105 results about "Entropy model" patented technology

The invention provides a multi-temperature-salinity-depth sensor cooperative detection task planning method, and aims to improve the detection precision and resource utilization efficiency of a sensor network. The method comprises the following steps: firstly, acquiring initial detection parameters of a plurality of temperature-salinity-depth sensors, and calculating a detection capability matrix of each sensor based on the information; then, a target detection area is partitioned through a regional environment model, and a target information entropy model is constructed based on dynamic changes of environment parameters; and combining the target information entropy model and the detection capability matrix to construct a task allocation optimization model, and generating a preliminary sensor task allocation scheme. Furthermore, on the basis of considering environment change, detection data quality and sensor resource constraints, the preliminary task scheme is optimized, and finally a sensor collaborative detection task planning scheme is formed. Through the information entropy model and the optimization task allocation method, the detection efficiency and data quality of the sensor are improved, and the method can be widely applied to marine environment monitoring and underwater detection tasks.

The invention belongs to the field of image / video compression, and discloses an intelligent image compression encoder based on a conditional reversible neural network, which comprises an image enhancement module, a reversible multi-frequency fusion network RMFFN and an entropy model, and is characterized in that the image enhancement module comprises a multi-expansion channel refiner module MDCR and an expansion residual attention module RDAM; the image enhancement module optimizes input features and constructs a conditional reversible neural network by using a flow model, namely, multi-stage nonlinear mapping is realized by stacking four reversible multi-frequency fusion networks; the super-prior codec of the entropy model extracts hidden variable features by stacking multi-scale residual attention blocks, and introduces a discrete Gaussian mixture likelihood model in an entropy coding stage. According to the method, high-fidelity reconstruction is still kept at a low bit rate, and the technical problems of bit rate-distortion balance, high-frequency detail retention and calculation efficiency of a traditional method are solved.

The invention relates to the technical field of industrial Internet of Things data processing, in particular to a dual-channel anomaly detection method based on local entropy and isolated forest, and the method comprises the steps: collecting multi-dimensional time sequence data of equipment operation in real time; extracting local data of the multi-dimensional time series data in each sliding window by using a variable step size sliding window technology, calculating four statistical indexes according to the local data to construct a multi-dimensional entropy model, and obtaining a local entropy score according to the multi-dimensional entropy model; utilizing an isolated forest algorithm to calculate an isolated forest abnormal score of the multi-dimensional time series data; and constructing a joint scoring model, dynamically adjusting the weight of the local entropy score and the weight of the isolated forest abnormal score to obtain a joint score, and judging the abnormal level of the equipment according to the joint score and a preset abnormal threshold value. Therefore, the problems of insufficient real-time performance, insufficient feature utilization, low decision collaboration efficiency and the like in an industrial scene are solved.

Methods and systems implement a pleno-generation face video compression framework with bandwidth intelligence for generative models and compression. Heterogeneous-granularity facial description regularizes long-term dependencies between video frames and compensates for motion estimation errors caused by compact representations of motion information. A generative decoder reconstructs heterogeneous-granularity visual representations, providing auxiliary visual signals for attention-based recalibration of a GFVC-reconstructed face signal. A coarse-to-fine generation strategy avoids error accumulation. High efficiency for heterogeneous-granularity signal compression is achieved by two different entropy-based signal compression methods: heterogeneous-granularities feature representation from the key-reference frame as hyperpriors to optimize the entropy model for compressing heterogeneous-granularity feature from subsequent inter frames, and a feature difference operation for heterogeneous-granularities feature representation between key-reference and subsequent inter frames, such that the entropy model only compresses heterogeneous-granularities feature residual for redundancy reduction. Mixed-model dataset generation and training and model-specific dataset generation and training are also provided.

Systems, methods, and instrumentalities are disclosed herein for reducing the amortization gap in end-to-end image compression and / or video compression. In examples, a video decoder may obtain an entropy model indication in video data. Based on the entropy model indication, the decoder may determine an entropy model to use for decoding a current picture. The current picture may be decoded based on the determined entropy model. In examples, the entropy model indication may indicate whether to use an updated entropy model or a prior entropy model for decoding the current picture. In examples, the entropy model indication may indicate an updated entropy model or a learned entropy model to use for decoding the current picture.

The invention discloses an end-to-end video coding method based on a scale-spatial stream and hierarchical hyper-prior, and the method employs a hierarchical hyper-prior coding and decoding network to introduce hyper-prior information to assist the generation of entropy model parameters, thereby enabling an entropy model to obtain more accurate estimation probability distribution, and improving the compression efficiency. And meanwhile, the motion information of the video is better estimated in combination with a scale-spatial stream, and the video coding performance is improved by estimating a more accurate entropy model and motion information. Experimental results show that when SSIM and MS-SSIM are respectively adopted as evaluation standards, compared with an HEVC (High Efficiency Video Coding) standard, the coding method provided by the invention has the advantage that the average code rates are respectively saved by 16.40% and 15.63% under the same distortion condition for video sequences with different resolutions. Compared with a traditional video compression method, the quality of the compressed low-resolution video can be effectively improved, and a good visual effect can be obtained through the method provided by the invention.

The invention discloses a point cloud end-to-end compression method based on efficient sampling and an optimized entropy model, and the method comprises the steps: designing a neural map sampling module through employing the construction of a local map, the embedding of map features and the sampling based on attention, and enabling the neural map sampling module to be embedded into a feature extraction network in a coding process, potential key points of original point clouds are extracted in a multi-module stacking mode, and the sampling efficiency is effectively improved. A deformable dynamic kernel is used for expanding the space aggregation capacity through an asymmetric space channel entropy module, potential variables are grouped along the channel dimension, and the statistical redundancy is reduced while the coding efficiency is kept. Based on the addition of the neuro-map sampling module and the asymmetric space channel entropy module, the effective optimization of the acquisition of the most sampling value points and the distribution parameter estimation is realized, so that the end-to-end compression of the massive point cloud data is realized.

The invention provides a compressor intelligent control method and system based on dynamic pressure relief and a storage medium, and the method comprises the steps: firstly, capturing the multi-physical field coupling characteristics of a compressor in a shutdown state in real time by deploying a pressure, temperature, vibration and current four-dimensional sensing network; secondly, based on a preset pressure entropy model, fusing the pressure fluctuation intensity, the temperature anomaly index, the high-frequency vibration factor and the current attenuation coefficient to quantitatively evaluate a pressure entropy value, and triggering a three-stage pressure relief mode based on the pressure entropy value; then, based on the deviation between the cavity pressure and the target pressure, pressure relief parameters are dynamically adjusted, and stable pressure transition is achieved; and finally, collecting full-link data of pressure relief operation, and dynamically optimizing pressure entropy model parameters based on reinforcement learning, so that the system continuously adapts to equipment aging and working condition changes. Through dynamic closed-loop adjustment, pressure relief energy consumption waste and equipment loss are reduced, and the stability of the pressure relief process is improved.

The invention provides a channel context-based full-I-frame video compression method, which mainly comprises the following steps of: firstly, converting an input frame # imgabs0 # into a potential feature y through an encoder network, and introducing a hyper-prior model to capture hidden information of potential representation so as to assist in parameter generation of an entropy model; then, the information is converted and expressed as new intermediate information # imgabs1 #. Intermediate information # imgabS2 # is quantized and then enters an arithmetic encoder to obtain a bit stream, a super-prior decoding network reconstructs super-prior information to obtain entropy model estimation parameters, the data passes through a decoding network, the decoding network decodes by using the parameters to obtain reconstructed data, the reconstructed data passes through an image enhancement module, and image data reconstructed under deep learning is output. Experimental results show that the method can effectively improve the compression effect, and original information is reserved as far as possible after the image is compressed.

The invention belongs to the field of artificial intelligence and financial science and technology, and relates to a dynamic form generation method, which comprises the following steps: preprocessing unstructured text data requested by a client form to obtain preprocessed text data; inputting the preprocessed text data into an intention recognition model to obtain an intention category; form keywords are selected according to the weights, and the incidence relation of the form keywords is obtained; retrieving a candidate record set from historical query records according to the form keyword; calculating probability distribution of the candidate record set by adopting a maximum entropy model to obtain a prediction probability; according to the prediction probability, determining a candidate record most related to the client form request as a form generation reference knowledge base; and generating a personalized form based on the form generation reference knowledge base. The invention further provides a dynamic form generation device, computer equipment and a storage medium. In addition, the invention further relates to a block chain technology, and historical query records can be stored in a block chain. According to the invention, personalized services can be provided, the business process is simplified, and the error rate is reduced.

This invention discloses a method, device, and product for hyperspectral remote sensing image compression based on a hybrid hyper-prior. This method utilizes a CNN- and MLP-based backbone analysis network and synthesis network to transform the original image into a compact latent representation, obtaining superior latent representation features. Furthermore, a spatial-pass-through attention module is used to further enhance the representational capabilities of the CNN-MLP hybrid network module, enabling in-depth exploration of redundant information within hyperspectral images. When constructing the entropy model, a CNN-based hyper-prior is used to mine local redundant information, while a Transformer-based hyper-prior is used to extract non-local redundant information. The two are then combined to predict the parameters of the entropy model, improving the accuracy of the entropy model construction and, consequently, the final image compression performance.

The invention discloses a Mama-based entropy model and an image compression method, and mainly solves the problem of limited compression performance caused by poor entropy estimation precision in the prior art. The scheme comprises the following steps: 1) constructing a two-dimensional state space hyper-prior network by a hyper-prior encoder and a decoder to obtain global hyper-prior features; 2) adopting a double-branch gating architecture to obtain a mixed context feature containing local and global dependency information; 3) constructing an entropy parameter fusion and probability modeling network, carrying out channel splicing and feature fusion on the mixed context features and global super-prior features, and outputting a conditional Gaussian distribution parameter of each latent variable position; according to the method, the entropy model estimation precision can be remarkably improved, meanwhile, the compression rate distortion performance is improved, and the balance between the rate distortion performance and the calculation complexity is achieved.

The invention discloses a photovoltaic cleaning robot path planning method fusing string level power generation analysis, and belongs to the technical field of data processing, and the method specifically comprises the steps: dividing historical string power generation data according to the same interval to construct an entropy model, collecting the data in real time, calculating an interval entropy, comparing the interval entropy with the model, and recognizing a suspected low power generation interval; key cleaning group strings are screened in combination with spatial neighborhood information, and a posterior probability is calculated through a Bayesian probability model to determine a cleaning priority; power station layout constraints and robot motion characteristics are combined, and an optimal cleaning path is generated by adopting a path planning algorithm; through multi-dimensional data fusion and an intelligent algorithm, precise identification and path optimization of photovoltaic string cleaning requirements are realized, the cleaning efficiency is effectively improved, and the operation and maintenance cost is reduced.

The invention relates to a man-machine mixed vision-oriented semantic perception video compression method and system, and the method comprises the steps: extracting the dynamic semantics of a video sequence, and generating a region of interest; generating a focusing frame with consistent vision according to the input frame and the corresponding interested area mask; predicting feature probability distribution of the focusing frame through an entropy model, and compressing the feature probability distribution into a code stream; decoding the code stream through a conditional decoder to obtain a semantic compressed reconstructed video; performing feature alignment on the decoded frames in the decoded frame buffer areas of the basic branch and the auxiliary branch to generate prediction features; inputting the prediction frame of the prediction feature and the video sequence into an entropy model, and compressing the prediction frame and the video sequence into a code stream through entropy coding; decoding the code stream to obtain reconstruction features; and converting the reconstruction features into fine reconstruction features to obtain a final compressed and reconstructed video. Compared with the prior art, the method has the advantages that high machine vision task accuracy can still be maintained under the condition of low code rate, and higher rate accuracy performance is achieved in the machine vision task.

The invention provides a point cloud data processing method and device, electronic equipment and a medium. The problem that existing point cloud data is challenging in the aspect of attribute compression is effectively solved. The method comprises the following steps: performing multi-level division on point cloud data to obtain a first multi-resolution grid; quantizing the latent variable based on the geometric information of the latent variable, constructing different types of context information based on the quantized latent variable, and constructing an entropy model based on the different types of context information; processing the different types of context information based on the entropy model to obtain a prediction result, and coding the quantized latent variable based on the prediction result to obtain bit stream information; and decoding the bit stream information to obtain a second multi-resolution grid, obtaining corresponding point data based on a latent variable of the second multi-resolution grid, splicing point cloud feature information carried by the point data, and reconstructing point cloud data according to the spliced point cloud feature information.

The invention discloses a remote sensing image compression method and device based on frequency domain enhancement and adaptive optimization. The method comprises the following steps: compressing and decompressing a remote sensing image through a pre-training first model deployed at a compression end and a decompression end at the same time; the first model comprises an encoder, an entropy model and a decoder; the encoder and the decoder perform frequency domain analysis on input through a frequency domain enhanced selective state space module, and model a long-distance dependence and frequency domain structure based on a state space mechanism; training the first model according to a preset training set and the loss function to obtain a pre-trained first model; compression includes: extracting, by an encoder, a main feature of an input image; probability modeling is carried out on the distribution information of the main features through an entropy model, a bit stream is generated, and the bit stream is sent to a decompression end; the decompression comprises the following steps: a decompression end receives a bit stream; restoring the bit stream into a main feature through an entropy model; and inputting the main features into a decoder to obtain a remodeled image.

The invention discloses a video compression method based on deep learning, particularly relates to the field of video compression, is used for solving the problem that dynamic migration of residual distribution cannot be tracked in real time by entropy modeling in deep learning video compression, and accurately depicts complex distribution of residual in space and time dimensions through a multi-scale self-attention mechanism. The adaptive entropy model configuration on each piece of data is realized by using a content-based mixed density generator, a bidirectional evaluation quantity is sent to a pre-training attention model to dynamically generate a threshold value to drive a high-precision coding template to implement fine compression on a high-frequency region, and meanwhile, entropy parameters are locally contracted based on real-time bit feedback to balance output. And the consistency of a decoding end model is guaranteed through parameter freezing synchronous metadata, so that synchronous optimization of compression efficiency and reconstruction quality is realized, fluctuation is converged, artifact deviation is suppressed, and the distortion limit is approached.

The invention relates to a three-dimensional point cloud prediction geometric coding method based on deep learning, and the method comprises the steps: 1, forming a prediction tree through points collected by each laser transmitter, and converting an original laser radar point cloud LPC into a plurality of prediction trees for representation; 2, respectively designing different predictors and entropy encoders for each component of the coordinates of the points, and compressing each component of the coordinates of the points by applying different quantization step lengths; 3, selecting a quantization step size for quantifying each component by adopting a quantization step size selection strategy; 4, adopting an entropy model to model the probability distribution of the residual error of each component so as to encode the residual error entropy; and step 5, decoding is realized through a reverse process from the step 1 to the step 4. Compared with other methods, the method provided by the invention obtains the best rate distortion performance.

The present invention relates to a machine vision encoding method and system based on self-supervised learning. The method includes the following steps: randomly sampling image information into sub-blocks, inputting the sub-blocks into a backbone network head to extract and transform feature channels to obtain a first feature; transforming the first feature to obtain a feature in a low-dimensional space, adding uniform noise to the feature in the low-dimensional space through a quantizer to obtain a quantized feature, and reconstructing a compressed feature to obtain a second feature; transforming the second feature to a low-dimensional space, adding uniform noise to the feature in the low-dimensional space through a quantizer to reduce redundancy, extracting and encoding side information, decoding the side information, using a mixed Gaussian entropy model to predict the probability distribution parameters and bit rate of the second feature, and reconstructing the dimension of the encoded feature as a third feature; extracting and transforming the dimension of the third feature, extracting and weighting the convolution feature to form a heat map, obtaining valid positive samples through the heat map, and obtaining an encoding result. Compared with the existing technology, the present invention has the advantages of low encoding complexity and high semantic reliability.

The invention relates to the technical field of education artificial intelligence, and provides a self-adaptive learning path recommendation method for artificial intelligence general recognition education, which comprises the following steps of: 1, calling a large language model to analyze unstructured course resources of AI general recognition education, mining semantic association and implicit dependency of knowledge points, and constructing a learning path; generating a space-time knowledge graph containing confidence scores, automatically capturing the latest literature and tool update of the field every 72 hours, and iteratively optimizing the topological structure of the graph; 2, deploying data acquisition points to acquire multi-source learning data of a user in real time, and calculating a dynamic cognitive state vector with knowledge points as dimensions by mastering an entropy model; the knowledge graph is automatically iterated through LLM, and the problem of update lag is solved; precise personalized recommendation is realized according to multi-dimensional data and a dynamic algorithm, and cognitive differences are adapted; visually presenting decision logic, and cracking a decision black box; the method can be migrated to multiple fields, supports multi-terminal and offline learning, exceeds an expected adaptive scene, and improves the learning efficiency and credibility.

An entropy-based technique is used to select a large language model capable of generating fluent natural language text. An entropy model, trained on fluent natural language samples, is used to determine the entropy of a large language model based on an output text generated by the large language model. The entropy of a machine-generated natural language text is used to quantify the amount of information that the large language model holds with respect to the tokens and context of an input text segment. The entropy score of a model is then used to select a large language model capable of generating fluent text or to select the most fluent machine-generated output text produced by a set of large language models.

The invention discloses a full-stream entropy encoding and decoding accelerator, which comprises an encoder and a decoder, the encoder comprises an encoding prefetching module and an encoding pipeline calculation core, the decoder comprises a decoding prefetching module and a decoding pipeline calculation core, the encoding prefetching module is used for reading information of a plurality of sets of entropy models at a time, and the decoding pipeline calculation core is used for decoding the encoding pipeline calculation core. Determining that the coding process and the decoding process of the to-be-coded characteristic value are not skipped according to the confidence coefficient, splitting the multiple sets of entropy model information into single sets of entropy model information in a parallel-to-serial mode, and inputting the single sets of entropy model information into a coding pipeline calculation kernel in combination with the corresponding to-be-coded characteristic value for coding to obtain a code stream; the plurality of decoding prefetching modules are used for polling and inputting the multiple segments of code streams and the corresponding entropy model information into the decoding pipeline calculation core for decoding, the decoding pipeline calculation core is used for intercepting the code streams corresponding to the reserved digits, calculating the anti-cumulative probability distribution value of the code streams, carrying out quantization processing on the anti-cumulative probability distribution value to obtain decoded characteristic values, and sending the decoded characteristic values to the decoding prefetching modules; the problems of computing resource waste and low storage efficiency are solved.

The application provides a digital semantic perception OFDM subcarrier allocation method based on an entropy model, and relates to the field of wireless communication. The method comprises the following steps: a semantic encoder in a joint coding and modulation module is used to perform semantic coding on an original image to obtain semantic features; the joint coding and modulation module is used to modulate the semantic features to obtain a modulation symbol sequence; a semantic importance evaluation module is used to process the semantic features to obtain a semantic importance vector; an orthogonal frequency division multiplexing transmitter is used to process the modulation symbol sequence into M to-be-transmitted orthogonal frequency division multiplexing data packets; the joint coding and modulation module is used to allocate the M to-be-transmitted orthogonal frequency division multiplexing data packets to M subcarriers according to the semantic importance vector and feedback channel state information; and the orthogonal frequency division multiplexing transmitter is used to send the M to-be-transmitted orthogonal frequency division multiplexing data packets to an orthogonal frequency division multiplexing receiver in a receiving end, so as to realize efficient and reliable transmission of digital semantic information in a wireless channel.

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for compressing and decompressing data. In one aspect, a method comprises: processing data using an encoder neural network to generate a latent representation of the data; processing the latent representation of the data using a hyper-encoder neural network to generate a latent representation of an entropy model; generating an entropy encoded representation of the latent representation of the entropy model; generating an entropy encoded representation of the latent representation of the data using the latent representation of the entropy model; and determining a compressed representation of the data from the entropy encoded representations of: (i) the latent representation of the data and (ii) the latent representation of the entropy model used to entropy encode the latent representation of the data.

An entropy-based technique is used to select a large language model capable of generating fluent natural language text. An entropy model, trained on fluent natural language samples, is used to determine the entropy of a large language model based on an output text generated by the large language model. The entropy of a machine-generated natural language text is used to quantify the amount of information that the large language model holds with respect to the tokens and context of an input text segment. The entropy score of a model is then used to select a large language model capable of generating fluent text or to select the most fluent machine-generated output text produced by a set of large language models.

An apparatus includes first and second circuitry, such as a first memory and a second memory. The first circuitry is configured to store compressed information representing a set of distribution functions associated with a set of entropy models. The second circuitry is configured to store mappings of requests for subsets of the set of distribution functions to one or more offsets in the first circuitry corresponding to the compressed information representing the subsets. The apparatus also includes circuitry that implements the set of entropy models. This circuitry is configured to perform entropy coding based on the subsets and a first entropy model selected from the set of entropy models. This circuitry can include one or more entropy engines that have a set of ports for encoding or decoding one or more streams of bits or symbols. Multiple streams can therefore be encoded or decoded concurrently or in parallel.

The invention discloses a long tribulus terrestris risk prediction method based on a multi-source data fusion maximum entropy model, and belongs to the technical field of long tribulus terrestris risk prediction. The objective of the invention is to solve the problem of accurate risk prediction of tribulus terrestris. The method comprises the following steps: collecting species distribution data and environment variable data of tribulus terrestris, and preprocessing; constructing a maximum entropy model; a three-level screening strategy is adopted, key environment variable screening is carried out on the preprocessed environment variable data in combination with the constructed maximum entropy model, and key link variables after screening are obtained; training a maximum entropy model by using the preprocessed species distribution data and the screened key link variables to obtain a trained maximum entropy model, and outputting a adaptability index of a grid unit; and carrying out potential suitable growth area division and early warning area delimitation on the tribulus terrestris. According to the invention, the space-time accuracy and prevention and control guidance value of the early warning result are greatly improved.

The invention discloses a multi-source entropy model image compression method based on a non-uniform channel. The image compression method comprises the following steps: obtaining potential feature representation information y after image information x passes through an encoder; the potential feature representation information y is divided into five non-uniform feature segments yi, the five non-uniform feature segments yi and prior information HyperParams provided by a hyper-prior network are combined to be fed into a multi-source entropy model MSEM (Multisource Enhanced Model) together to obtain generated feature segment information # imgabs0 # and delta i, and the delta i is the mean value mu and the variance sigma of the feature segments yi; and according to the mean value mu and the variance sigma of the feature segment yi, obtaining feature segment information conditional probability distribution # imgabs1 #, after the ith feature segment is processed, feeding generated feature segment information # imgabs2 # HyperParams and yi + 1 into an MSEMi + 1 module together, and performing circular processing. The number of channels fed to subsequent modules is reduced, so that the calculation amount of the model is reduced, and the efficiency of the model is improved.

The invention relates to a machine vision coding method and system based on self-supervised learning, and the method comprises the following steps: randomly sampling image information into sub-blocks, inputting the sub-blocks into a backbone network head part, extracting and transforming a feature channel, and obtaining a first feature; transforming the first feature to obtain a feature of a low-dimensional space, adding uniform noise to the feature of the low-dimensional space through a quantizer to obtain a quantized feature, and reconstructing the compressed feature to obtain a second feature; converting the second feature into a low-dimensional space, increasing uniform noise through a quantizer, reducing redundancy, extracting and coding to obtain side information, decoding the side information, predicting a probability distribution parameter and a code rate of the second feature by adopting a Gaussian mixture entropy model, and reconstructing a dimension of a coding feature as a third feature; and extracting and transforming the dimension of the third feature, extracting convolution features, weighting the convolution features into a thermodynamic diagram, obtaining an effective positive sample through the thermodynamic diagram, and obtaining a coding result. Compared with the prior art, the method has the advantages of low coding complexity, high semantic reliability and the like.

The invention discloses a three-dimensional point cloud geometric compression method and device based on a diffusion model and a medium, and the method comprises the steps: enabling an input point cloud to pass through a dual-domain encoder, obtaining low-frequency features and high-frequency features, and extracting skeleton points for space guidance; probability distribution of the low-frequency features is modeled through a full decomposition entropy model, probability distribution of the high-frequency features is modeled under low-frequency guidance, arithmetic coding is carried out, and skeleton points are subjected to lossless compression through G-PCC to jointly form a final compressed code stream; and carrying out multi-round iterative denoising reconstruction through a diffusion model, taking low-frequency features, high-frequency features and skeleton points as conditional information, inputting the conditional information into a decoding process in a mode of affine transformation injection and attention mechanism fusion, and guiding gradual recovery of a point cloud structure and details. According to the method, more robust point cloud geometric feature extraction is realized through a high and low frequency feature double-path fusion mode, a diffusion reconstruction process based on affine transformation and an attention mechanism is further guided, and the compression efficiency and the reconstruction quality are effectively improved.