Abstract: Provided are a base editor, specifically, a base editor in the form of a ribonucleoprotein (RNP) complex, and uses in gene correction in vivo using the same. When using the base editor in the form of an RNP complex of the present disclosure, target genes can be effectively corrected by reducing off-target effects.

Abstract: An image processing apparatus, including: at least one processor; and a memory configured to store one or more instructions which, when executed by the at least one processor, cause the image processing apparatus to: obtain a neural network model for performing image quality processing on an input image; calculate a plurality of gradients by partially differentiating weights of the neural network model with respect to a loss of the neural network model, by applying training data corresponding to the input image to the neural network model; remove at least one gradient from among the plurality of gradients by applying a gradient mask including gradient pruning information to the plurality of gradients; train the neural network model by updating the weights of the neural network model based on one or more remaining gradients from among the plurality of gradients; and obtain a quality-processed output image based on the input image, using the trained neural network model

Abstract: The present invention relates to a customized stretchable wearable device and a manufacturing method therefor, enabling the manufacturing of a customized stretchable wearable device which fits the body characteristics of a wearing subject who is wearing the wearable device, thereby enabling good wear comfort and sense of operation, and also a beautiful exterior.

Abstract: A method of designing an optical system including a dome light and coaxial light, includes: determining a distance between the dome light and the coaxial light based on a radius reduction according to a hole of the dome light; determining a size of the coaxial light based on an optical path of a light ray emitted from the coaxial light; and determining a structure of a printed circuit board (PCB) in the dome light based on an optical path through which the light ray emitted from the coaxial light is reflected by an object.

Abstract: A neural network operation method includes: receiving an input vector sequence including a plurality of channels; performing a first convolution operation on a first input vector of the input vector sequence; and performing a second convolution operation on a second input vector of the input vector sequence that is adjacent to the first input vector in a channel direction.

Abstract: The present disclosure relates to a triboresistive touch sensor capable of generating electric power by electrification and electrostatic induction, and thus sensing a touch position without a grid.

Abstract: System for measuring the thickness and properties of a thin film using a spatial light modulator according to the inventive concept includes: a spatial light modulator; a first beam splitter; an objective lens; a second beam splitter; a first camera configured to acquire a surface image of the sample; an optical fiber equipped with a light receiving unit to receive reflected light passing through a certain area of a back focal plane of the objective lens; and a spectrometer configured to measure an intensity of light received from the optical fiber and output the measured intensity as an electrical signal.

Abstract: An electronic device includes: a host processor configured to control an operation of the electronic device; accelerators of heterogeneous hardware types configured to exchange data with each other through direct communication; and a control unit configured to convert a command received from the host processor, based on a type of each of the accelerators and transfer a result of the converting to a corresponding accelerator among the accelerators.

Abstract: An operation method of performing a neural network operation of fully homomorphic encrypted data is provided. The operation method includes: receiving data for performing the neural network operation and receiving a parameter for generating an approximation polynomial corresponding to the neural network operation; obtaining layer information corresponding to layers of a neural network model, the layer information based on the data; determining importances of the layers, respectively, wherein the determining of the importances is based on the parameter and the layer information; generating an approximation polynomial approximating the neural network operation for each of the layers, wherein the generating is based on the layer importance; and generating an operation result by performing the neural network operation based on the approximation polynomial, wherein the parameter includes a computation time condition that the neural network operation must satisfy.

Abstract: The present disclosure relates to successive approximation register analog-to-digital converters. An example successive approximation register analog-to-digital converter includes a first sampling and holding circuit that samples an analog signal at a first point in time and generates a first input voltage, a second sampling and holding circuit that samples the analog signal at a second point in time and generates a second input voltage, and a first analog-to-digital converter. The first analog-to-digital converter performs a feed forward equalization function by receiving the first input voltage and the second input voltage, sampling the first input voltage and the second input voltage, and outputting a multi-bit digital signal based on a sampling result of the first input voltage and a sampling result of the second input voltage.

Abstract: A method and apparatus for measuring properties of a magnetic body are provided. The method includes applying a uniform magnetic field to generate a magnetic space; rotating a magnetic body in the magnetic space about a predetermined axis; measuring a first resistance value; measuring a second resistance value; and calculating a magnetic anisotropy constant of the magnetic body, based on the first and second resistance values, wherein the first resistance value is a measured resistance value of the magnetic body according to a magnetic field sweep in a direction of a magnetization hard axis of the magnetic body, and wherein the second resistance value is another measured resistance value of the magnetic body according to an angle of an external magnetic field, the angle of the external magnetic field being an angle formed by a magnetization easy axis of the magnetic body and a direction of the uniform magnetic field applied to the magnetic body.

Abstract: A processor-implemented method includes receiving data for performing a neural network operation of homomorphic encrypted data and a parameter for generating an approximate polynomial corresponding to the neural network operation, obtaining layer information corresponding to each of a plurality of layers configuring a neural network based on the data, determining layer importance corresponding to each of the plurality of layers, based on the parameter and the layer information, generating an approximate polynomial approximating the neural network operation for each of the plurality of layers, based on the layer importance, and generating an operation result by performing the neural network operation based on the approximate polynomial.

Abstract: A method for processing homomorphic ciphertexts includes: receiving an input of an instruction for sorting regarding a plurality of homomorphic ciphertexts; sorting the plurality of homomorphic ciphertexts by using a sorter which can sort 3 more homomorphic ciphertexts in a single stage; and outputting the sorting result. The sorter performs sorting by using a comparison function that selectively outputs a bigger value or a smaller value between two input values.

Abstract: A video processing method is provided. The video processing method obtaining a clip representation based on a frame index and a global representation of an input video, obtaining a residual frame representation based on the frame index and a residual representation of the input video using a residual neural network and outputting a final frame representation generated based on the clip representation and the residual frame representation.

Abstract: A cryptography processor includes at least one interface circuit configured to communication with at least one of a host device and a semiconductor memory device, at least one register configured to store at least one of intermediate data of a common operation and control data received from the host device, a plurality of common operation circuits, respectively configured to perform different operations, and a plurality of function circuits configured to control at least one of the plurality of common operation circuits to perform a workload allocated from the host device. At least one of the plurality of common operation circuits may be configured to be controlled by different function circuits, among the plurality of function circuits, to execute different functions of a cryptographic operation, and the workload may be configured to execute a function of the cryptographic operation.

Abstract: Disclosed is an electronic device that includes a window, a display panel that is disposed under the window and that has a display area and a non-display area defined therein, a lower member that is disposed under the display panel and that includes a lower conductive layer, a transmission line disposed between the display panel and the lower member, a conductive partition wall spaced apart from the transmission line and disposed to overlap the display area, and a housing coupled with the window to define a receiving space in which the display panel, the lower member, and the transmission line are disposed.

Abstract: A processor-implemented method includes, while training a neural network (NN) using a current training mode selected from a plurality of training modes for training of the NN, measuring time data of a plurality of sub-operations for the training of the NN; based on the time data, determining a computation time to perform computation operations among the plurality of sub-operations and a communication time to perform communication operations among the plurality of sub-operations; based on a comparison result of the computation time and the communication time, selecting a next training mode from the plurality of training modes; and training the NN based on the next training mode.

Abstract: The present disclosure provides an all-in-one image quality improvement model providing method, which restores quality of images degraded due to multiple image quality degradation factors and is performed by a server, including collecting low-quality images of which image quality is degraded due to at least one of the multiple image quality degradation factors and determining an image quality degradation factor that is a cause of image quality degradation of a corresponding image, constructing an integrated restoration network and a single restoration network to restore the low-quality images with respect to the determined image quality degradation factors.

Abstract: A double sided meta lens includes a substrate having a first surface and a second surface facing each other, a first meta lens that includes a plurality of first nanostructures arranged on the first surface of the substrate and exhibits a first phase profile, and a second meta lens that includes a plurality of second nanostructures arranged on the second surface of the substrate and exhibits a second phase profile. The first phase profile may be configured such that a central region of the first meta lens operates as a convex lens and a peripheral region of the first meta lens operates as a concave lens, and the second phase profile may correspond to a focusing lens and has a section in which an absolute value of a phase slope increases with distance from a center of the second meta lens.

Abstract: Proposed is a deployable high-rigidity robot arm in a folding and rolling type. the deployable high-rigidity robot arm includes a hub capable of being rotated forward or backward, a plurality of panels having a sliding and folding structure that is capable of being wound or unwound on the hub, a shape-forming apparatus coupled to the plurality of panels and configured such that the plurality of panels is unwound from the hub and a bent structure is formed, and a shape-maintaining apparatus configured to maintain and extend the bent structure formed by the shape-forming apparatus. Furthermore, the plurality of panels is configured to be telescoped by a rotation of the hub while the plurality of panels maintains the bent structure.