Abstract: Embodiments of the disclosure relate to learning new watermark algorithms for artificial intelligence (AI) models for a data processing (DP) accelerator. In one embodiment, a system trains a watermark algorithm based on a predetermined set of criteria, where the watermark algorithm is trained to generate variations of the watermark algorithm. The system configures the watermark unit at runtime with a variation of the watermark algorithm for the watermark algorithm to be used by the DP accelerator.

Abstract: The invention discloses an exopolysaccharide from Rhodopseudomonas palustris and a method for preparing and use thereof, and the method for preparing comprises the steps of: 1) keeping a seed solution from Rhodopseudomonas palustris GJ-22 in a fermentation medium for fermentation culture to obtain a fermentation broth; 2) centrifuging the fermentation broth to take the supernatant, which is treated by alcohol precipitation after filtration, and then collecting the pellet from alcohol precipitation by centrifugation to obtain crude polysaccharide; 3) removing proteins from the rude polysaccharide using protease enzymolysis method and Sevag method, followed by dialysis treatment with distilled water to remove small molecules and organic solvent to obtain a polysaccharide sample; 4) purifying the polysaccharide sample through an anion exchange column and a molecular exclusion chromatography column obtain the exopolysaccharide from Rhodopseudomonas palustris.

Abstract: The present disclosure provides neodymium-iron-boron magnetic body having gradient distribution, comprising an ease-to-demagnetize zone and a hard-to-demagnetize zone, wherein in a direction perpendicular to magnetization direction, remanence of the ease-to-demagnetize zone is less than remanence of the hard-to-demagnetize zone, and coercivity of the ease-to-demagnetize zone is greater than coercivity of the hard-to-demagnetize zone; and along the direction perpendicular to magnetization direction, the remanence and the coercivity of the ease-to-demagnetize zone are respectively a constant value, and the remanence and the coercivity of the hard-to-demagnetize zone are respectively a constant value. Due to the gradient distribution of remanence and coercivity of the neodymium-iron-boron magnetic body provided by the present application, the remanence, coercivity, magnetic flux and surface magnetic field of the neodymium-iron-boron magnetic body are optimized.

Abstract: A reference signal sending or receiving method includes: determining, by the network device, a plurality of resource elements REs used to carry a first CSI-RS, where the plurality of REs are distributed in a plurality of resource units, and in each resource unit, a plurality of REs used to carry the first CSI-RS are located on a plurality of subcarriers in a same symbol, values of the first CSI-RS carried on at least two REs are different, and values of the first CSI-RS are loaded to the plurality of REs in the resource unit by using a first multiplex code; and sending, by the network device, the first CSI-RS to the terminal device by using the plurality of REs.

Abstract: An adjustable deforming composite structure based on a hydrogen-induced expansion effect and a preparation method therefor are provided. The hydrogen-induced expansion effect means metals absorb hydrogen under a hydrogen-containing atmosphere and at a temperature to produce a volume expansion effect. Reactions between the metals and hydrogen are reversible reactions. When a hydrogen partial pressure is reduced or the temperature is increased, the hydrogen in the metals is removed, and the metals are restored to an original shape. Under a stimulation of external hydrogen and heat, a composite of a hydrogen-absorbing metal and other non-hydrogen-absorbing materials undergo an adjustable deformation according to a design, and a material undergoes reversible shape changes. The preparation method is applied to composite materials for a 4D printing and is used for an intelligent shape adjustment at a medium to high temperature.

Abstract: The presently disclosed subject matter relates to multi-material electrical contacts, and methods of making multi-material electrical contacts.

Abstract: Embodiments of this application disclose a resource configuration method and a device. The method includes: generating, by a transmit end device, configuration information, where the configuration information is used to indicate at least one resource bundling size, the resource bundling size belongs to a resource bundling size set, and the resource bundling size set corresponds to at least one system configuration parameter; and sending, by the transmit end device, the configuration information.

Abstract: The present invention relates to a device for integrally processing a housing of a Computing, Communication, and Consumer (3C) product. The device includes a loading frame and an upper frame, where a material receiving device is arranged in the loading frame, a width of the loading frame is greater than that of the housing, a grinding device is arranged below the upper frame, and the grinding device includes a grinding movement device; the grinding movement device is connected to a grinding installation block, an outer side of the grinding installation block is provided with grinding lifting balls, a grinding arc block is embedded in an inner side of an upper portion of the grinding installation block, and the upper frame is also provided with a material pressing device. An outer side of the loading frame is provided with a feeding and discharging device.

Abstract: The present invention provides a quantum dot composite brightness enhancement film and a method for manufacturing same, relating to the field of optical thin films. The quantum dot composite brightness enhancement film includes a quantum dot film layer formed by a back coating layer, a substrate layer, a first barrier layer, a quantum dot layer, and a second barrier layer which are sequentially attached, a composite brightness enhancement film layer formed by a diffusion layer, a core layer, and a prism layer which are sequentially attached, and an Optically Clear Adhesive (OCA) layer connecting the second barrier layer and the diffusion layer.

Abstract: An electrode for display includes at least two display electrodes arranged at intervals along first direction, each including a body electrode and an end electrode; the at least two display electrodes include first and second display electrodes, body electrodes of the first and second display electrodes being parallel in first direction with a first preset pitch; among two side edges of the end electrode of the first display electrode opposite in first direction and two side edges of the end electrode of the second display electrode opposite in first direction, at least two side edges are unparallel; a minimum distance in first direction between one side edge of the end electrode of the first display electrode close to the second display electrode and one side edge of the end electrode of the second display electrode close to the first display electrode is greater than or equal to first preset pitch.

Abstract: According to one embodiment, a broadcast request is received from a host that hosts an application that initiated a broadcast message to be broadcast to one or more DP accelerators of a plurality of DP accelerators coupled to the host, where the broadcast request includes one or more DP accelerator identifiers (IDs) identifying the one or more DP accelerators. A broadcast session key for a broadcast communication session to broadcast the broadcast message is received from the host. For each of the one or more DP accelerator IDs, a public key of a security key pair corresponding to the DP accelerator ID is identified. The broadcast message is encrypted using the broadcast session key. The broadcast session key is encrypted using the public key. The encrypted broadcast message and the encrypted broadcast session key are transmitted to a DP accelerator identified by the DP accelerator ID.

Abstract: In one embodiment, a computer-implemented method of a DP accelerator performing an encryption or decryption operation includes receiving, by the DP accelerator, a command and input data for the DP accelerator to encrypt or decrypt. The command is one of: encrypt the input data or decrypt the input data. The method further includes encrypting, or decrypting, by the DP accelerator, the input data according to the command; and providing the encrypted or decrypted input data to the host device. The host device and DP accelerator may exchange one or more keys and such keys can be used to establish a secure link between the host device and DP accelerator and/or to use for encryption or decryption. One or more of the keys may be based upon a root key or key pair of the DP accelerator and can be stored in a secure storage of a security unit of the DP accelerator.

Abstract: A packet transmission method includes generating, by a first router, a first routing information update packet; generating, by the first router, a first source-tracing packet, where the first source-tracing packet is used to indicate that the first router is a source router generating the first routing information update packet; and determining, by the first router, that a second router adjacent to the first router has a source-tracing capability, and sending the first source-tracing packet to the second router.

Abstract: Embodiments of the present invention provide a method and an apparatus for sending a reference signal, and a method and an apparatus for obtaining a reference signal. The method for sending a reference signal includes: determining, based on a resource that is allocated to a reference signal in a basic pattern corresponding to the reference signal and a resource that is allocated to at least one basic pattern in a physical layer transmission unit, a resource occupied by the reference signal in the physical layer transmission unit; and sending the reference signal through the determined resource. It can be learned that, according to the technical solutions provided in the embodiments of the present invention, arrangement of the reference signal in the physical layer transmission unit can be flexibly set.

Abstract: A method for establishing mathematical model of relationship between spontaneous imbibition volume and time of porous medium includes sample pretreatment, fully-saturation and centrifugal experiments and NMR T2 measurement. First, two rock core samples of predetermined size are selected for cleaning and drying. The first rock sample is vacuumed and injected with water to obtain a saturated sample for NMR T2 measurement. Then, spontaneous imbibition experiment is conducted on another sample, and T2 measurements are conducted to obtain the water distribution and migration characteristics during the imbibition process. Next, the calculation of the imbibition permeability, average capillary pressure and surface relaxivity are conducted based on the NMR data obtained from two samples. Finally, substitute these parameters into the Handy relationship to obtain a new NMR-based mathematical spontaneous imbibition model.

Abstract: A flip LED chip module and a method for manufacturing LED chip module are provided. In the manufactured flip LED chip module, the second glue layer for reducing light energy is formed on the front surface of each of the flip LED chips. Therefore, the difference between the light intensity of the light along the normal direction of the flip LED chip and the light intensity of the light at the four sides of the flip LED chip can be reduced, thus improving the uniformity of the color of the light. Hence, the color of the light of the LED device manufactured by the LED chip module can be more uniform, thus improving the quality of the LED device and the user satisfaction.

Abstract: In one embodiment, a computer-implemented method performed by a data processing (DP) accelerator, includes receiving, at the DP accelerator, first data representing a set of training data from a host processor; receiving, at the DP accelerator, a watermark kernel from the host processor; and executing the watermark kernel within the DP accelerator on an artificial intelligence (AI) model. The watermark kernel, when executed, is configured to: generate a watermark, train the AI model using the set of training data, and implant the watermark within the AI model during training of the AI model. The DP accelerator then transmits second data representing the trained AI model having the watermark implanted therein to the host processor. In an embodiment, the method further includes receiving a pre-trained AI model and the training is performed for the pre-trained AI model.