9results about How to "Improve execution" patented technology

The application discloses a farmland planning analysis method and system based on satellite data, relates to the technical field of farmland planning analysis, and comprises the following steps: obtaining multi-satellite image data of a target farmland area, pre-processing and performing semantic segmentation, constructing a binary image, fitting line segment endpoints, forming a farmland closed polygon, obtaining agricultural machinery parameters, forming constraint conditions, performing regional decomposition on the farmland closed polygon based on the constraint conditions, forming a plurality of sub-regions with continuous operation attributes, constructing a bidirectional optional coverage path and a state set, judging the direction consistency of the path connection relationship between the sub-regions according to the coverage path, generating a complete path for division and evaluation based on the judgment result, and performing local path reconstruction according to the evaluation result to obtain a complete continuous path as the farmland planning analysis result. The application significantly improves the intelligent level of farmland planning analysis.

The application discloses an immersive neural rehabilitation training method based on virtual reality, and particularly relates to the technical field of neural rehabilitation training, and comprises the following steps: acquiring action data, tactile data, proprioceptive data and gaze data when a trainee performs a target action, performing uniform time sequence alignment on the data in a virtual reality environment, reorganizing according to a starting section, a pushing section and a reaching section of the target action, and outputting a target action response sequence; performing segmented comparison on the differences of multi-source responses of the target action under the conditions of visual weakening, tactile weakening and proprioceptive weakening to determine a perception dependency relationship, rearranging the prompt items and feedback items in the virtual reality scene according to the perception dependency relationship, and then performing recovery checking on the rearranged results.

The invention discloses an optimal path planning method of a pulse coupling neural network based on double constraints. The method comprises the following steps: mapping a path planning environment to a DC-PCNN network; a DC-PCNN neural network model is constructed, and all neurons are initialized; activating a target neuron, and recording the current neuron as a father node; calculating an exponential decay function of the torque deviation, and multiplying the calculated value as a penalty factor by an update item of the internal activity item; calculating a gravitational function value of the flow field constraint, and using the calculated value to update a current neuron dynamic threshold value; comparing the internal activity item with a dynamic threshold; when gt; if yes, activating the neuron, and recording the current neuron as a father node; repeating the steps S4-S6 until the initial neuron is activated; and backtracking all activated nodes, and planning an optimal path. According to the method, the search efficiency is remarkably improved while the path optimality is ensured.

The application relates to a compiled graph fusion optimization method, device, equipment, storage medium and computer program product. The method comprises the following steps: acquiring breakpoint information corresponding to a plurality of operation logic segments generated in a front-end tracking process of a computation graph compilation, the breakpoint information comprising operation information triggering a breakpoint, context tensor attributes and a dependency relationship, the dependency relationship being an association relationship between the operation triggering the breakpoint and the context tensor; based on the operation information triggering the breakpoint, the context tensor attributes and the dependency relationship, performing fusion reconstruction on the plurality of operation logic segments to obtain at least one fused logic segment; and performing back-end compilation based on the at least one fused logic segment to obtain compilation information of the computation graph. The method can reduce graph breakpoints.

The application discloses a supply chain link risk analysis method based on a block chain and big data, and comprises the following steps: S1, collecting multi-source heterogeneous data in a physical supply chain network, constructing a distributed ledger and a unified knowledge structure, and chaining and solidifying the cryptographic hash digest of core data, the application has the beneficial effects that: by collecting multi-source heterogeneous data in the physical supply chain and constructing a graph structure knowledge representation, the temperature, humidity, vibration, position and other physical states are mapped to the environment on the alliance chain in real time; the system can not only perceive the risk in advance, but also automatically execute multi-objective optimization and parallel simulation verification through the smart contract, so that the emergency scheme with better comprehensive performance considering the cost, time efficiency and reputation is screened out before the physical loss occurs; the mechanism helps to alleviate the problems of perception lag, decision dependence on artificial and slow response in the traditional supply chain, realizes the full-process automation closed loop from risk early warning to scheme execution, and is favorable for reducing the actual loss caused by supply interruption or quality accidents.

Embodiments of the present specification provide a method and device for training a multi-task model, wherein the multi-task model comprises a backbone network for determining a user representation, and k head networks for performing k user prediction tasks based on the user representation. The method comprises: determining, based on m user samples, k sets of original gradient vectors of the k user prediction tasks with respect to the backbone network, wherein each user sample comprises a user feature and k user labels; mapping the k sets of original gradient vectors to a subspace of an original space in which the k sets of original gradient vectors are located, to obtain k sets of mapped gradient vectors; determining r weights corresponding to the r spatial dimensions of the subspace based on the component distribution of the k sets of mapped gradient vectors on the r spatial dimensions of the subspace, and performing weighted processing on the r dimensional components of each mapped gradient vector respectively by using the r weights to obtain k sets of weighted gradient vectors; and mapping the k sets of weighted gradient vectors back to the original space to obtain k sets of processed gradient vectors, which are used to update the network parameters of the backbone network.

This application discloses a reminder method, device, electronic device, and computer-readable storage medium, relating to the field of health management technology. The method includes: determining the proximity of a user's medication anchor point based on the location of the target medication and the user's location; predicting the risk of medication unavailability based on the target medication's dosing plan, the user's historical outing patterns, calendar busyness, historical medication carrying success rate, and proximity to the medication anchor point; predicting a target medication preparation time window based on the distance between the user and the medication anchor point, the user's busyness level, and current activity status; continuously monitoring the real-time distance between the user and the medication anchor point, and the user's real-time busyness within the target medication preparation time window; and pushing a medication preparation reminder to the user when the real-time distance is less than a distance threshold and the real-time busyness is less than a busyness threshold for a specified duration. Thus, this solution ensures that the medication is readily available to the user when the scheduled dosing time arrives.

The present disclosure provides a blockchain-based Web3.0 system registration method and a spatial information network system, and is applied to a spatial information network system. When a first consensus ground node receives a user registration request, the first consensus ground node generates a blockchain identity and determines assistance nodes, each of which generates a key fragment. The client and the first consensus ground node generate a user public-private key pair based on system parameters and each key fragment. The first consensus ground node generates user storage information of a mobile user based on the public key in the user public-private key pair and the blockchain identity, and stores the user storage information in the blockchain. Through performance adaptation, scientific screening of assistance nodes is achieved, which not only avoids the efficiency bottleneck of a single node bearing too much load, but also replaces the traditional centralized mechanism with a distributed key fragment generation mode, greatly reducing the risk of single-point failure and the probability of key attack as a whole, while dispersing the computing pressure and improving the efficiency and security of key generation.

The application provides an industry knowledge graph-based large model enhancement training method, which comprises the following steps: constructing an industry task structure graph, structurally encoding task nodes to obtain a task structure vector set and a task structure hop number matrix; dividing a training task into stages based on the graph to form a stage task set and a corresponding stage sample set; training a large model stage by stage, and introducing a structural consistency constraint in each stage to align the model internal task representation with the task structure vector and make the relationship between tasks comply with the hop number matrix; and finally, fusing the model parameters and task intermediate representations of each stage, optimizing in a unified structure space, making the task representation of the final model close to the original graph structure, retaining the stage training memory and maintaining the hop number distance relationship between tasks. The method effectively improves the structural understanding and generalization ability of the large model in the industry task.