Abstract: Provided are a method and apparatus for compressing a neural network model by using hardware characteristics. The method includes: obtaining a target number of output channels of a target layer to be compressed, from among layers included in the neural network model that is executed by hardware; adjusting the target number of output channels to meet a certain purpose, based on at least one of a first latency characteristic for output channels of the target layer and a second latency characteristic for input channels of a next layer, according to the hardware characteristics of the hardware; and compressing the neural network model such that the number of output channels of the target layer is equal to the adjusted target number of output channels.

Abstract: A cover assembly includes a mouthpiece including one end, that is configured to come into contact with a mouth of a user, and another end opposite to the one end, the mouthpiece configured to move to an open position and a closed position by rotation of the mouthpiece around the other end; and an accommodation unit for accommodating the one end of the mouthpiece in the closed position. The accommodation unit includes a button unit including a body, the button unit configured to contact the one end of the mouthpiece and to be displaced in a longitudinal direction based on the one end being pressed; and a locking unit including a body, the body of the locking unit including a first portion that is configured to contact the one end of the mouthpiece and a second portion that is configured to contact the button unit.

Abstract: Disclosed are wood preforming devices for manufacturing a crash pad for a vehicle including a real wood sheet. A wood preforming device for manufacturing a crash pad for a vehicle includes a real wood sheet includes a lower press mold comprising a debossed portion provided on a portion on which a product is formed, and a support portion configured to support an upper press mold, in response to the lower press mold and the upper press mold pressing each other, the support portion having a protrusion for fixing a real wood sheet, the upper press mold having an embossed portion corresponding to the debossed portion of the lower press mold, and a movable core provided on the debossed portion of the lower press mold and being configured to guide the real wood sheet by moving upward from the debossed portion, in response to the upper press mold moving downward.

Abstract: A display device comprising: first and second pixels; a first data line connected to the first pixel and configured to have data voltages applied thereto; and a second data line connected to the second pixel, the second data line being adjacent to the first data line, and configured to have the data voltages applied thereto, wherein the first data line includes a 1A-th data line which is in a first data layer, and the second data line includes a 2B-th data line which is in a second data layer different from the first data layer.

Abstract: The semiconductor device includes a substrate, a stack structure including gate patterns and interlayer insulating films that are alternately stacked on the substrate, an insulating pillar extending in a thickness direction of the substrate within the stack structure, a polycrystalline metal oxide film extending along a sidewall of the insulating pillar between the insulating pillar and the stack structure, a liner film having a transition metal between the insulating pillar and the polycrystalline metal oxide film, and a tunnel insulating film, a charge storage film, and a blocking insulating film which are disposed in order between the polycrystalline metal oxide film and the gate patterns.

Abstract: The present invention relates to a positive electrode active material and a lithium secondary battery using a positive electrode containing the positive electrode active material. More particularly, the present invention relates to a positive electrode active material that is able to solve a problem of increased resistance according to an increase in Ni content by forming a charge transport channel in a lithium composite oxide and a lithium secondary battery using a positive electrode containing the positive electrode active material.

Abstract: Described are anti-doppel antibody-drug conjugates, compositions comprising them, and related methods of treating doppel-associated diseases and conditions, including cancer.

Abstract: A multilayer electronic component has a body and a non-conductive resin layer. The non-conductive resin layer includes a body cover portion disposed in a region of an external surface of the body in which an electrode layer of an external electrode is not disposed, and an extending portion extending from the body cover portion between the electrode layer and a conductive resin layer of the external electrode, to thereby suppress arc discharge, improve bending strength, and improve moisture resistance.

Abstract: Disclosed is a window film for a display device including a light transmitting substrate including a polyamide including a structural unit represented by Chemical Formula 1 and a structural unit represented by Chemical Formula 2, and a hard coating layer laminated on one surface of the light transmitting substrate: in Chemical Formula 1 and Chemical Formula 2, each of A1, A2, B, and D is the same as defined in the detailed description.

Abstract: Described herein are doppel-targeting molecules (e.g., antibodies) useful for inhibiting pathological angiogenesis and treating diseases and conditions associated with pathological angiogenesis, such as tumors, cancers, atherosclerosis, tuberculosis, asthma, pulmonary arterial hypertension (PAH), neoplasms and neoplasm-related conditions, and for detecting doppel expression in a subject. Related compositions and methods also are described.

Abstract: Provided are a method and apparatus for performing a convolution operation for optimizing the arithmetic intensity of device.

Abstract: Disclosed are a training data generation device, a method for driving the same, and a computer-readable recording medium. The training data generation device according to an embodiment of the present inventive concept includes a data receiver configured to receive, as a feedback, a training result output by applying, to a learning model, medical data generated by pre-processing an assigned amount of pathological images; and a controller configured to secure training data for artificial intelligence (AI) or image analysis by correcting medical data of pathological images having errors checked based on the training result received as a feedback, and then reapplying the corrected medical data to the learning model.

Abstract: A semiconductor package and a method of manufacturing a semiconductor package. As a non-limiting example, various aspects of this disclosure provide a semiconductor package, and method of manufacturing thereof, that comprises shielding on multiple sides thereof.

Abstract: A semiconductor package and a method of manufacturing a semiconductor package. As a non-limiting example, various aspects of this disclosure provide a semiconductor package, and method of manufacturing thereof, that comprises shielding on multiple sides thereof.

Abstract: The invention is directed to an isolated population of mammal cells comprising about 75% or higher B-1a regula e PBS-treated tory cells expressing the cell surface inhibitory receptors lympho-cyte-activation gene 3 (LAG-3), programmed cell death protein 1 (PD-1), and C-X-C chemokine receptor type 4 (CXCR4), and secreting interleukin-27 (IL-27). The invention is also directed to methods of preparing and using the cell population to suppress the immune system and/or to treat or prevent diseases.

Abstract: Disclosed are a coupling agent for a rubber composition and a rubber composition for a tire containing the same. The rubber composition according to the present disclosure contains a coupling agent that may induce interactions between a rubber as a raw material and carbon black, thereby improving dispersibility of the carbon black in the rubber composition. Thus, a tire with low rolling-resistance and excellent wear resistance may be realized.

Abstract: An operation method of a first AN for a hybrid beamforming-based cooperative transmission in a C-RAN includes obtaining a spatial channel covariance with terminals subject to a service; selecting a set of terminals to be serviced from among the terminals subject to the service; determining a radio frequency (RF) precoding matrix by which interference between the terminals to be serviced is minimized; transferring information on the determined RF precoding matrix to a centralized processor (CP); and generating an effective channel by configuring an RF precoder based on the determined RF precoding matrix.

Abstract: A semiconductor package and a method of manufacturing a semiconductor package. As a non-limiting example, various aspects of this disclosure provide a semiconductor package, and method of manufacturing thereof, that comprises shielding on multiple sides thereof.

Abstract: Disclosed are methods for transmitting and receiving beamforming-based signals in a communication system. An operation method of a terminal may comprise transmitting capability information of the terminal to an AN; receiving first configuration information of a network cluster configured based on the capability information from the AN; transmitting an SRS in a beamforming scheme through resources indicated by the first configuration information; receiving second configuration information of a user cluster configured based on a measurement result of the SRS from the AN; and performing cooperative communication with ANs belonging to the user cluster, which are indicated by the second configuration information. Accordingly, the performance of the communication system can be improved.

Abstract: An operation method of a first AN for a hybrid beamforming-based cooperative transmission in a C-RAN includes obtaining a spatial channel covariance with terminals subject to a service; selecting a set of terminals to be serviced from among the terminals subject to the service; determining a radio frequency (RF) precoding matrix by which interference between the terminals to be serviced is minimized; transferring information on the determined RF precoding matrix to a centralized processor (CP); and generating an effective channel by configuring an RF precoder based on the determined RF precoding matrix.