14 results about "Tree code" patented technology

The invention discloses a sewage high-risk pollutant screening and identification method based on fragmentation tree pre-training, and the method comprises the steps: constructing a fragmentation tree pre-training data set according to the second-level mass spectrum data of known high-risk pollutants, so as to carry out the self-supervision pre-training of a graph neural network encoder, and obtaining a fragmentation tree editor; the method comprises the following steps: acquiring secondary mass spectrum data of suspected high-risk pollutant related compounds in a to-be-detected sewage sample, constructing a fragmentation tree set, constructing a high-risk pollutant screening model based on a fragmentation tree encoder, performing high-risk pollutant signal screening on the fragmentation tree set to obtain a high-risk candidate fragmentation tree set, and performing high-risk pollutant signal screening on the high-risk candidate fragmentation tree set. And generating a candidate molecule set of each high-risk candidate fragmentation tree, and identifying a target molecular structure and a matching score of the high-risk candidate fragmentation tree from the candidate molecule set. According to the invention, rapid screening and priority ranking of high-risk pollutants in sewage are realized, and candidate output of structure identification is further provided on the basis of screening.

The application discloses a sewage high-risk pollutant screening and identification method based on fragmentation tree pre-training, comprising the following steps: constructing a fragmentation tree pre-training data set according to secondary mass spectrum data of known high-risk pollutants, pre-training a graph neural network encoder in a self-supervised manner to obtain a fragmentation tree editor, obtaining secondary mass spectrum data of suspected high-risk pollutant-related compounds in a sewage sample to be tested, constructing a fragmentation tree set, constructing a high-risk pollutant screening model based on the fragmentation tree editor, screening high-risk pollutant signals from the fragmentation tree set to obtain a high-risk candidate fragmentation tree set, generating a candidate molecule set of each high-risk candidate fragmentation tree, and identifying a target molecular structure and a matching score of the high-risk candidate fragmentation tree from the candidate molecule set. The application realizes rapid screening and priority sorting of high-risk pollutants in sewage, and further provides a candidate output of structure identification on the basis of screening.

The embodiment of the application provides a kind of based on syntax semantic fusion sentiment tuple extraction method and system, it is related to natural language processing technical field.The method can be after obtaining input text, using structure perception syntax encoder calculates the encoding embedding data of input text, again through bidirectional cross attention fusion, obtains the fusion representation of syntax representation and semantic representation depth interaction.Based on double similarity comparison learning framework, the unified representation of fusion representation is calculated, and candidate example is retrieved according to the unified representation.The method can be based on the structure perception graph converter of dependency tree coding to generate encoding embedding data, and through the syntax and semantic fusion realized by bidirectional cross attention mechanism, and carry out double similarity comparison learning, for the structured task of large language model Provide explicit syntax guidance, can simultaneously integrate syntax and semantic information, correctly identify complete sentiment tuple, improve the output precision of sentiment tuple extraction result.

Post training compression of 3DGS data is agnostic to training in a traditional signal compression perspective. Gaussian parameters are treated as signals. Pre-processing and transform coding techniques are used to compress the signals effectively. Firstly, lossless / lossy compression is performed on 3DGS geometry (positions, scales, rotations) using a point cloud coding-based (e.g., G-PCC, GeS) framework. Positions are compressed using occupancy tree coding. Scales and rotations are encoded as attributes using transform coding. The widely used block-based graph Fourier transform (GFT) is used to compress the attributes (base colors, spherical harmonic coefficients and opacities). In addition, a graph construction strategy is used for 3DGS data that computes the edge weights based on similarity (or dissimilarity) between the 3D Gaussian distributions using KL-divergence. Alternatively, positions can be encoded using occupancy tree (e.g., G-PCC, GeS) or AI-based PCC methods, and any subset of Gaussian parameters or the transformed coefficients of Gaussian parameters can be mapped into 2D frames and encoded by video coders.

The application provides a data transmission method, device, user equipment, base station and system. The method is applied to user equipment in the coverage of a base station. There are multiple groups of user equipment in the coverage of the base station. The method comprises obtaining uplink resource information corresponding to a group to which the user equipment belongs. The uplink resource information comprises the frequency of L subcarriers used for uplink communication and the position information of n OFDM symbols corresponding to the group on a single subcarrier. The tree code encoder is used to encode the to-be-sent data and the check data, so as to obtain L groups of encoded data. Each group of encoded data in the L groups of encoded data is compressed into a code word with a length of n by using a compression matrix. The L groups of code words are subjected to OFDM modulation. After modulation, each group of code words with the length of n occupies one subcarrier in the L subcarriers, and each code word corresponds to one of the n OFDM symbols on the occupied subcarrier. The modulated data is sent. The uplink communication without random access can be realized.

The invention discloses a zero-sample Chinese character recognition method based on radical structure tree graph level representation, and belongs to the technical field of Chinese character recognition. The method comprises the following steps: establishing a Chinese character database and constructing a directed radical structure tree of each Chinese character from top to bottom; adding direction position codes to nodes in the tree to distinguish the spatial relationship of brother nodes; inputting the radical structure tree with the direction and position codes into a GraphSAGE-based tree encoder to obtain normalized vector representation of the Chinese characters; extracting vector representation of the sample image by using ResNet; aligning the Chinese character vector and the image vector through comparative learning; and finally, zero sample recognition is realized by calculating the similarity between the to-be-recognized image and all candidate Chinese character vectors, and the highest similarity is selected as a recognition result. According to the method, the structure perception capability is improved through the directed tree, fine-grained spatial differences are perceived through directional coding, and the recognition capability of the model on unseen Chinese characters is effectively improved.

The application discloses a point cloud encoding method, a point cloud decoding method and a terminal, and belongs to the technical field of point cloud processing. The point cloud encoding method comprises the following steps: obtaining a first identification parameter of a first target point cloud to be encoded; and performing an encoding operation on the first target point cloud based on the first identification parameter. The encoding operation comprises at least one of the following: in the case that the first identification parameter is used to represent parallel encoding, performing geometric encoding and attribute prediction encoding on the first target point cloud in parallel to obtain an encoding result of the first target point cloud; and performing geometric prediction encoding on at least part of the to-be-encoded points of the first target point cloud. In the embodiment of the application, the geometric encoding and the attribute prediction encoding are performed on the first target point cloud in parallel based on the first identification parameter of the first target point cloud. By performing the geometric prediction encoding on at least part of the to-be-encoded points of the first target point cloud, the multi-tree encoding on all the to-be-encoded points of the first target point cloud is avoided.

The invention provides a distributed drainage basin model multi-flux parallel computing method. The method comprises the following steps: S1, performing multi-way tree coding on mutually independent computing units according to parallel features of a distributed drainage basin model; s2, creating a parallel computing system according to the total process number, wherein the parallel computing system comprises a master control process which is globally and uniquely responsible for task allocation regulation and control, a plurality of computing processes which are responsible for physical processes and a plurality of transfer processes which are responsible for data transmission among computing units; s3, the main control process sequentially distributes the calculation tasks of all the units to the calculation process according to the multi-way tree codes; s4, the calculation process receives the task and receives upstream data from the transfer process, and calculation is started after unpacking; s5, after the calculation process task is completed, a result is packaged and transmitted to the transfer process; and S6, repeating the steps S3-S5 until all the calculation units are calculated. According to the method, the multi-flux parallel computing method for any distributed drainage basin model discrete in a tree structure is realized.

The invention discloses a platform data trusted sharing method based on a block chain and zero knowledge proof, and relates to the technical field of data security sharing, the method comprises the following steps: when data enters a sharing platform, executing structure analysis processing on data records to form a field set, extracting field structure parameters, and constructing a structure tree node sequence; calculating a field structure difference quantity Sdif, a structure hierarchical expansion quantity Sh and a structure connection stability quantity Sc, generating a structure consistency index Cstr, constructing a structure tree coding sequence, generating a structure commitment value through Hash operation, and writing the structure commitment value into the block chain for data registration; when a data sharing request occurs, structure analysis processing is executed again on to-be-shared data, a structure consistency index and a structure commitment value are calculated again, consistency judgment is carried out through a structure consistency difference quantity and a structure code difference quantity, comprehensive structure consistency judgment is executed according to a judgment result, and data sharing is allowed when structures are consistent. The sharing request is terminated upon a change in structure.

The application discloses a point cloud encoding method, a decoding method, a point cloud encoding device and a decoding device. The point cloud encoding method comprises the following steps: calculating the distance from a first vertex to the parent node of the first vertex as a first distance; determining the point closest to the first vertex among the uncoded vertices as a second vertex, and calculating the residual value of the second vertex on N coordinate components of the first vertex; encoding the absolute value of the residual value of the second vertex on N coordinate components of the first vertex; and encoding the sign of the residual value of the second vertex according to the residual value of the second vertex and the first distance. The application encodes the feasible sign bit of the absolute value of the residual value and the residual value by using the geometric position relationship between the vertices in the point cloud, improves the encoding efficiency of the residual value, and improves the performance of the point cloud prediction tree encoding by optimizing the efficiency of the residual value encoding.

The application belongs to the technical field of image processing, and proposes a brain tumor image segmentation method based on a lightweight multi-modal, the main scheme is: acquiring a multi-modal brain magnetic resonance image dataset, and constructing a multi-modal data fusion feature extraction network based on a binary tree structure, the feature extraction network includes a tree coding layer consistent with the number of modes; constructing a local feature extractor for each mode of brain magnetic resonance image, and performing feature extraction; constructing a multi-modal fusion unit for the feature extraction results of two adjacent modes, and performing inter-modal feature fusion; generating a skip connection parameter for each tree coding layer, and performing up-sampling on the output of all tree coding layers before utilizing the skip connection parameter for splicing and fusion operation; constructing a brain tumor image segmentation network, and obtaining the segmentation result of the brain tumor image; training, verifying and testing the two networks by using the multi-modal brain magnetic resonance image dataset.

This invention discloses an automatic annotation generation method based on multimodal code representation. It employs multiple encoders to learn the source code representation, including a graph encoder to holistically represent the API context graph, a tree encoder to represent the abstract syntax tree, and a field encoder to represent the token sequence, thereby learning a holistic and comprehensive code representation. Unlike existing annotation generation techniques based on encoders and decoders, this invention models the code from a holistic perspective, focusing on API usage and control flow information. Furthermore, it designs a feature fusion method to combine multiple intermediate encoding results, which are then fed into a joint decoder to generate annotations word by word. This invention can effectively generate high-quality annotations for specified code snippets, thereby promoting software maintenance and significantly improving software productivity.

Post training compression of 3DGS data is agnostic to training in a traditional signal compression perspective. Gaussian parameters are treated as signals. Pre-processing and transform coding techniques are used to compress the signals effectively. Firstly, lossless / lossy compression is performed on 3DGS geometry (positions, scales, rotations) using a point cloud coding-based (e.g., G-PCC, GeS) framework. Positions are compressed using occupancy tree coding. Scales and rotations are encoded as attributes using transform coding. The widely used block-based graph Fourier transform (GFT) is used to compress the attributes (base colors, spherical harmonic coefficients and opacities). In addition, a graph construction strategy is used for 3DGS data that computes the edge weights based on similarity (or dissimilarity) between the 3D Gaussian distributions using KL-divergence. Alternatively, positions can be encoded using occupancy tree (e.g., G-PCC, GeS) or AI-based PCC methods, and any subset of Gaussian parameters or the transformed coefficients of Gaussian parameters can be mapped into 2D frames and encoded by video coders.

The application belongs to the technical field of video coding, and particularly relates to a sample adaptive offset mode decision method and device, a computer readable storage medium and a terminal device. In the method, an optimal non-parametric fusion offset mode can be selected from various non-parametric fusion offset modes according to the division depth of a tree coding unit, and then an optimal offset mode can be selected from the optimal non-parametric fusion offset mode and a parametric fusion offset mode. Through the method, the optimal non-parametric fusion offset mode can be selected according to the division depth of the tree coding unit, the decision time in the non-parametric fusion offset mode is saved, the efficiency of the sample adaptive offset mode decision method is improved, and the method has strong usability and practicality.