Abstract: A device and a computer-implemented method for machine learning. The method includes: providing a latent variable of a diffusion model representing the synthetic digital image, wherein providing the latent variable comprises sampling the latent variable from random noise, or providing a digital reference image, and adding noise with the diffusion model to the digital reference image to determine the latent variable, or mapping the digital reference image with an encoder to an input of the diffusion model, and adding noise with the diffusion model to the input to determine the latent variable, mapping the latent variable with the diffusion model depending on parameters of the diffusion model to features of the diffusion model, mapping the features with the diffusion model depending on parameters of the diffusion model to an output of the diffusion model.

Abstract: A method for generating an image. The method includes: providing a randomly drawn image or representation thereof as input of a sequence of layers of a neural network which includes a cross-attention layer; providing a first input, which is a representation of the input of the sequence determined by layers preceding the cross-attention layer or is the input to the sequence of layers, and providing a text embedding characterizing a description of the image to be generated as a second input; determining, by the cross-attention layer, an attention map based on the first and second inputs; optimizing the input provided to the sequence of layers based on a loss function which includes a term characterizing a negative total variation of the attention map; determining an output of the sequence of layers using the optimized input; and determining the image based on the determined output.

Abstract: A high-coercivity Nd—Fe—B sintered magnet and a preparation method and use thereof are provided. The sintered magnet contains the following components in percentage by a mass: 100%: 26-37 wt % of R, R being at least one rare earth element including Nd; 0.07-0.23 wt % of Mn; 0.8-1 wt % of B; 0.5-4 wt % of M, M comprising Cu and/or Al, and at least one selected from Co, Ti, Ni, Zr and Ga; and the remaining being Fe. A Mn-containing auxiliary alloy powder is mixed with a Mn-free neodymium-iron-boron main alloy powder to prepare the Nd—Fe—B sintered magnet. The Mn-containing auxiliary alloy powder, also contains at least one of metals Cu and Al. Mn can replace a part of Fe in the main phase, so that the amount of solid solution of beneficial elements in grain boundaries in the main phase is reduced, and the coercivity is improved.

Abstract: The present disclosure provides a sintered neodymium-iron-boron permanent magnet, a preparation method and use thereof. The permanent magnet described herein comprises a grain and a grain boundary phase. The grain boundary phase is located on an epitaxial layer of the grain. The grain boundary phase comprises at least an RH. The grain comprises at least Nd2Fe14B. In the grain boundary phase within a depth of 100 ?m from the surface to the center of the sintered neodymium-iron-boron permanent magnet, the area of the grain boundary phase with an RH content of more than 6 wt % accounts for 50% or more of the total area of the grain boundary phase. The present disclosure adopts an RH and an RL as the diffusion source for composite diffusion, significantly improving the coercivity of the permanent magnet and the utilization rate of the RH in the diffusion source.

Abstract: A junction termination with an internal field plate, the field plate structure and the junction termination extension region are folded inside the device to make full use of the thickness of the drift region in the body, thereby reducing the area of the termination and relieving the electric field concentration at the end of the PN junction. The breakdown position is transferred from the surface into the body of the original PN junction, and the withstand voltage of termination can reach to the breakdown voltage of the parallel plane junction. Under such design, a smaller area can be obtained than that of the conventional structure at the same withstand voltage.

Abstract: An image forming apparatus which is capable of reading an original, temporarily storing image data in a storage unit, and then transferring the image data to an image forming unit to improve FCOT in forming an image. A memory stores image data output from image reading units. An image forming unit forms an image on a recording medium based on the image data stored in the memory. A first memory output frequency in a first mode in which one side of an original is read and an image is formed on one side of a recording medium is lower than a second memory output frequency in a second mode in which both sides of an original are read and an image is formed both sides of a recording medium.

Abstract: A junction termination with an internal field plate, the field plate structure and the junction termination extension region are folded inside the device to make full use of the thickness of the drift region in the body, thereby reducing the area of the termination and relieving the electric field concentration at the end of the PN junction. The breakdown position is transferred from the surface into the body of the original PN junction, and the withstand voltage of termination can reach to the breakdown voltage of the parallel plane junction. Under such design, a smaller area can be obtained than that of the conventional structure at the same withstand voltage.

Abstract: An image reading apparatus includes a first unit, a second unit that is pivotable against the first unit, and a controller. The first unit includes a platen, a member, and a movable first reader. The second unit includes an original tray, a conveyance unit that conveys an original from the original tray, a static second reader. The controller determines an opening/closing state of the second unit against the first unit based on read data obtained when the second reader reads the member. The first reader reads the first side of an original conveyed by the conveyance unit and the second reader reads the second side of the original when the original placed on the original tray is read. The first reader reads the first side while moving under the original when the original placed on the platen is read.

Abstract: An image reading apparatus includes a first unit, a second unit that is pivotable against the first unit, and a controller. The first unit includes a platen, a member, and a movable first reader. The second unit includes an original tray, a conveyance unit that conveys an original from the original tray, a static second reader. The controller determines an opening/closing state of the second unit against the first unit based on read data obtained when the second reader reads the member. The first reader reads the first side of an original conveyed by the conveyance unit and the second reader reads the second side of the original when the original placed on the original tray is read. The first reader reads the first side while moving under the original when the original placed on the platen is read.

Abstract: An image forming apparatus which is capable of reading an original, temporarily storing image data in a storage unit, and then transferring the image data to an image forming unit to improve FCOT in forming an image. A memory stores image data output from image reading units. An image forming unit forms an image on a recording medium based on the image data stored in the memory. A first memory output frequency in a first mode in which one side of an original is read and an image is formed on one side of a recording medium is lower than a second memory output frequency in a second mode in which both sides of an original are read and an image is formed both sides of a recording medium.

Abstract: The present invention relates to IP network communication technology. It provides a kind of IP multicast data layer distribution method and system. The method includes: first code original multicast data into multiple sub-flow, separately specify different shunt sub-source for every sub-flow, distribute sub-group address, create sub-group state, etc, form shunt structure. Then the sink send join news to each shunt sub-source, join each shunt sub-source sub-group and become the leaf node of each sub-tree. And then each shunt sub-source determine the intermediate node according to the port receiving join news and construct multicast distribution-tree. If the overlapping portion between the multicast distribution trees produce congestion, in the congestion nodes network coding is used to eliminate congestion and construct coding sub-tree according to the way of building multicast distribution tree from the below of the coding nodes.