11results about How to "Solve the cold start problem" patented technology

The invention discloses a multi-user sharing-oriented multimedia network video recommendation method, which comprises the following steps: firstly, constructing multi-user characteristics by utilizing collected program information in a multi-user sharing environment, and constructing a leading user label according to the similarity of the program characteristics and the continuity of user watching behaviors, so that separation of multi-user mixed logs is realized; performing periodic multi-user identification prediction of future sessions; secondly, building a user interest mining model based on a time-varying LinUCB algorithm to learn interest changes of a user for each program theme, and enhancing the personalized ability and efficiency of a recommendation system from three aspects of parallel calculation, adaptive control of an exploration coefficient and incremental updating based on LSTM; and finally, establishing an article quality model based on a non-time-varying LinUCB algorithm to further ensure the program quality, and integrating the two algorithms into a final recommendation system model by adopting a cross weighting strategy to form a final program recommendation list. The novelty and accuracy of the recommendation result are ensured.

The invention discloses a large-model-driven urban railway newly-opened station flow prediction method, and relates to the technical field of urban railway traffic operation management. The technical problems of low prediction precision, poor interpretability of a traditional model and insufficient reasoning flexibility in a complex urban scene caused by lack of historical passenger flow data of a newly opened station of an urban rail are solved. The method comprises the following steps: firstly, collecting pre-processed passenger flow related multi-dimensional data, and then constructing a multi-dimensional reference system to retrieve a target station similar reference or extract a macroscopic index reference; judging a scene category through a rule engine, calling a corresponding template, and filling information to generate a structured cue word; inputting the cue word into the large language model, and outputting a content prediction value and a result of a natural language reasoning chain according to the reasoning thinking chain template; and finally, checking the result, and completing optimization of the method based on feedback. The method has a good application prospect in the technical field of urban rail transit operation management.

The invention belongs to the technical field of multimedia content pushing, and discloses a multimedia content accurate prediction and pushing method based on a user portrait. Comprising the following steps: collecting data information of a user on a login content platform, and constructing a user portrait; obtaining content data of the multimedia content, and constructing a multimedia content portrait; matching the user portrait with the multimedia content portrait, and determining to-be-pushed content; based on feedback data analysis of the user, judging whether the pushed content to be pushed is accurate or not; and when the to-be-pushed content is inaccurate, correspondingly adjusting the to-be-pushed content. According to the method, the hidden core content preference characteristics of the user are mined according to the dominant core content preference characteristics of the user, the potential preference requirements of the user can be mined, the subsequently recommended multimedia content fits the real interest of the user and covers the preference of the user which is not directly exposed, recommendation can accurately reach the deep preference of the user, and the user experience is improved. Therefore, the user experience and the stickiness are improved.

The application discloses a plateau meteorological data downscaling method based on cross-terrain knowledge transfer, and belongs to the technical field of data-driven meteorological modeling and calculation. The method comprises the following steps: obtaining low-resolution meteorological data and corresponding high-resolution meteorological data of a source domain, low-resolution meteorological data and corresponding high-resolution meteorological data of a target domain, and terrain data; training a basic downscaling network model based on the low-resolution meteorological data and the corresponding high-resolution meteorological data of the source domain; performing adaptive training on the model based on the low-resolution meteorological data and the terrain data of the target domain and in combination with a composite target function containing a terrain difference regularization term; fine-tuning the model based on the meteorological data of the target domain to obtain a final downscaling model, and realizing downscaling of the low-resolution meteorological data by using the model. The application can improve the model performance, generalization ability and physical consistency.

The application provides a fluid electromagnetic heater of a methanol engine, comprising: a heater body arranged on an engine body and connected with an inlet of a fuel supply pipeline; an electromagnetic induction heating module arranged on an outer wall of the heater body and capable of converting electric energy into heat energy and transmitting the heat energy to fluid flowing through the heater body; an electric control unit electrically connected with a control unit of the engine and capable of controlling and adjusting a heating power of the electromagnetic induction heating module according to an engine working condition; wherein the electromagnetic induction heating module comprises a magnetic and heat conductive layer and an electromagnetic coil arranged around the magnetic and heat conductive layer. The application can realize uniform, on-demand and efficient heating operation through cooperation of the electric control unit, the electromagnetic induction heating module and the heater body, greatly improves the combustion efficiency of the methanol fuel, and solves the problems of cold start difficulty, insufficient combustion and high emission of the methanol engine.

The application discloses a user cold start recommendation system and method based on similarity matching, constructs a knowledge graph database according to users, products and the interaction relationship between the users and the products, and carries out standardization processing. The processed data is extracted to score information and mapped to a user like-dislike type space, so that the like-dislike product type of each user is determined, and a similarity matching algorithm is designed to find the old user with the highest similarity to the new user. The old user product type score vector and the new user product type vector score are processed by an association function and used as the recommendation basis of the system, the reasonable expansion of the new user data is completed, and the cold start problem of the recommendation system is effectively alleviated. The application solves the user cold start problem caused by the introduction of a new user into an industrial recommendation system. The application can be used for the scene of effectively recommending the product types that meet the preferences of new users in the industrial recommendation system.

The one or more embodiments of the specification provide a label data management method and device based on a multi-tenant environment, which comprises: a platform receiving an information query request sent by a first intelligent agent, analyzing user information of a target user carried therein, and determining an accessible label set thereof; then searching a label database according to the user information, obtaining label information corresponding to the accessible label of the user, the label information being generated based on historical dialogue data of a second intelligent agent and the target user; and finally returning the accessible label and the corresponding label information to the first intelligent agent to support the first intelligent agent to generate personalized dialogue content.

This invention belongs to the field of distributed artificial intelligence and decentralized network technology, and solves the problems of trust cold start, Sybil attacks, and cross-domain trust breakdown in A2A (Agent-to-Agent) economic networks. Based on the DHUC self-other-fault model, a trust weight formula is proposed: W_trust=ε×C_obs×10 3 ×(1+HCC_valid / HCC_total)×G_heritage achieves zero-collateral trust establishment; the uniqueness of the 428571 seven-dimensional genes is used to calculate the witch weight W_sybil=(C×S×G) / 10□, identifying and isolating cloning attacks; cross-chain trust lossless transfer is achieved through Trust Capsule and zero-knowledge proof; a triadic consensus mechanism is adopted to coordinate multi-agent cross-domain collaboration. This invention constructs a meta-layer trust infrastructure for A2A networks, applicable to Sigil Wen-style Automaton interaction networks, Justin Sun's Web 4.0 multi-chain AI infrastructure, and ERC-8004 standard agent payment networks, defending against dimensionality fall attacks and genetic contamination.

The application discloses a reentry vehicle reachable domain rapid prediction method based on a GA-TD3 hybrid algorithm. In view of the problems of time-consuming calculation of a traditional numerical iteration method and blindness of single deep reinforcement learning in exploration under a sparse reward environment, a GA and TD3 co-evolution strategy is adopted to design a trajectory planning system. In view of the multiple path constraint challenges such as heat flow, dynamic pressure and overload faced by a hypersonic aircraft, a composite reward shaping mechanism and a GA start-up preheating strategy are adopted to guide and correct the hypersonic aircraft, and finally, an evolution-gradient hybrid training framework is adopted to design a reachable domain prediction algorithm, so that the prediction accuracy of the reachable domain boundary of the hypersonic aircraft and the robustness of the algorithm convergence are successfully improved.

The invention discloses an intelligent detection method and system for electric power field safety operation. Historical image data of an electric power field safety operation scene are collected; image data is used to carry out fully supervised training on the YOLO model, and a fixed category target is identified; based on the image data, constructing a multi-modal data set based on an electric power security rule knowledge graph; performing fine tuning training on a YOLO-World model by using the preprocessed multi-modal data set, and identifying an open category target; acquiring real-time image data, preprocessing the real-time image data, and inputting the preprocessed real-time image data into the trained YOLO model and the trained YOLO-World model; removing overlapping frames from the recognition results of the two models and then outputting a final detection result; performing compliance comprehensive analysis on the final detection result; and according to an analysis result, judging whether to prompt or give an alarm. According to the method, a fixed category target can be accurately identified, a variable category or a new target can be flexibly handled, and fine tuning training of the model is carried out by constructing a data set driven by a knowledge graph based on an electric power safety rule, so that the adaptability and accuracy of the system are improved.