94 results about "Rope" patented technology

This specification discloses systems and methods for enhancing the efficiency of transformer models during inference and training by precomputing and storing in memory a significant portion of operations in the first transformer layer. The stored precomputed outputs are retrieved from memory during runtime, reducing computational complexity and memory bandwidth requirements. This approach results in decreased latency, increased throughput, and lower cost-per-token. The disclosed techniques are particularly advantageous for transformer models that incorporate positional encodings within the attention mechanism, such as Rotary Position Embedding (RoPE) and other relative position encoding schemes. The method of offline precomputing involves calculating the outputs of the eliminated operations and components for each of the original vocab_size embedding-vectors, where vocab_size is the size of the embedding vocabulary. One embodiment of the invention removes the feedforward network and the attention query, key, and value projections from the first transformer layer of the encoder and the decoder stacks.

The invention discloses a fault-tolerant reconstruction method for sensing faults of a spatial dexterous hand, and relates to the field of robot control. In order to solve the defects that in the prior art, a dexterous hand on-orbit sensor is weak in fault-tolerant capability, insufficient in model generalization capability and incapable of being adaptive to complex space environment changes, the technical scheme provided by the invention comprises the following steps: constructing a conditional variation auto-encoder model, inputting a joint position and a rope length sequence, and constructing a model; time sequence features are extracted, and a reconstructed joint position mean value and a prediction variance are output; the reconstructed joint position mean value is compared with actual joint data, a dynamic confidence interval is generated, and whether the data breaks down or not is judged; for the data judged to be normal, information entropy is calculated based on the prediction variance, and data samples with the information entropy exceeding a set threshold value are screened and written into a dynamic buffer area; and when the buffer data is fully loaded, incremental learning is triggered, and model parameters are updated by utilizing the buffer data. The method is suitable for real-time detection and fault-tolerant reconstruction of sensor faults in the long-term on-orbit operation process of the space dexterous hand.

The invention relates to the technical field of intelligent traffic, and discloses a multi-modal large model context understanding method for a high-speed abnormal exposure scene, which comprises the following steps of: 1, aiming at a complex space-time structure of video data, encoding by adopting a RoPE (Rotation Position Element) variant to obtain a time sequence characteristic, and 2, combining the time sequence characteristic with a text characteristic to obtain a multi-modal large model context understanding result; the method comprises the following steps of: 1, inputting time sequence + text multi-modal data, 3, dynamically predicting the weight of a Laplacian pyramid decomposition kernel according to the characteristics of the input multi-modal data, 4, extracting and converting the characteristics in each layer of a characteristic pyramid, reconstructing the characteristics to original resolution layer by layer, and generating a corrected image, and 5, in the training process, generating a corrected image. And 6, inputting the corrected image into an event understanding module, identifying and classifying various events in the scene, and realizing semantic understanding.

A machine learning legal natural language generation computing system for developing a pre-trained and instruction fine-tuned MLLNLG generative language model that aligns with human instructions and preferences using causal language modeling for next-token prediction. The training method employs a scaled version of RoPE to compress token position indices, to manage longer sequences beyond the computing hardware devices (GPU's) DRAM memory capacity. This system is optimized for various legal NLP and NLG tasks by receiving input tasks from users through an I / O component. The system enables efficient pre-training of a generative legal language model from scratch on a single computing hardware device (GPU), achieving an efficient MFU of 41.35 and through weight tying, the system reduces model parameters compared to LLMs, eliminating the need for multiple GPUs. Tailored for the legal domain, the model excels in processing legal jargon and demonstrates advanced reasoning capabilities, outperforming existing models in summarization tasks.

The invention discloses a control method of a space rope-driven mechanical arm based on model-free reinforcement learning, which comprises the following steps: S1, a simulation environment is established for the space rope-driven mechanical arm by utilizing multi-joint contact dynamics, the space rope-driven mechanical arm comprises a floating base and a rope-driven mechanical arm, and the floating base is connected with the rope-driven mechanical arm; the first end of the rope-driven mechanical arm is connected to the floating base, and the other end of the rope-driven mechanical arm is the tail end point of the space rope-driven mechanical arm. And S2, a reinforcement learning framework based on a dense reward function is constructed, the reward function is designed so as to control the space rope-driven mechanical arm, and the reward function is related to the distance between the tail end point of the space rope-driven mechanical arm at each moment and the target point. According to the control method of the space rope-driven mechanical arm based on model-free reinforcement learning, generalization of strategies can be improved, and the capacity of an intelligent agent for executing complex tasks is enhanced.

The invention discloses a vulnerability repair model construction method based on abstract syntax tree position coding. The method comprises the following steps: step 1, obtaining a 2D position sequence; and step 2, expanding rotation position encoding (Rotation Position Encoding, RoPE) to adapt to the 2D position sequence, so that a position fusion module is constructed. And step 3, inserting the position fusion module into a transformer-based model to form a final vulnerability repair model. By means of the method, the understanding ability of the model for the code structure is enhanced, and irrelevant attention calculation is reduced.

The invention provides a multi-branch sequence recommendation method based on dynamic channel fusion. The method aims at solving the problems that in sequence recommendation, data have a large amount of noise and are sparse, and existing model prediction is smooth. The method comprises the following steps: on one hand, enabling a user sequence to pass through an embedding layer, introducing position information by utilizing RoPE rotation position coding, and extracting long-term dependency of a user in the sequence through a multi-feature channel feature network to obtain a new embedded Eglobal, a short-term interest and long-term dependency interactive embedded Ecross, an embedded Eema after exponential smoothing filtering and an embedded Efreq after frequency domain modulation; and on the other hand, in order to fully fuse the characteristics of different channels, a traditional hyper-parameter fusion mode is abandoned, and a pooling network layered adaptive fusion mode is adopted. Firstly, four channel features are divided into two groups including a self-attention group and a trend group, features of different channels are learned through Squeeze average pooling, learned parameters are activated through Excitation by adopting sigmoid to obtain channel weights, and feature representations of different groups are sent into a gating network for weighted summation loss calculation.

1. The name of the design product: data line (hanging rope 1). 2. The use of the design product: for data line. 3. The design points of the design product: in shape. 4. The picture or photo that best indicates the design points: perspective view.

1. The name of the design product: rope cover. 2. The use of the design product: for high-altitude operation, protect the rope from damage. 3. The design points of the design product: in shape. 4. The picture or photo that best shows the design points: perspective drawing.

1. The name of the design product: keyboard bottom shell (beautified slot shape). 2. The use of the design product: the design is mainly used for stabilizing the keyboard. Design 1 is a bottom shell without a rope buckle. Design 2 is a bottom shell with a rope buckle and a bottom shell without a rope buckle. 3. The design points of the design product: in shape. 4. The picture or photo that best indicates the design points: design 1 perspective view. 5. Design 1 is designated as the basic design.

The application relates to a long text extrapolation method and system based on period continuation of a large model, which is suitable for a ROPE position coding-based Transformer architecture model and comprises the following steps: 1) adaptively modifying a ROPE strategy; firstly, the ROPE coding strategy in the source code of the Transformer architecture model needs to be positioned; secondly, the ROPE coding strategy is updated by using an Extra-PE strategy; 2) fine-tuning the Transformer architecture model; and 3) performing reasoning by using the trained Transformer architecture model. The application converges the position coding distribution domain in a high-dimensional subspace to realize the long text extrapolation capability of the large model, and can guarantee that the attention distribution of the model in the training length window is not damaged.

1. Name of the product in this design: Carabiner (A2503TN). 2. Purpose of this design: For rope connection. 3. The key design feature of this product is its shape. 4. The image or photograph that best illustrates the key design points: Design 2 3D diagram. 5. Design 2 is designated as the basic design.

1. The name of the design product: string insertion buckle (M1372). 2. The use of the design product: buckle. 3. The design points of the design product: the combination of shape and pattern. 4. The picture or photo that best shows the design points: combined use state reference drawing.

1. The name of the product of this design: electronic label box. 2. Purpose of this product design: Used for event recording, a reusable electronic timing label box that is fixed on a rope by clamping it with a buckle on the back of the product. 3. The design key point of this appearance design product lies in the combination of shape and pattern. 4. The picture or photo that best illustrates the design points: Stereoscopic diagram 1.

1. Name of the product in this design: Rope Adjuster. 2. Purpose of this design: As a rope adjuster. 3. The key design feature of this product is its shape. 4. The image or photograph that best illustrates the design's key points: a 3D model.

The invention discloses an even-number multi-strand braided rope action recognition and deviation rectification method for AR-assisted teaching, and belongs to the technical field of computer vision and augmented reality teaching. The method comprises the steps that a rope braiding operation video stream of a user is collected through an RGB-D camera, hand skeleton key points and rope tracks are jointly extracted, a finger-rope interaction topological graph is constructed, and the finger-rope interaction topological graph is obtained; spatial topological features and time sequence dependence features are extracted by using a graph convolution time sequence attention network, the rope braiding step category and completion degree are accurately identified, a standard track is obtained according to a rope braiding knowledge graph, and the deviation between user operation and the standard track is calculated. And superposing a virtual guide arrow and a virtual hand demonstration model in real time in an augmented reality environment to perform action correction. According to the method, indexes such as deviation accumulated values and correction response time are counted, skill scores are calculated through a braided rope quality evaluation model, teaching guide strength and difficulty levels are adjusted in a self-adaptive mode, a complete closed-loop intelligent teaching system is formed, and braided rope skill learning efficiency and effects are remarkably improved.

The present invention proposes an inverse kinematics data processing method suitable for a manipulator with a rope-driven pure rolling joint, aiming to solve the problems of strict joint limitations and equal angle constraints that are difficult to overcome with traditional methods. The method mainly comprises: firstly establishing a kinematic model of the manipulator, then constructing a proxy model to analyze the multiple branches of the inverse kinematics solution; then solving through geometric iteration, including the forward and backward reaching stages and the state updating stage, during which the out-of-limit joint angles are corrected; further, by detecting the out-of-limit situation and the iterative progress, random perturbations or branch switching measures are taken to avoid falling into local optimality; finally, according to the preset iterative termination conditions, the optimal joint angle solution is returned. The present invention improves the success rate and solution speed of the joint angle solution, is suitable for high-precision automated production and complex operation tasks, enhances the reliability of the manipulator operation, and is of great significance to the development of the rope-driven manipulator technology.

The invention discloses a two-channel electric power entity recognition method and system based on BIBASRU and improved Reformer, and relates to the technical field of natural language processing and electric power information extraction, and the method comprises the steps: employing a MacBERT model to carry out pre-training semantic coding, and extracting context semantic features; constructing a BIBASRU structure as a first channel; a Reformer model based on ROPE optimization is constructed to serve as a second channel, and learnable position embedding is replaced with an ROPE coding mechanism; designing a dual-channel fusion mechanism, and dynamically weighting and integrating two paths of feature outputs; and finally, introducing a CRF layer to carry out sequence labeling decoding. According to the method, high-precision and high-efficiency entity recognition is realized in professional texts such as electric power regulations, scheduling logs and equipment reports, the method is remarkably superior to traditional BERT-BiLSTM-CRF models and the like, and the method is suitable for intelligent power grid knowledge graph construction and automatic information extraction scenes.

1. The name of the design product: easy rope puller. 2. The use of the design product: the design product is used for the puller of rope or braid. 3. The design points of the design product: in shape. 4. The picture or photo that can best show the design points: perspective drawing.

1. Name of the product in this design: Rope cutter (2251). 2. Purpose of this design: To cut ropes. 3. The key design feature of this product is its shape. 4. The image or photograph that best illustrates the design's key points: a 3D model.

The invention provides a drawing rope tying binding device which comprises a punching assembly and a knotting assembly, meanwhile, through rotation of a third stepping motor, accurate reciprocating movement of an arc-shaped variable-diameter needle head in the axis direction of a through hole of a punching platform is achieved, it is ensured that the punching position is accurate each time, and under driving of a second stepping motor, a fluted disc is driven to rotate, so that the binding efficiency is improved. The rope clamping mechanism and the knotting mechanism are driven to act through the first sector rack, the second sector rack and the third sector rack correspondingly, and clamping, knotting and other operations of a rope are achieved. The arc-shaped block drives the rope releasing mechanism to act at a proper time to complete releasing of a rope knot so as to complete binding of the pulling rope knot, compared with a buckled stitching needle, the binding mode is easier to manually open, book paper which is mistakenly fixed can be easily adjusted or added paper can be disassembled, 180-degree full-width opening can be achieved after binding, the rope material is soft, and the binding effect is good. And more comfortable writing and reading experience can be provided.

1. Name of the product in this design: Carabiner (A2510TN). 2. Purpose of this design: For rope connection. 3. The key design feature of this product is its shape. 4. The picture or photo that best illustrates the key design points: Design 1 3D diagram 1. 5. Design 1 is designated as the basic design.

The present invention discloses a method for constructing a privacy data perception model. By using a simplified RoformerBERT model, combining the attention mechanism and RoPE encoding to achieve the automatic perception of social network privacy information, accurately locate the position of privacy information, and using the BI-LSTM model to encode the text information multiple times, extract context features hierarchically, and then use the EGP algorithm to discriminate privacy entities, effectively improving the parameter utilization rate and the recognition accuracy of the model. For the self-constructed Weibo privacy dataset and privacy data in the real social network, the FDPER model can accurately identify and locate the user's privacy leakage information and feedback this privacy data to the user. In the specific comparative analysis of identifying privacy entity data, the effectiveness and accuracy of privacy recognition are also confirmed by the reproduction rate.

The invention provides a positioning data processing method, system and device based on an intelligent rope belt adjusting device, relates to the technical field of high-precision positioning, and particularly relates to a rope belt data management platform which preprocesses received positioning data, stores effective information and carries out trajectory prediction according to the received data when communication is interrupted, so that the positioning data can be accurately positioned. The positioning fault problem caused by data missing is effectively avoided, the continuity of the motion track is guaranteed, deviation is monitored and predicted in real time by means of the speed-motion direction coupling anomaly, track anomaly can be accurately recognized even if noise interference exists, an optimization mechanism is triggered in time to prevent error accumulation, and the accuracy of the motion track is improved. The method comprises the following steps of: receiving off-line positioning data stored by an intelligent rope belt adjusting device in an interruption period, comparing and analyzing the actually measured data and the predicted trajectory after communication is recovered, realizing dynamic iteration of a prediction algorithm, and finally maintaining high reliability and accuracy of the positioning data in a complex abnormal scene.

1. The name of the product of this design: Rope Puller (PS02). 2. Purpose of this design product: for fitness. 3. The design key point of this appearance design product lies in the shape. 4. The picture or photo that best illustrates the design points: three-dimensional picture.

The invention discloses a variable-weight comprehensive error compensation method for a rope-driven super-redundant mechanical arm, and belongs to the technical field of robots. According to the method, modeling compensation is conducted on rope hole gaps and rope elastic deformation, and a variable-weight geometric error compensation model is provided by combining the characteristics of a series structure of the rope-driven super-redundant mechanical arm. According to the method, a rope driving error compensation method is put forward according to two key factors in rope driving, namely a rope hole gap and rope elastic deformation; secondly, constructing a mechanical arm geometric error analysis model; secondly, a variable-weight geometric error compensation model is provided for the geometric error of the mechanical arm; and finally, parameter identification is completed, and therefore comprehensive error compensation comprehensively considering the rope driving error and the mechanical arm geometric error is achieved.

1. The name of this design product: Pull head with rope (NAZ8684). 2. Application of this design product: zipper pull. 3. The key design feature of this product is its shape. 4. The image or photograph that best illustrates the design's key points: a 3D model.

1. Name of the product in this design: Rope retractor (single). 2. Application of this design: This design is used for rope retrieval in industrial, sports, fire rescue and other fields. 3. The key design feature of this product is its shape. 4. The image or photograph that best illustrates the design's key points: 3D view 1.

1. Name of the product in this design: Cord-threading buckle component (M1350). 2. Purpose of this design: Fastener. 3. The key design feature of this product is its shape. 4. The image or photograph that best illustrates the design's key points: 3D rendering 1.