduces the complexity of model calculation in the process of feature forgetfulness.

This specification discloses an adaptive operation control method and device for an air purifier. The method constructs multi-source time-series data by acquiring indoor environmental parameters, user operation behavior data of the target air purifier, and time context data. After preprocessing, a sequence of samples is constructed using a sliding time window, with the user's actual operations at future moments serving as a monitoring signal. The sequence of samples is input into a trained Long Short-Term Memory (LSTM) network model, which, through its gating mechanism and state calculation, outputs a predicted probability distribution of the user's expected operating mode. Based on the predicted mode and its confidence level, and combined with real-time indoor environmental parameter thresholds, a final control strategy is generated through a strategy decision engine to control the air purifier's operating mode and fan speed. User manual intervention is monitored; when the system's automatic control actions deviate from the user's expectations, the LSTM model parameters are updated based on this deviation.

The invention discloses a learning assistant cognitive load adaptive method based on few samples and related equipment, and relates to the field of artificial intelligence. The method comprises the following steps: providing a candidate explanation group consisting of a plurality of candidate explanations for a user, recording selection behaviors of the user in multiple rounds of graphic questions and answers, extracting cognitive preference characteristics of a plurality of style dimensions, and generating style preference vectors; selecting a basic multi-modal model with a freezing parameter and an adjustable LoRA parameter, mapping the style preference vector into a gating coefficient, only training the LoRA parameter and keeping the basic model parameter freezing, obtaining a low-rank increment and forming a user personalized parameter set; the method comprises the following steps: analyzing input characteristics before task input, generating a task complexity score, generating a temporary adjustment coefficient based on the difference between task complexity and user expected information density, correcting a gating coefficient, and calculating an effective low-rank increment in a reasoning process; and discarding the temporary correction parameters after reasoning is completed so as to maintain the stability of the model structure.

The system according to the embodiment aims to enable a user to find a product that meets the desired purchase conditions in a short time. [Solution] A system according to an embodiment includes a receiving unit, a recommendation unit, and a providing unit. The receiving unit receives desired purchase conditions from a user. The recommendation unit recommends products based on the information received by the receiving unit. The providing unit provides the user with the products recommended by the recommendation unit.

This application provides a graphical user interface (GUI) operation method, apparatus, device, and medium. The method includes: performing intent recognition and parsing on natural language commands to obtain a structured intent sequence; performing parsing and semantic mapping operations on the candidate GUI operation sequence to obtain a structured GUI operation sequence; constructing a matching score matrix between the structured GUI operation sequence and the structured intent sequence based on a local sequence alignment algorithm; determining the element with the largest value in the matching score matrix as the matching degree; and executing the candidate GUI operation sequence if the matching degree is greater than a threshold. This solves the semantic difference between the GUI operation sequence and the natural language commands, and robustly handles the mismatch in granularity and length between the two sequences, thereby improving the accuracy of the calculated matching degree and thus increasing the consistency between the GUI operation result and the user's expectations, enhancing the user experience.