6results about How to "Meet real-time needs" patented technology

The present application provides an endoscope adjusting system and method, which comprises a button and a controller connected in communication; the button is arranged at a preset position of a control end of a surgical robot for an operator to input endoscope control instructions and send the endoscope control instructions input by the operator to the controller, the endoscope control instructions comprising endoscope movement instructions and endoscope parameter adjustment instructions; the controller is arranged in communication with the endoscope and controls the endoscope to move or adjust corresponding parameters according to the received endoscope control instructions. The present application enables the operator to control the movement of the endoscope and adjust the parameters of the endoscope directly through the button, so that the endoscope can be controlled according to the operator's will, the communication cost between doctors and nurses is reduced, and the adjustment of the endoscope can be realized without a large movement of the main control arm, ensuring the continuity of the operation.

According to the edge device retrieval enhancement generation system and method based on super-dimensional calculation, the medium, the program product and the terminal, the characteristics of sparse super-dimensional vector coding, lightweight similarity calculation and the like of a super-dimensional calculation technology are deeply fused with an RAG system, and an adaptive edge device RAG system architecture is constructed. Sparse super-dimensional vector coding of super-dimensional calculation is used for RAG knowledge base construction, storage overhead is reduced through binary bit compression storage, and edge device storage constraints are solved. The Hamming distance algorithm is adopted to replace traditional cosine similarity calculation, the calculation power requirement in the retrieval process is reduced, time delay is shortened, and the real-time requirement is met. And on the basis of clustering management of the HDC vector, new knowledge does not need to be subjected to total recalculation, and the updating efficiency is improved.

The invention provides a display device and a live program searching method. After a display device opens a live broadcast application, if a user wants to quickly search for an interested live broadcast program, a program search instruction can be input into the display device, and specified information needing to be matched with the program to be searched is indicated. And the display device responds to the program search instruction and can automatically call the intelligent agent, so that the target live channel conforming to the specified information can be screened out through the intelligent agent based on the channel information of the live channel provided by the live application and the program information of the live program corresponding to the live channel. Therefore, the display device can generate the corresponding target channel list based on the target live channels, and the user can search the interested live program in the target channel list, so that the search range of the user can be greatly reduced, the provided live program can meet the real-time requirement of the user, and the user experience is improved. The efficiency and the accuracy of searching the live program by the user can be effectively improved.

This application provides a dynamic charging method, apparatus, and storage medium for lithium batteries, relating to the field of battery management. It solves the technical problem in existing technologies where, when charging lithium batteries at battery swapping stations, it is difficult to dynamically select the charging mode of the lithium battery based on the expected demand of each swapping station and the state of the lithium battery, and to set adaptive charging methods. The method includes: acquiring target data for each lithium battery within the battery swapping station; dividing a day into several time periods, and extracting the reserved fully charged battery capacity for the current time period from historical data based on the time range of the current time period; selecting high-power charging batteries and dynamic power charging batteries based on the reserved fully charged battery capacity and target data; performing high-power charging on the high-power charging batteries; and adaptively charging the dynamic power charging batteries based on the grid load of the current time period. This application can reduce battery reactive power loss while improving the lifespan of lithium batteries and the safety of battery swapping stations.

This invention discloses a method and apparatus for dynamically allocating GPU computing power for real-time tasks, comprising: acquiring priority information, preset execution deadline information, and execution parameter information of each parallel task currently being executed by the GPU; determining the current remaining computing power value of the GPU; inputting the execution parameter information and the sub-computing power value into a time prediction model to obtain the remaining execution time required to complete each of the parallel tasks; determining a second computing power value required by the GPU to complete the parallel tasks; and dynamically allocating computing power values ​​to each of the parallel tasks based on a comparison between the second computing power value of each parallel task and the remaining computing power value, as well as the priority information. This method can dynamically adjust quotas for applications to meet real-time requirements and ensure concurrent execution of multiple tasks sharing the GPU.

The application relates to the technical field of personalized recommendation, in particular to a dynamic knowledge graph and large model personalized recommendation system, which comprises an intention representation module, a graph calling module, a graph screening module and a large model reasoning generation module. The intention representation module receives user historical behavior logs and real-time interaction signals, generates user current intention representation through semantic coding; the graph calling module calls a dynamic knowledge graph containing entity nodes, relationship edges and time limit weight matrices; the graph screening module performs graph traversal based on the current intention representation, screens out a candidate entity set with real-time correlation scores; and the large model reasoning generation module inputs the two into a large language model, combines entity causal relationships and time limit weights in the graph for reasoning, and generates a candidate recommendation list containing recommendation reasons. The system can improve recommendation accuracy and interpretability and adapt to real-time user needs.