Abstract: A quantization method of a neural network, and an apparatus for performing the quantization method are provided. The quantization method includes obtaining parameters of the neural network, quantizing the parameters using a quantization scheme in which at least one positive quantization level and at least one negative quantization level symmetric to each other by excluding zero from quantization levels, and outputting the quantized parameters.

Abstract: A method of removing harmful substances in blood according to an embodiment of the present disclosure is performed by a fluidic device including an inlet into which isolated blood is injected and at least one blood-clot generating and fixing unit, and the method includes injecting isolated blood into the inlet and removing harmful substances in blood by generating blood clots in the at least one blood-clot generating and fixing unit from the injected blood and fixing the blood clots.

Abstract: The present invention relates to an electrode structure, a method of manufacturing the same, and a secondary battery including the same, and the electrode structure may include a negative electrode part; a positive current collector which is formed of a fabric material and surrounds an outer surface of the negative electrode part; and a positive electrode coupled to an edge of the positive current collector.

Abstract: A method for designing a magnetic plate allowing magnetic particles contained in magnetic ink distributed at different densities includes estimating a distribution of the magnetic particles allowing the magnetic ink to be spread, estimating forms of magnetic field applying the magnetic force to the magnetic particles in order for the magnetic ink to be spread in a desired magnetic printing pattern within the printing layer, obtaining adjustment factors of the magnetic plate corresponding to the estimated forms of the magnetic field by using a model for which deep learning or machine learning has been performed using design conditions including at least one of an upper surface structure of the magnetic plate and a magnetization property of the magnetic plate and obtaining a parameter for the magnetic plate configured to form the magnetic printing pattern based on the obtained adjustment factors.

Abstract: The present invention relates to a reaction method for forming a layered structure of immiscible liquid-phase concentric layers within a rotary reactor and a reaction system including the layered structure, and may provide a basis capable of efficiently performing a multistage reaction in terms of time and space.

Abstract: An image deblurring method and an apparatus performing the same are disclosed. The image deblurring method according to an example embodiment includes receiving a blurred image, and deblurring of outputting a sharp original image based on the blurred image. The deblurring includes obtaining a second image having the same scale as that of a first image by inputting the first image to a neural network, obtaining a third image having the same scale as that of the first image by concatenating the first image and the second image, and obtaining a fourth image having the same scale as that of the first image by inputting the third image to the neural network.

Abstract: An evaluation method capable of evaluating whether or not waste plastic-derived porous carbon can be applied on an industrial scale, according to the present disclosure, may include the steps of evaluating CO2 capture performance using a 5-step temperature vacuum swing adsorption (TVSA) process, assessing economic feasibility in an industry using a techno-economic assessment (TEA) method, and quantifying environmental impacts of the porous carbon production pathway and global warming potential (GWP) using cradle-to-gate life-cycle assessment (LCA). A method for manufacturing porous carbon, according to the present disclosure, may include the steps of carbonizing a polyethylene terephthalate plastic, activating the carbonized plastic with CO2, and performing cooling.

Abstract: A separator for an electrochemical device provided with a porous coating layer including multiple types of binder resins, and a method for manufacturing the same. The separator shows high adhesion between an electrode and the separator, even when any separate adhesive layer is not disposed on the surface of the separator. In addition, the separator shows higher adhesion to an electrode, as compared to a separator using a fluorinated binder resin, such as polyvinylidene fluoride, used conventionally in the art. Therefore, when introducing the separator to a battery, it is possible to manufacture an electrode assembly under a mild temperature and pressure condition, to improve the productivity of assemblage, to reduce defect generation, and thus to increase the yield. Further, the separator shows high affinity to an electrolyte, as compared to a semi-crystalline polymer, such as a fluorinated binder resin, and thus improves the output characteristics of a battery.

Abstract: Provided are a non-volatile memory device and a method of operating the same. The non-volatile memory device includes a substrate, a plurality of word lines extending in a first direction on the substrate, a plurality of ferroelectric patterns respectively provided on the word lines, a blocking insulating film covering the ferroelectric patterns, a plurality of bit line pairs including a first bit line and a second bit line extending in a second direction crossing the word lines and the ferroelectric patterns on the blocking insulating film and intersecting the first direction, and a channel pattern provided between the first bit line and the second bit line of each of the bit line pairs on the blocking insulating film, wherein the channel pattern has an ambipolar conduction characteristic.

Abstract: Disclosed are an organic thermoelectric material and a thermoelectric generator including the same. More particularly, the thermoelectric generator includes an ionically conductive active layer containing a polyanion including an anionic group and a counter cation in a repeat unit thereof; a conductive polymer; and a polyvalent crosslinking agent as a single molecule including a plurality of acid functional groups. First and second electrodes are disposed to be connected to the ionically conductive active layer.

Abstract: Disclosed is a method of manufacturing a solid electrolyte using water as a solvent. The method includes dissolving a precursor in water to form a slurry, drying the slurry to form granules, pressing the granules to form a pressed solid body, and sintering the pressed solid body to manufacture a solid electrolyte.

Abstract: The present invention relates to a real-time parathyroid imaging apparatus including: a light source including an excitation filter capable of exciting parathyroid glands; and a camera including an image sensor and an emission filter of which a transmissivity ratio between a visible light region and a near-infrared emission wavelength region is N:1 (here, N<1). Through the present invention, a system may be implemented whereby a surgeon may acquire, in real-time during an operation, an autofluorescence image of the parathyroid glands by using a near-infrared light source, and an auto focus function may be used, and visible light and near-infrared autofluorescence images may be simultaneously fused and acquired without having to turn off the lights in an operating room.

Abstract: A boron nitride layer and a method of fabricating the same are provided. The boron nitride layer includes a boron nitride compound and has a dielectric constant of about 2.5 or less at an operating frequency of 100 kHz.

Abstract: Provided is a microfluidic film including a first microfluidic film including a first base film, a first microchannel, which is formed on the first base film and through which a fluid flows, and a first through passage, which is configured to pass through the first base film, and a second microfluidic film including a second base film being stacked on the first base film and a second through passage, which is configured to pass through the second base film and communicates with the first through passage.

Abstract: Provided is an apparatus for generating a microfluidic concentration field, the apparatus including: a substrate; a base film disposed on the substrate; a microchannel, which is formed in a space between the substrate and the base film and through which a fluid flows; a through passage, which communicates with the microchannel and is configured to pass through the base film; and a membrane, which is formed at a portion where the microchannel and the through passage communicate with each other and allows the fluid flowing along the microchannel and the through passage or a material flowing together with the fluid to selectively pass through the membrane, wherein a concentration field is formed between the fluid of the through passage and the fluid of the microchannel by the membrane.

Abstract: Provided is a microfluidic film including a base film, a microchannel, which is formed on the base film and through which a fluid flows, and a through passage, which is configured to pass through the base film and through which the base film stacked on an upper portion or a lower portion of the base film and the fluid communicate with each other.

Abstract: An amorphous silicon-carbon composite, a method for preparing the amorphous silicon-carbon composite using a pyrolysis method, a negative electrode for a lithium secondary battery, and a lithium secondary battery including the same.

Abstract: Disclosed are an implantable sensor driven by an alignment key, an implantable device comprising the implantable sensor, and a biometric data measurement system comprising the implantable device. The implantable device according to the present embodiment may comprise an implantable sensor forming a magnetic dipole moment in one direction from the inside to the outside of the body, and may be inserted into the body to measure biometric data by means of the implantable sensor.

Abstract: A processor-implemented method includes determining a first quantization value by performing log quantization on a parameter from one of input activation values and weight values in a layer of a neural network, comparing a threshold value with an error between a first dequantization value obtained by dequantization of the first quantization value and the parameter, determining a second quantization value by performing log quantization on the error in response to the error being greater than the threshold value as a result of the comparing; and quantizing the parameter to a value in which the first quantization value and the second quantization value are grouped.