Abstract: This application discloses example data management methods. One example method includes obtaining, by a first terminal device, metadata of a target file in response to an access operation performed by a user on the target file through a first application, where the target file is stored in a local storage device, the first application is different from a second application, the second application is a dedicated application that is of the local storage device and that is installed on the first terminal device, and the first terminal device stores the metadata of the target file. The first terminal device sends an access request to the local storage device, where the access request includes the metadata of the target file. The first terminal device receives the target file sent by the local storage device.

Abstract: Embodiments described herein provide for optimizing a language model (LM) agent. In at least one embodiment, and LM agent comprises an “actor” LM and a “retrospective LM which provides reflections on attempts by the actor LM. The reflections are used to update subsequent prompts to the actor LM. Optimizing the LM agent comprises fine-tuning parameters of the retrospective LM while keeping parameters of the actor LM frozen. A gradient may be determined by a change in reward from the environment based on actions taken by the actor LM with and without a reflection of the retrospective LM. Using this gradient, parameters of the retrospective LM may be updated via backpropagation.

Abstract: A magnesium/titanium composite plate with large thickness ratios and a gradient heterothermal rolling bonding method are provided by the present disclosure, relating to the technical field of rolling bonding plates. The method includes following steps: assembling blanks according to a sequence of titanium strip, transition layer foil and magnesium alloy plate to obtain a composite blank; carrying out induction heating treatment on one side of the titanium strip of the composite blank, then rolling, and carrying out heat treatment on a composite plate blank after rolling to obtain the magnesium/titanium composite plate with large thickness ratios; and a thickness ratio of the magnesium alloy plate to the titanium strip is greater than or equal to 20:1.

Abstract: A magnesium/titanium composite plate with large thickness ratios and a gradient heterothermal rolling bonding method are provided by the present disclosure, relating to the technical field of rolling bonding plates. The method includes following steps: assembling blanks according to a sequence of titanium strip, transition layer foil and magnesium alloy plate to obtain a composite blank; carrying out induction heating treatment on one side of the titanium strip of the composite blank, then rolling, and carrying out heat treatment on a composite plate blank after rolling to obtain the magnesium/titanium composite plate with large thickness ratios; and a thickness ratio of the magnesium alloy plate to the titanium strip is greater than or equal to 20:1.

Abstract: A backlight module and a display apparatus. The backlight module includes: a back panel, including a base plate and a vertical plate connected with an edge of the base plate; the base plate is matched with the vertical plate to form an accommodation space; a light guide plate, located in the accommodation space and having a safety gap with the vertical plate of the back panel; and an optical film, located in the accommodation space and located on a side of the light guide plate away from the base plate. A surface of the light guide plate facing the optical film has a first positioning portion and a second positioning portion; the first positioning portion positions the optical film in a first direction, and the second positioning portion positions the optical film in a second direction. The first direction and the second direction are perpendicular to each other.

Abstract: The subject technology provides frequency-aware masked autoencoders for multimodal pretraining on frequency-based signals. An apparatus receives input data comprising frequency-based signal information associated with one or more modalities. The apparatus transforms the input data from a time domain to a frequency domain. The apparatus generates a frequency-embedded latent representation of the input data comprising time-domain and frequency-domain information. The apparatus also generates a masked frequency-embedded latent representation by masking one or more frequency components in the frequency-embedded latent representation. The apparatus produces a trained machine learning model by training a neural network to predict one or more masked frequency components of the frequency-embedded latent representation.

Abstract: A robust nonnegative matrix factorization (RNMF) method, in which an image sample set is split into a training set and a test set. The training set and the test set are normalized to map the image data from [0, 255] to [0, 1]. The training set matrix is pretrained by RNMF for decomposition. l2,1-deep incremental nonnegative matrix factorization (l2,1-DINMF) model is construed. The l2,1-DINMF model is configured to decompose the training set matrix into l+1 factors. After the basis matrix has been updated, and the samples of the training set and samples to be recognized are projected into a feature space. Feature representations of the test set are classified by a trained SVM classifier to obtain a predicted label, and the predicted label is compared with an actual label of the test set to calculate a recognition accuracy.

Abstract: A system and method for joint dynamic interest request forwarding and dynamic cache placement and eviction and provided within the context of the Named Data Networking (NDN) architecture. The system and method employ a virtual control plane that operates on the user demand rate for data objects in the network, and an actual plane that handles Interest Packets and Data Packets. Distributed algorithms within the virtual plane achieve network load balancing through dynamic forwarding and caching, thereby maximizing the user demand rate that the NDN network can satisfy. A method of congestion control is also provided to achieve optimal network fairness using the VIP framework.

Abstract: A mechanical lock for building component connection includes a plug element (1) and a lock bod part (2). The plug element (1) includes a plug shell (10), a plug pin structure (11), and a plug transmission structure. The plug pin structure (11) is movably installed in the plug shell (10). The plug transmission structure is in transmission connection with the plug pin structure (11). An end of the plug pin structure (11) is provided with a stop part (110). The lock bod part (2) includes a lock body shell (20), a bracing structure (21), and a lock body transmission structure. An insertion channel (23) matched with the plug pin structure (11) is formed in the lock body shell (20). The lock body transmission structure is in transmission connection with the bracing structure (21).

Abstract: Systems, methods and apparatus for sematic segmentation of 3D point clouds using deep neural networks. The deep neural network generally has two primary subsystems: a multi-branch cascaded subnetwork that includes an encoder and a decoder, and is configured to receive a sparse 3D point cloud, and capture and fuse spatial feature information in the sparse 3D point cloud at multiple scales and multi hierarchical levels; and a spatial feature transformer subnetwork that is configured to transform the cascaded features generated by the multi-branch cascaded subnetwork and fuse these scaled features using a shared decoder attention framework to assist in the prediction of sematic classes for the sparse 3D point cloud.

Abstract: A controllable rapid pressure loading technology for a large volume press is provided. A regular octahedra, plugs, diamond pistons, and a standard are included. The regular octahedra is provided with a cavity with openings at two ends of the cavity. The standard is placed in the cavity. The openings at two ends of the cavity are respectively blocked by the conductive plugs. The diamond piston is arranged between the standard and the plug. The diamond pistons arranged between the standard and the plug has high hardness and has the pressure transmission efficiency superior to that of a common ceramic plug, improving compression efficiency in sample cavity. A loading method of pre-charging a pressure to a pressure loading device first and then rapidly releasing to a pressure loading mold is used to match an improved pressure transmission component, which shortens the pressure loading time of the sample cavity.

Abstract: A macro image shooting method uses at least two macro cameras having different focal lengths. When a zoom ratio is within a first ratio range, an image captured by a first macro camera is displayed in a viewfinder frame. When the zoom ratio is a camera switching ratio, the image displayed in the viewfinder frame is switched from the image captured by the first macro camera to an image captured by a second macro camera. When the zoom ratio is greater than the camera switching ratio and is within a second ratio range, the image captured by the second macro camera is displayed in the viewfinder frame.

Abstract: The present disclosure provides a magnetic nano-drug with double targeting vascular endothelial growth factor (VEGF)-vascular endothelial growth factor receptor (VEGFR), and a preparation method and application thereof, and belongs to the technical field of biomedicine. The magnetic nano-drug of the present disclosure includes aminated ZnFe2O4 hollow porous magnetic nano-particles, lactate dehydrogenase-silk fibroin (LDH-SF), ethylene dichloride (EDC), and n-hydroxy succinimide (NHS). A mass ratio of the aminated ZnFe2O4 hollow porous magnetic nano-particles to the EDC to the NHS to the LDH-SF is (5-10):(1-5):(1-6):(1-8). The magnetic nano-drug according to the present disclosure can position a VEGF ligand through magnetic targeting of hollow zinc ferrite particles, use magnetic particles to regulate reprogramming of macrophages in an intervertebral disc so as to indirectly regulate down-regulation of the VEGF, and inhibit angiogenesis.

Abstract: A method, system, and computer program product that is configured to: receive at least one workload of a mixed addressing mode application; classify the at least one workload with artificial intelligence (AI) including a support vector machine (SVM) algorithm; match at least one agent to the at least one workload based on a workload class and tuning policies; execute workload polices of the at least one workload based on the workload class and the tuning policies; evaluate a transaction per second (TPS) and response time of the at least one workload; calculate a reward of the at least one workload; and train a plurality of models based on historical data corresponding to the evaluated TPS, the evaluated response time, and the calculated reward.

Abstract: The present disclosure relates to a method for high pressure regulation and control of photoelectric detection based on BiOBr, and relates to the technical field of photoelectric detection. An exemplary method includes inserting an insulation layer into a pressure chamber of a diamond anvil cell and adding BiOBr, putting a pressure-calibrating substance on a culet of the diamond anvil cell; pressurizing the pressure chamber by rotating a press bolt on the diamond anvil cell; and conducting photoelectric detection using the pressurized BiOBr, where two platinum sheets are disposed on the BiOBr as an electrode. The present disclosure enhances the photo-response speed and photo-responsivity of photoelectric detection.

Abstract: A system and method for joint dynamic interest request forwarding and dynamic cache placement and eviction and provided within the context of the Named Data Networking (NDN) architecture. The system and method employ a virtual control plane that operates on the user demand rate for data objects in the network, and an actual plane that handles Interest Packets and Data Packets. Distributed algorithms within the virtual plane achieve network load balancing through dynamic forwarding and caching, thereby maximizing the user demand rate that the NDN network can satisfy. A method of congestion control is also provided to achieve optimal network fairness using the VIP framework.

Abstract: Embodiments of the present disclosure provide a method, an electronic device, and a computer program product for synchronously accessing data. The method may include sending metadata associated with data to be moved by a user to a programmable network device associated with a persistent memory containing the data, so as to enable the programmable network device to move the data based on the metadata, wherein the programmable network device is a smart network interface card having a remote direct memory access function. The method may also include entering a hibernation state. In addition, the method may include exiting from the hibernation state in response to receiving a confirmation of operation completion from the programmable network device, so as to notify the user that an operation of moving the data is complete.

Abstract: Embodiments of the present disclosure relate to a method, a device, and a computer program product for data access. The method includes acquiring a priority of a workload associated with an input/output (I/O) instruction of a user, and determining, based on the priority, whether to perform data access to a persistent memory indicated by the I/O instruction by using a central processing unit (CPU). If it is determined not to perform the data access by using the CPU, the data access is performed by using a programmable data moving apparatus. The method according to the embodiments of the present disclosure can avoid that important workloads compete for CPU resources equally with secondary workloads, and alleviate blocking of workloads due to insufficient CPU resources, thereby improving the overall performance of persistent memory access.

Abstract: The present disclosure provides a method for predicting carbon consumption and carbon emission of a hydrogen-rich blast furnace (BF) based on a C-rd theory.

Abstract: A gas diaphragm driven rapid supercharging device based on a diamond anvil cell is provided which includes a gas diaphragm, a diamond anvil cell press, a hard steel shell and a gas release component. The gas diaphragm and the diamond anvil cell press are positioned inside the hard steel shell, the gas diaphragm is positioned below the diamond anvil cell press, a first opening is formed in a top of the hard steel shell, and a second opening is formed in a bottom of the hard steel shell. The gas release component passes through the second opening of the hard steel shell to be connected with the gas diaphragm through a gas path, and the gas release component is configured for releasing gas at a first inflation speed in a standard pressure experiment, and inflating the gas diaphragm at different driving pressures to obtain a pressure calibration curve.