47 results about "Preference learning" patented technology

Information retrieval systems face challenging problems with delivering highly relevant and highly inclusive search results in response to a user's query. Contextual personalized information retrieval uses a set of integrated methodologies that can combine automatic concept extraction / matching from text, a powerful fuzzy search engine, and a collaborative user preference learning engine to provide accurate and personalized search results. The system can include constructing a search query to execute a search of a database parsing an input query from a user conducting the search of the database into sub-strings, and matching the sub-strings to concepts in a semantic concept network of a knowledge base. The system can further map the matched concepts to criteria and criteria values that specify a set of constraints on and scoring parameters for the matched concepts. Furthermore, the system can learn user preferences to construct one or more profiles for producing personalized search results.

Information retrieval systems face challenging problems with delivering highly relevant and highly inclusive search results in response to a user's query. Contextual personalized information retrieval uses a set of integrated methodologies that can combine automatic concept extraction / matching from text, a powerful fuzzy search engine, and a collaborative user preference learning engine to provide accurate and personalized search results. The system can include constructing a search query to execute a search of a database. The system can parse an input query from a user conducting the search of the database into sub-strings, and can match the sub-strings to concepts in a semantic concept network of a knowledge base. The system can further map the matched concepts to criteria and criteria values that specify a set of constraints on and scoring parameters for the matched concepts.

Systems, methods, and apparatus are provided for equalization preference learning for digital audio modification. A method for listener calibration of an audio signal includes modifying a reference sound using at least one equalization curve; playing the modified reference sound for a listener; accepting listener feedback regarding the modified reference sound; and generating a weighting function based on listener feedback. A listener audio configuration system includes an output providing a sound for listener review; an interface accepting listener feedback regarding the sound; and a processor programming an audio device based on listener feedback.

The present invention relates to a self-service intelligent vertical search method and includes the following steps: cookies files of users, registered information, historical search information and ordered attention module can be utilized to study preferences of users, and the preferences of the users are set as statistical models of the users which are real-timely stored to a database of a search engine dynamically. A final key word / words collection file can be obtained by studying high speed Chinese word segmentation and search habits of users. The search engine can search all the information that is relevant to the inertial key sentence / word through a network database. Meanwhile, the search results are matched with the statistical models of the users and the search results which fit for the preferences of the users can be returned to the users. The present invention has the beneficial effects that the users can find out the needed information from the huge information collection and study using preferences and habits of the users actively, and then the search results can cater to ''tastes'' of the users more and the users make judgment to the values of the search results totally.

In embodiments of the present invention improved capabilities are described for a computer program product embodied in a computer readable medium that, when executing on one or more computers, helps target responses returned to a user through the use of a computer facility by performing the steps of (1) providing a user preference learning API to a third-party website to determine preferences of a user as related to a market of a third-party, wherein the preference learning API is executing as an extension of the computer facility; (2) receiving third-party information related to the market of the third-party; (3) collecting the preferences of the user and storing them as a user preference profile; (4) receiving a query from the user at the third-party website; and (5) using in the user preference learning API the third-party information related to the market of the third-party and the preferences of the user as stored in the user preference profile to provide a response back to the user that relates to the query from the user.

The invention provides an interest point recommendation method based on user positive and negative preference learning, and the method comprises the steps: learning deep features of interaction between a user and a scenic spot through employing a neural network, and training two neural network models: a positive preference neural model and a negative preference neural model; wherein the positive preference neural network model generates a list of scenic spots loved by the user, and then acquiring a final recommendation list through optimization of the negative preference neural network model,so that more accurate scenic spot recommendation is provided for the user. The problems that traditional interest point recommendation precision is not high, and the individuation degree of recommendation results is low are solved.

An algorithm for selecting audio content including an audio database, a user preference learning routine, and an audio selection routine. The user preference learning routine is capable of explicitly learning by a manual entry made by the user concerning his / her favorite audio selections. The user preference learning routine is capable of implicitly learning by recording the selections made by the user through a preference input. The preference input queries the user whether he / she liked or disliked a particular audio selection. Initially, random selections are provided to the user. If the user does not like the audio selection it is skipped. If the user liked the selection it is recorded. As the user listens or doesn't listen to audio selections, a profile is developed that identifies the user's preferences. The audio selection routine chooses the next audio selection to be played for the user. There are at least two methods of making the audio selection. In the first method, a user sends his / her listening history to a central location. In the second method, a table of at least four items is used to determine the user listening preferences.

An information processing apparatus learning a preference of a user for a content item includes acquiring means for acquiring an operation or expression of the user for a certain content item as feedback information; training data generating means for generating training data for the preference learning from the feedback information acquired by the acquiring means; and learning means for learning the preference of the user and how to attach a meaning to the feedback information in association with the training data by using multiple pieces of training data generated by the training data generating means.

The invention provides a personalized knowledge recommendation method and system for different scenes. The method comprises the following steps: obtaining user data information, and judging a usage scene where a user is located; if the scene is a short-term session type scene, establishing a session learning model, extracting user interaction items in the session as input, modeling each interaction behavior through an attention mechanism, and outputting preference of session content; establishing an edge information learning module, and predicting the current user demand by mining the edge information of the project by utilizing the graph structure information; if the scene is a long-term continuous scene, using a classification-layering attention mechanism for encoding, and carrying out preference learning on two levels of knowledge items and categories to predicate user requirements; and recommending related information according to predicted user requirements. The recommended information is more accurate.

The invention discloses a multi-behavior migration recommendation method based on deep learning, and the method comprises the steps: firstly obtaining and processing a plurality of implicit feedback data sets of a user; constructing a base network Gb and a plurality of behavior networks G (k), and learning low-dimensional embedded representations of users and article nodes in each network by usinga network representation learning method; then, based on different influence of multiple implicit behavior feedbacks of the user on user preference modeling, using an attention mechanism for automatically learning the weight of each behavior, and acquiring fused low-dimensional embedded representation of the user and the object finally, naturally splicing and sending low-dimensional embedding vectors of the user and the articles and to a full-connection embedding layer, adopting and feeding back a preference learning method based on a deep neural network to a feedforward neural network witha hidden layer, wherein the preference of the user for articles is learned on an output layer. The method can better capture the preference of the user and realize personalized recommendation, and has the advantages of high recommendation accuracy, strong generalization ability, easiness in realization and the like.

The invention provides a backward learning-based dynamic multi-attribute service selecting method. The method comprises the following steps: 1, user preferential learning based on backward learning: initializing a service set, a user set, a service evaluation level set and a user non-functional attribute evaluation link (UQEL) table; calling service several times and giving evaluation by a user, mapping the user evaluation on service to evaluation on corresponding non-function attribute, adding to the UQEL table of the user to acquire a user preferential table finally; and 2, weight-based dynamic multi-attribute service selection: generating a candidate service set according to user requirement to acquire a user preferential set, calculating the weight of each non-functional attribute to generate a dynamic decision matrix sequence, calculating weight included angle cosine of a user preferential vector and a candidate non-functional attribute vector, the weight of each observation time, and the weight cosine sum of each candidate service in all observation times, and recommending the service with the maximum weight sum to the user. The method is used for realizing adaptive service selection without more participation of users, is convenient to use, and has good service selection adaptability.

According to one embodiment, a broadcast receiver includes a tuner which receives and demodulates a broadcast signal, an extracting section which extracts electronic program information from the demodulated broadcast signal, a generating section which displays a setting screen and generates input information for the setting screen, a communication section which transmits the input information to an external server device and receives watching history data corresponding to the input information from the server device, and a preference learning section which creates a recommended program list based on the watching history data and the electronic program information and generates program recommendation screen information for displaying recommended programs according to the recommended program list.

In embodiments of the present invention improved capabilities are described for targeting web-based responses returned to a user through the use of a computer facility by utilizing a user preference learning facility to determine preferences of a user.

A personalized advertising system for sending personalized advertisements to a user comprises a positioning device, a track database, a customer database, a preference learning module and a personalized advertising module. The positioning device calculates a moving trajectory according to the wireless signal sent by a mobile device of the user and stores the moving trajectory in the track database. The customer database stores a shopping profile of the user. The preference learning module generates a predicted shopping preference information for the user according to the moving trajectory and the shopping profile. The personalized advertising module sends a personalized advertising message to the user according to the predicted shopping preference information. Compared with the prior art, the system of the present invention utilizes the location tracking technology to detect the behavior of the user in the store, and sends the appropriate advertisements to the user by automatic preference learning.

The invention discloses an education software data processing method based on big data. The method comprises the steps: carrying out the classification marking of all courses on education software, and screening each historical learning record in a preset time period according to the login account of a student user in the education software; analyzing the preference course category, the preference learning duration and the preference course teacher of the student user so as to screen the courses meeting the preference of the student user for recommendation, meanwhile, associating the learning and testing conditions of the student user together; and according to the test scores corresponding to the historical learning records in the screened preset time period, analyzing skilled course category and the weak course category of the student users; therefore,ethe learning data of the student user on the education software are analyzed and processed comprehensively, the defects that existing student network education software is single in function and low in intelligent level are overcome, the function of the network education software is perfected, and the intelligent use requirements of users are satisfied.

Systems, methods, and apparatus for equalization preference learning are provided. An example method includes receiving a first label for a first audio concept for a media object and applying active learning to select a first example not yet rated by a first current user. The example method includes collecting a first user rating, by the first current user, of the first example compared to the first audio concept and applying transfer learning to combine the first user rating with ratings from prior users of examples not yet rated by the first current user to build a model of the first audio concept. The example method includes creating a tool operable by the first user to generate examples close to and far from the first label to modify the media object.

The invention provides an intelligent photographing method based on user preference multi-dimensional driving, and the method comprises the steps: firstly obtaining a preview picture of a camera, andanalyzing a scene to which the preview picture belongs; if the scene exists in the scene database, continuing to inquire the professional portrait scenery picture data set under the scene category; and performing a matching operation according to the picture scene similarity to find a group of pictures with the highest background similarity; finally, calculating a picture most likely to be liked by the user through the user preference learning model, and taking the picture as a final shooting reference picture to guide the shot person to lay down the posture, thereby improving the shooting experience of ordinary people.

Systems, methods, and apparatus are provided for equalization preference learning for digital audio modification. A method for listener calibration of an audio signal includes modifying a reference sound using at least one equalization curve; playing the modified reference sound for a listener; accepting listener feedback regarding the modified reference sound; and generating a weighting function based on listener feedback. A listener audio configuration system includes an output providing a sound for listener review; an interface accepting listener feedback regarding the sound; and a processor programming an audio device based on listener feedback.

The invention discloses a user interface design determination method and device, and relates to the technical field of artificial intelligence and deep learning. According to the scheme, a screening model of a design scheme of a user interface module is obtained through L times of iterative training; randomly generating N design schemes of the user interface, inputting the N design schemes into ascreening model; obtaining M design schemes, generating a plurality of training data pairs according to the M design schemes, wherein each training data pair comprises two design schemes and a comparison result, a preference learning method of a Gaussian process is used for training the multiple training data pairs to obtain an evaluation model of the design schemes, the evaluation model is a Gaussian regression model, and the design schemes to be evaluated are input into the evaluation model to obtain scores of the design schemes to be evaluated. After the evaluation model is obtained throughoptimization, the evaluation model is used for automatically evaluating the design scheme, evaluation of the design scheme can be quickly completed, and the evaluation time is very short.

The invention discloses a personalized costume matching recommendation method and system using a time factor. The method comprises the steps of collecting a data set, constructing a heterogeneous graph, learning feature representations of clothes and a user and general rules of clothes matching, learning preference rules of the user for a suit, calculating personalized matching scores, training a heterogeneous graph network, and obtaining a scoring model integrating a clothes matching degree and user personal preferences, and finally, the user obtaining personalized costume matching recommendation according to the model. According to the method, the heterogeneous graph neural network is constructed, different preference degrees of different interactive behaviors of a user on clothes are considered, and the influence rule of time factors on user preferences is learned by integrating user interactive behavior time and clothes node time; and finally, the preference of the user to the suit is learned according to the historical record and the clothing information of the user and the preference of the user to the single item, the feature representation and the internal relation of the clothing and the user are modeled, and the personalized clothing matching recommendation method is realized.

The invention proposes a tourist preference learning system based on a sightseeing behavior. The system comprises an iBeacon module, an intelligent module, an application and a server; the iBeacon module is used for identifying position information of sightseeing spots in a scenic region or exhibits inside an exhibition hall; the application is used for receiving and parsing a broadcast data packet of the iBeacon module, calculating the distance between the intelligent terminal and iBeacon equipment, and identifying a current sightseeing spot of a tourist; the intelligent terminal is used foracquiring sightseeing behavior data of a certain sightseeing spot of the tourist and uploading the sightseeing behavior data to the server; and the server is used for processing and storing the sightseeing behavior data of the tourist, and acquiring the preference of the tourist to different sightseeing spots according to a preference learning model. According to the method of the invention, the iBeacon module and the intelligent terminal are combined to obtain the fine-grained tourist sightseeing behavior of the tourist, and the fine-grained sightseeing preference of the tourist for all sightseeing spots in a certain scenic region can be obtained.

The invention discloses a method for predicting the popularity of a shop based on singular value decomposition (SVD) and solves a technical problem of low efficiency of a conventional shop address selecting system and method. The method uses user preference as a medium, adds shop characteristic fusion and shop preference learning on the basis of SVD, and acquires effective information by using social media and location-based service, wherein the information includes the information of commercial areas around the shop, traffic information around the shop, and passenger flow information around the shop. The method for predicting the popularity by using the SVD not only embodies recessive characters but also embodies extracted dominant characters. In a matrix decomposition process, a left singular vector, a right singular vector, and a shop feature vector are iteratively computed. A neighbor shop of a newly-opened shop is computed by using the shop, the vector parameter value of the newly-opened shop is acquired by fitting. The method solves a problem of inaccurate newly-opened shop parameter due to matrix sparesity and is improved in efficiency.

The invention relates to the field of overlapping community discovery in complex networks, discloses a preference learning mechanism-based overlapping community discovery method and system, and aims at rapidly, efficiently, correctly and stably detecting overlapping communities in complex networks through considering preference relationships. The method comprises the following steps of: distributing a unique label for each node, setting a membership degree, to the label, of the node, and defining significance of the node; calculating preference degrees between the nodes; selecting a learning object to normalize the nodes according to the preference degrees, so as to obtain each node relational graph after iteration; and continuously carrying iteration on the basis of each node relational graph after the last iteration, doing the rest in the same manner until iteration is carried out for a set sequence to obtain each final node relational graph, dividing the nodes with same labels intoa same community, considering the nodes with at least two labels as overlapping nodes, and considering communities where the overlapping nodes are located as overlapping communities.

The disclosure provides 'preference learning for adaptive OTA notifications'. A vehicle includes a controller and a processor. The processor is programmed to prompt the user with a selection to install a software update to the controller responsive to a user preference value, computed from user selections whether or not to install software updates associated with contextual information that matches a software update to be installed, exceeding a threshold defined by a priority of the software update, and update the user preference value per the selection.

Systems, methods, and apparatus for equalization preference learning are provided. An example method includes receiving an audio input with respect to a target sound. The example method includes extracting one or more features from the audio input to provide one or more examples for rating based on the audio input. The example method includes generating a query based on the audio input and the one or more rated examples. The example method includes providing one or more synthesizer suggestion results identified in a search based on the query. The example method includes evaluating the one or more results with respect to the target sound. When one of the results matches the target sound, the example method includes outputting synthesizer parameters associated with the result. When none of the results matches the target sound, the example method includes refining the query for a second search based on feedback with respect to the one or more results.

The invention discloses a method for predicting color harmony degree according to user preference learning. The method comprises the steps: based on user comments of a large website data set, carryingout semantic classification on comments of users in the data set on color themes, and dividing the comments into multiple different harmony degree values according to specific keywords; extracting multiple color themes from the color themes, and performing training in a three-layer back propagation neural network; and generating an aesthetic ability value conforming to user aesthetics from a result output from the hidden layer through a probability density model based on Kernel distribution; and finally, acquiring a final color harmony degree value through linear calculation. According to theinvention, the harmony degree value estimated by the method has high accuracy due to consideration of color aesthetic appreciation of different users, and the method can be applied to practical application scenes such as color suggestions with color palettes, image re-coloring and color style conversion.

The invention discloses a power dispatching communication network node importance identification and evaluation method, which comprises the steps of S1, dividing a node importance identification criterion of a power dispatching communication network into three levels, namely a topology layer, a flow layer and a service layer; S2, respectively evaluating the importance of the power dispatching communication network nodes in the topology layer, the flow layer and the service layer; S3, performing normalization processing on importance indexes of the power dispatching communication network nodesin each layer; S4, obtaining importance evaluation of the nodes through a preference learning multi-attribute decision algorithm; or obtaining the importance measurement value of the communication network node through direct summation calculation. According to the evaluation method, importance evaluation of the nodes in the network topology level, the flow level and the service level is comprehensively considered, the importance degree of the nodes in the power dispatching communication network can be accurately and reasonably identified and evaluated, and the evaluation method has great significance in improving the network robustness, reducing the network failure risk and improving safe and reliable operation of a power grid.

The invention discloses an advertisement click rate estimation method based on user preferences. The method comprises the following steps: acquiring data to obtain an advertisement click sequence of each user; preprocessing the advertisement click sequence to obtain a sequence training data set; constructing a click rate prediction model based on a user preference network; carrying out iterative training on the click rate prediction model based on the user preference network by adopting the training data set obtained by preprocessing; and inputting a complete advertisement click sequence of the user, learning user preference characteristics by utilizing the click rate prediction model based on the user preference network obtained by training, and obtaining a click rate prediction result. According to the invention, a novel attention mechanism is introduced, implicit long-term and short-term preferences in a user sequence are considered at the same time, user drift preferences and fusion preferences are further learned, the defects of an existing method in user preference diversity multi-dimensional and drift modeling can be overcome, and the advertisement click rate estimation accuracy is effectively improved from the perspective of user preference learning.

According to one embodiment, a broadcast receiver includes a tuner which receives and demodulates a broadcast signal, an extracting section which extracts electronic program information from the demodulated broadcast signal, a generating section which displays a setting screen and generates input information for the setting screen, a communication section which transmits the input information to an external server device and receives watching history data corresponding to the input information from the server device, and a preference learning section which creates a recommended program list based on the watching history data and the electronic program information and generates program recommendation screen information for displaying recommended programs according to the recommended program list.