486results about How to "Lower deployment costs" patented technology

The invention discloses a classroom teaching quality auxiliary evaluation method, device and system based on cloud edge collaboration, and the method comprises the steps: obtaining video image data, collected by a camera, of students in a classroom in real time, and transmitting the video image data to an edge device; and in the edge device, performing student classroom behavior identification and tracking on the received video image data by using a pre-trained student classroom behavior identification model, and fusing the identification results of adjacent multiple frames with the identification result of the current frame in the tracking process to obtain the final identification result of the current frame. Evaluating the classroom teaching quality according to the student classroom behavior recognition result, wherein the student classroom behavior recognition model is obtained by training based on a multi-label target detection algorithm and by adopting a knowledge distillation structure in advance; and sending a student classroom behavior identification result and an evaluation result to a cloud video server for storage. The method has the advantages of being simple in implementation method, low in cost, high in real-time performance and precision, high in robustness and the like.

The embodiment of the invention relates to the field of large model facilities, and discloses a Serverless large model reasoning service system, method and equipment and a medium. Receiving an inference service creation request sent by a user, and forwarding the inference service creation request to a corresponding cluster; processing the north-south traffic of the cluster; converting the reasoning service creation request into an object, and configuring a running environment; according to the object, a Pod is created, the Pod is used for controlling a model service workload, the Pod comprises a user container, and the user container is constructed based on a basic mirror image comprising an operating system, a Supervisor and a Docker and is used for running a large model service; and the queue agent is responsible for processing routing, load balancing and request queuing of an incoming request based on a lightweight agent Sidecar. The technical problem that large model reasoning service deployment is difficult can be solved at least.

The invention discloses a brain-like edge calculation-driven multi-mode intelligent security method and system, and aims to improve the security capability in a complex dynamic scene. According to the method, multi-modal data of videos, audios, sensors and the like are collected, preprocessing is carried out by utilizing a brain-like edge computing gateway, dynamic behavior characteristics are extracted, secondary confirmation is carried out through a lightweight visual large model, and then formal early warning is generated. And the system executes intelligent linkage operation according to the early warning information risk level, and records false alarm information for model optimization. According to the invention, brain-like calculation and multi-modal fusion are adopted to improve the recognition precision, edge calculation reduces delay, an intelligent closed loop realizes differentiated treatment, a low-power design and a modular architecture reduce the cost, and the system has a self-learning capability. The defects of an existing security and protection system in the aspects of dynamic behavior recognition, response speed, disposal intelligence, cost control and the like are effectively overcome. The system is suitable for scenes such as a campus and a community, and provides an accurate, efficient and intelligent security solution.

The invention belongs to the technical field of artificial intelligence, and relates to an Android mobile phone intelligent control method and system based on a large model agent, and the method comprises the steps: 1) obtaining a task inputted by a user; 2) judging whether the task input by the user is within the processing range of the Agent model, if so, entering step 3), otherwise, ending the execution of the task input by the user; 3) obtaining page information of the Android mobile phone; (4) judging and processing the non-standard state of the page; 5) carrying out compression simplification and page completion processing on the obtained page XML; 6) inputting the processed page XML and the processed page screenshot into an Agent model, and outputting the type and the position of the operation executed by the Android mobile phone in the next step by the Agent model; and (7) according to the output of the Agent model, operation execution is carried out on the Android mobile phone on the basis of the accessible service (AccessibilityService). The mobile phone page state can be recognized and judged, and operation is planned and executed in combination with the intention of a user.

The invention provides a two-stage training method for a lightweight Chinese-English bilingual large language model. The method comprises the following steps: step 1, collecting and preprocessing Chinese-English training data; the preprocessing process comprises special mark clearing, sensitive data clearing, data duplicate removal, advertisement removal and formatting processing; 2, performing first-stage high-resource language training, namely performing model training by using English; 3, second-stage bilingual mixed training is carried out, and the Chinese data and the English data are mixed for second-stage training by adopting language migration training; wherein the step 3 comprises the following steps: step 31, performing word list expansion; and step 32, language migration training. The bilingual processing capability can be improved.

The invention discloses a coal face working state detection method based on YOLOv8. The method comprises the following steps of 1, obtaining a data set of all working states of a coal face; step 2, a YOLOv8-EST network is constructed; 3, inputting the working state data set of the coal face into the improved YOLOv8 model for training to obtain a working state detection model of the coal face; and 4, inputting a to-be-identified coal face state picture into the coal face working state detection model to obtain a detection result of the working state of the to-be-identified coal face. According to the invention, the improved YOLOv8 target detection algorithm is constructed by adopting the YOLOv8 target detection algorithm, fusing the improved Swin Transform module and fusing the EMA attention mechanism, so that the real-time, efficient and accurate detection of the working state of the coal face is realized, and the production safety and efficiency are improved.

The invention relates to the technical field of text abstract generation, in particular to a dialogue abstract generation method and device based on a large language model, equipment and a storage medium. The method comprises the steps of obtaining a first dialogue data set; cleaning the first dialogue data set to remove redundant dialogues in the first dialogue data set to obtain a second dialogue data set; inputting the second dialogue data set into a first preset model, and extracting through the first preset model to obtain a plurality of initial triples; and inputting all the initial triads into a second preset model, and combining all the initial triads through the second preset model to obtain a first dialogue abstract corresponding to the first dialogue data set. According to the method, the model deployment cost required for generating the text abstract can be reduced, and the beneficial effects of balancing the accuracy, practicability and use cost of generating the text abstract are achieved.

The invention discloses a training method of an image generation model, and an image processing method and device. Relates to the field of artificial intelligence, and the method comprises the steps: obtaining a training sample image set which comprises a plurality of spliced feature maps; the plurality of spliced feature maps are input into an initial student model to obtain a first prediction result output by the initial student model, the plurality of spliced feature maps are input into a teacher model to obtain a second prediction result output by the teacher model, and the initial student model is obtained by initializing parameters of the teacher model; the teacher model is a pre-trained text graph generative model; and constructing a target loss function based on the first prediction result and the second prediction result, and training parameters of the initial student model based on the target loss function to obtain a target image generation model. The technical problem of low image generation efficiency caused by high reasoning delay due to more reasoning step number and parameter quantity of an image generation model in the prior art is solved.

The invention discloses a portable mobile terminal collaborative relative positioning method and system, and the method comprises the steps: obtaining multi-source data, and carrying out the preprocessing of the multi-source data; wherein the multi-source data comprises an RTK differential value, a GNSS observed quantity and a Bluetooth observed quantity; gross error detection is carried out on the preprocessed multi-source data; constructing a fusion observation equation according to the RTK differential value after gross error detection, the GNSS observed quantity and the Bluetooth observed quantity, and obtaining an estimated value of a positioning parameter through a Kalman filtering algorithm according to the fusion observation equation; and carrying out relative position calculation according to the estimated value of the positioning parameter to obtain a relative positioning result of the portable mobile terminal. By adopting the technical scheme of the invention, the relative positioning with higher precision and moderate cost is realized on the portable mobile terminal so as to meet the high-precision relative positioning requirement under diversified application scenes.

The invention provides a charging pile software updating method and device, computer equipment and a storage medium, local software version information of each module in a charging pile and cloud software version information of each module in a file server are obtained, and the software version information comprises updating file information and updating file state; performing version self-checking on local software version information of each module based on a hash table in which latest software version information of each module is stored, and updating the hash table based on a version self-checking result; performing version check on the cloud software version information of each module based on the hash table, and updating the hash table based on a version check result; and in response to a software updating signal, performing software updating on each module in the charging pile based on the hash table. By adopting the method, the deployment, operation and maintenance cost of the system during software updating is reduced, and meanwhile, the universality and success rate of software updating are improved.

The invention discloses a large-scale vector approximate nearest neighbor search heterogeneous system based on a GPU-NVMe framework, which is characterized in that the system is based on a direct GPU-NVMe framework, and a hybrid index combining a clustering algorithm and a graph algorithm is designed. According to the hybrid index, vectors with good clustering are efficiently processed by utilizing the parallel processing capacity of the GPU, and meanwhile, the search accuracy of noise vectors is improved by taking a graph index as a compensation mechanism. According to the method, collaborative optimization of hardware and software is fully considered, vector index data deployment is implemented at an NVMe driving level, meanwhile, CPU and GPU hardware are cooperatively utilized to carry out a gigabit-level vector retrieval calculation process, and an extensible large-scale retrieval system on a direct GPU-NVMe architecture is constructed. Compared with the prior art, the method has the advantages that the performance bottleneck in the gigabit-level vector data search is effectively solved, the search throughput and accuracy are remarkably improved, and the search delay is reduced.

The invention provides a method and a system for realizing virtual simulation of the internet of things on a Web page. The method comprises the following steps of: resource deployment: deploying a front-end display layer, a rear-end logic layer and a data storage layer on a server; simulation configuration: loading the content of the front-end display layer to a Web page of the user terminal, and setting the attribute of each simulation model in the Internet of Things system according to the actual demand; starting simulation: loading corresponding simulation logic and resources according to initialization information set by a user; initializing all equipment and environment parameters, and entering a simulation state; model resolving: simulating network operation by using a discrete time driven framework to ensure that all network events occur at a specific time; and carrying out picture presentation and dynamic editing. According to the method for realizing virtual simulation of the Internet of Things on the Web page, through the design of a multi-layer architecture, front-end display, rear-end logic processing and data storage are combined, efficient and flexible simulation of the Internet of Things system is realized, and the development and test cost and time of the Internet of Things system are reduced.

The invention provides a strip-shaped long-distance unmanned aerial vehicle sensing countering system and method along a railway, and relates to the technical field of rail traffic safety. The system comprises an unmanned aerial vehicle detection device arranged at a key node along a railway and used for detecting an unmanned aerial vehicle; the unmanned aerial vehicle countering device transmits a controllable interference signal to interfere with the unmanned aerial vehicle; the signal receiving and directional transmitting equipment is used for receiving, amplifying and redirectionally transmitting the interference signal or the radio signal transmitted by the unmanned aerial vehicle; the rear-end intelligent management and control platform is used for receiving the signal information, analyzing the threat level in real time, selecting a countering strategy and implementing directional interference through equipment; the method comprises the steps of equipment deployment, detection and early warning, data analysis, instruction issuing, signal emission and interference, and response and dynamic adjustment. According to the invention, the problems of insufficient coverage, poor identification precision and single countering means in long-distance and strip-shaped scenes along the railway in the prior art can be solved, and real-time sensing and dynamic countering of the unmanned aerial vehicle are realized.

The embodiment of the invention provides a vehicle abnormity processing system, which is applied to a new energy vehicle and comprises a plurality of sampling devices, a first transmission pipeline and a specified processing unit, each sampling device is arranged on a sampling point position in a vehicle body of the new energy vehicle, and the plurality of sampling point positions comprise sampling point positions corresponding to a battery pack of the new energy vehicle; the first transmission pipeline is used for connecting the plurality of sampling devices and the specified processing unit; the specified processing unit is used for acquiring the to-be-analyzed gas transmitted by the target sampling device; acquiring a gas analysis result obtained by analyzing the to-be-analyzed gas; according to a preset corresponding relation, executing a first processing operation corresponding to the gas type of the to-be-analyzed gas; the target sampling device corresponds to the target sampling point position, and the to-be-analyzed gas is the gas transmitted to the specified processing unit through the first transmission pipeline under the condition that the target sampling device judges that the target sampling point position is abnormal.

The invention discloses a control method of a multi-cabinet liquid cooling system, electronic equipment and a computer readable medium, all liquid cooling cabinets are communicated with communicating pipe sections through communicating vessels, the lengths of liquid inlet pipe sections are matched, and the lengths of liquid outlet pipe sections are matched; periodically detecting the environment temperature, the liquid level of each liquid cooling cabinet and the liquid pool temperature; determining each liquid-cooled cabinet as one of the four cabinets according to the liquid level and the liquid pool temperature of each liquid-cooled cabinet; controlling the opening of the communicating vessel of the liquid cooling cabinet with normal load and normal liquid level; for the liquid cooling cabinets with normal loads and abnormal liquid levels, liquid inlet valves of the liquid cooling cabinets are controlled according to the liquid levels of the liquid cooling cabinets; controlling a communicating vessel of the liquid cooling cabinet with abnormal load and normal liquid level to open, and controlling a liquid pump of the liquid cooling cabinet according to the temperature of a liquid pool and the environment temperature of the liquid cooling cabinet; the liquid pumps of the liquid cooling cabinets with abnormal loads and abnormal liquid levels are controlled according to the liquid pool temperature and the environment temperature, and the liquid inlet valves of the liquid cooling cabinets are controlled according to the liquid levels of the liquid cooling cabinets. And the control and cooling efficiency is improved.

The invention discloses a digital role audio and video generation method and a digital role live broadcast interaction method, and the generation method comprises the steps: generating explanation data of a to-be-explained object according to a digital role material generation task of a requester; performing data conversion processing on the explanation data to obtain audio data corresponding to the explanation data; performing role image digital processing on the obtained role image reference data to obtain a digital role image video; carrying out fusion processing on the audio data and the digital role image video to obtain a digital role driving video for explaining the object to be explained by using a digital role; returning the digital role driving video to the requester, so that the requester processes the digital role driving video before distribution to generate a digital role audio / video; therefore, computing resources are saved, the cost is reduced, and the coverage rate of digital role audio and video playing is improved.

The invention belongs to the technical field of intelligent mine communication positioning, and discloses a mining communication positioning system and method based on RIS, and the system comprises a mining communication base station which is used for transmitting and receiving a communication signal to user equipment, and / or is used for achieving a positioning function; the mining communication positioning terminal equipment is used for a user to access a network service and comprises the steps of generating, sending and receiving a communication signal; the mining RIS equipment is used for changing communication signal transmission characteristics and / or realizing a positioning function; and the mining communication transmission network is used for transmitting communication signals among the mining communication base station, the mining RIS equipment and the mining communication positioning terminal equipment. The mining communication positioning terminal can be positioned under a coal mine, the cost of network equipment can be reduced, network interference can be reduced, and the communication quality can be improved.

The invention discloses a multi-task electromagnetic model based on a hybrid expert network, and belongs to a wireless communication technology. The model comprises a preprocessing module, a feature extraction module, a task output module and a pre-training-fine tuning learning strategy. The preprocessing module carries out standardization processing on the multi-source electromagnetic signals; the feature extraction module is based on a Transform structure, introduces a hybrid expert network to replace part of a traditional feedforward neural network, and dynamically selects an expert sub-network through a task specific routing mechanism; the task output module configures a special structure according to different task targets; in the pre-training stage, a mask auto-encoder is used for pre-training large-scale label-free data, and a downstream task is subjected to full-amount fine adjustment through small-scale label data. According to the method, multi-task collaborative learning and differential expression are realized, and the recognition performance, robustness and processing efficiency of the model in a complex electromagnetic environment are improved.

The embodiment of the invention discloses a wave period identification method and device based on deep learning, and the method comprises the steps: collecting wave data in a preset time length at preset time length intervals, carrying out the image processing of the wave data, obtaining a target sea wave image, and obtaining the white-head wave value data; and performing calculation based on the target sea wave picture and the white head wave value data to obtain a wave period estimation value, and constructing a wave period identification model of the wave truth value and the wave period estimation value. A target sea wave picture is automatically extracted in a deep learning mode, and errors and cost of manual intervention are reduced; based on calculation of a target sea wave picture and white head wave value data, an accurate prediction basis is provided for wave period estimation, the accuracy and stability of the wave period estimation are further improved in combination with application of a deep learning algorithm, and a wave period recognition model constructed by the embodiment can automatically learn and optimize the relationship between sea wave features and wave periods. Through continuous learning and iteration, the model can adaptively adjust parameters, and the prediction precision is improved.

The invention discloses a navigation type temperature and humidity laser radar dynamic observation system and method, and the system achieves the large-range and high-precision dynamic observation through the deep integration of a temperature and humidity laser radar, a radar fixing and damping structure, a vehicle-mounted temperature control system, an optical skylight, a Beidou positioning system, a terminal display system and a communication system. The radar fixing and damping structure is combined with a vehicle-mounted temperature control system, so that the influence of road surface vibration and environment change on the monitoring precision is effectively eliminated, and stable operation of equipment under complex road conditions and extreme weather conditions is ensured. The optical skylight design not only ensures efficient transmission of laser signals, but also protects external pollution, and further improves the observation reliability. Through combination of a Beidou positioning system and a GIS map, a dynamic distribution diagram of temperature, humidity and aerosol data is generated in real time, and the multi-dimensional real-time monitoring requirement is met. According to the method, the observation flexibility and applicability are remarkably improved.

The invention discloses an optical communication system and an optical communication method. An optical communication system includes a first device, an optical splitter, and a plurality of second devices. The first device is used for sending a first optical signal, the first optical signal is respectively transmitted to the plurality of second devices after being shunted by the optical splitter, and the first optical signal carries an analog radio frequency signal. The plurality of second devices are used for sending second optical signals respectively, the plurality of second optical signals are transmitted to the first device after passing through the optical splitter, and each second optical signal carries an analog radio frequency signal. The optical signal for communication between the first device and the second device bears the analog radio frequency signal, and the analog radio frequency signal does not need to be converted into a digital signal, packaged into an optical protocol frame format and then modulated into the optical signal. In this way, communication between the first device and the second device does not involve the optical protocol conversion process of optical protocol frame format encapsulation and de-encapsulation, the service delay of the system is reduced, the system structure is simpler, and power consumption and cost are reduced.

A method, system, device and medium for detecting an arbitrary modal saliency target relate to the field of image processing, and the method comprises the following steps: collecting uncertain modal saliency target detection image data; constructing an arbitrary modal feature extractor, extracting multi-scale features from the collected image data, predicting the features belonging to modal categories, and calculating modal category loss; constructing a feature fusion model of any mode, and fusing the features of any input mode to obtain fused features; constructing a saliency target detection decoder, obtaining a saliency target detection result by using the multi-modal fusion features, calculating the cross entropy loss of the saliency target detection result and a real saliency target detection result, adding the cross entropy loss with the loss of a corresponding modal category, then calculating the gradient of the total loss, carrying out back propagation, and updating model parameters; the system, the equipment and the medium are used for implementing the method. The method reduces the training and deployment cost of various models, improves the calculation efficiency, and can adapt to various scenes.

The invention relates to the technical field of wireless network positioning, in particular to an adaptive positioning method based on CSI, electronic equipment and a storage medium. According to the method, high-discrimination fusion fingerprint data is generated through a fusion representation model according to amplitude information data and phase information data, and mapping between the high-discrimination fusion fingerprint data and corresponding physical positions in the environment is completed through a positioning model; the positioning model comprises an adaptive convolutional neural network composed of a source network and a target network, the source network processes data from a source domain, and the target network processes data from a target domain. The positioning model uses a self-adaptive convolutional neural network composed of a source network and a target network, so that the adaptability to the environment is improved.

The invention relates to the technical field of parking lot monitoring, and discloses a parking lot real-time monitoring system and method based on a sensor, and the system comprises a sensor array, a sensor module of an image information collection unit, and a weather compensation module which is used for compensating the influence of weather conditions on the detection precision of the sensor. Comprising an image priority unit, an image processing unit and a parking space state unit, the scheduling module comprises a parking space sorting unit, a parking space scheduling unit and a parking space scheduling feedback unit, and the optimization module comprises a parking space change monitoring model and a parking space prediction unit. And the accuracy of the identification result is ensured.

The invention discloses a device and a method for adaptively adapting a cloud platform to a domestic AI accelerator card, belongs to the technical field of cloud computing and artificial intelligence, and aims to solve the technical problems of how to realize rapid integration and deployment of the domestic AI accelerator card by the cloud platform, improve the flexibility and expandability of the cloud platform and further better meet diversified AI computing requirements. According to the technical scheme, the device comprises a script tool set, a resource management module, an adaptation task management module, a data storage module and an adaptation analysis module; wherein the script tool set integrates a CPU and a script tool for identifying, sensing, driving and monitoring physical machine accessories of a domestic AI acceleration card, and is used for automatically realizing hardware identification, driving installation and hardware state monitoring functions according to the characteristics of hardware; and the resource management module is used for realizing platform access of the AI server and management of the domestic AI acceleration card, and monitoring states and loads of the nodes and the AI acceleration card in real time.

The invention discloses a chirp interval modulation backscatter communication method and a chirp interval modulation backscatter communication system based on variable-length scrambling codes, which mainly solve the problems of high power consumption of backscatter labels and high system deployment cost in the prior art. According to the scheme, a sending end sends an environment LoRa signal; the backscattering label modulates information to be sent to an environment LoRa signal in a time interval modulation mode, and modulates a variable-length scrambling code to an anchor symbol; a receiving end performs frequency shift compensation on the backscattering signal by adopting a frequency shift compensation method, demodulates the backscattering signal by combining a sliding window with a variable-length scrambling code, intercepts a backscattering signal sample through the sliding window, and performs de-spreading and fast Fourier transform (FFT) on the signal in the window; the position of an anchor symbol is determined by detecting the maximum value of an FFT peak value, information sent by the backscattering label is recovered from the position of the anchor symbol, and data demodulation is completed. On the premise of keeping demodulation reliability, the power consumption of the tag is effectively reduced, the deployment cost of a backscatter communication system is reduced, and the method can be used for ultra-low power consumption Internet of Things communication.

The invention belongs to the technical field of databases, and particularly discloses a storage method and device based on cold and hot data separation and a multi-mode database engine, and the method comprises the steps: monitoring a row storage table and a column storage table; under the condition that the cold data exists in the row storage table, transferring and storing the cold data on the row storage table into the column storage table; under the condition that the hot data exists in the column storage table, transferring and storing the hot data on the column storage table into the row storage table; the row storage table is used for storing hot data, the column storage table is used for storing cold data, one piece of storage data is stored in the row storage table or the column storage table in a single copy mode, the access frequency of the hot data is higher than that of the cold data, the row storage table is borne by a row storage engine, and the column storage table is borne by a column storage engine. According to the method and the device, the row storage table for storing the hot data is borne by the row storage engine, and the column storage table for storing the cold data is borne by the column storage engine, so that the storage system can support mass data query and analysis and also can efficiently perform DML operation.

The invention relates to a quantum key management method and system for satellite internet biological characteristics, and the method comprises the following steps: carrying out the registration of biological characteristics and the initialization of communication protocol parameters for a user terminal which is accessed for the first time; for the registered user terminal, before initiating communication each time, binding the real-time biological characteristics with the quantum key, and carrying out encrypted transmission of the session key by using the bound quantum key, so as to carry out encrypted transmission of satellite internet transmission data by using the session key; and in the encryption transmission process of the satellite internet transmission data, according to a preset period, performing satellite link security level adjustment and quantum key dynamic updating. According to the invention, identity authentication and terminal authentication in a high-delay and narrow-bandwidth application scene can be realized, the security problems of identity counterfeiting, key leakage and the like are solved, and the method can be applied to satellite internet scenes such as emergency communication, field operation and the like, and has the advantages of strong compatibility, low deployment cost and the like.

The invention discloses a method for assembling a cross-platform full-stack application and a medium, and belongs to the technical field of software. The method comprises the steps that a cross-platform UI frame is selected, and the UI frame is configured; defining a WebApi interface for processing data interaction; writing a Shell instruction adaptation program for starting a back-end service program; configuring an interface route in a Web back-end service, and deploying an SPA front-end file in a static file directory; developing an SPA front end, and outputting the static file to a back-end static file directory; a cache management strategy is started in a front-end code, and browser cache control is forced; and embedding the back-end service and the SPA front-end file into the cross-platform application during packaging. According to the method, the cross-platform UI, the back-end service and the SPA front end are integrated and seamlessly become a full-stack application which simultaneously comprises the cross-platform UI, the back-end service and the SPA front end, so that complex business requirements are met, the deployment cost is reduced, a condition for being published as an application is achieved, and finally user experience, program simplicity and stability are improved.

The invention provides a training method of a lightweight Chinese-English bilingual large language model. The method comprises the following steps: step 1, collecting and preprocessing Chinese-English training data; step 2, carrying out pre-training; step 21, performing text word segmentation; step 22, text coding; step 23, rotation position coding: rotation position coding in an attention layer is adopted in an input layer; step 24, setting an attention mechanism, a residual network, a standardization layer, a full-connection network output layer and an autoregressive training mode; step 25, selecting pre-training data; step 26, carrying out pre-training; each batch is input in a Chinese and English data crossing mode, that is, one piece of data is sampled from each of English data and Chinese data in a crossing mode until the size of the training batch is reached, and then the data is input into the model for training. And step 3, instruction fine tuning training. The method is extremely high in flexibility and adaptability.