Abstract: A method of performing distributed matrix computation using task entanglement-based coding as a method of processing a huge amount of matrix computation in a distributed manner in a distributed computing environment is provided. A main server encodes information to be transmitted to a plurality of edge devices for distributed matrix computation on the basis of task entanglement-based coding employing a Chebyshev polynomial, thereby reducing the amount of information to be transmitted. Also, when the number of computation results received from the edge devices becomes a recovery threshold, the main server immediately performs decoding to derive a matrix computation result.

Abstract: A method and apparatus for determining sensory threshold information for a shoe-type device, the shoe-type device, and a method of controlling the shoe-type device are disclosed. The method of determining sensory threshold information includes executing a program for sensory threshold testing, determining sensory threshold information of a user based on a choice input by the user that is received in a process of the sensory threshold testing, and transmitting the determined sensory threshold information to the shoe-type device.

Abstract: A receiver includes a sampling circuit configured to sample a comparison result between an input signal and a plurality of threshold voltages according to a sampling clock signal; a clock controller configured to generate the sampling clock signal according to a clock control signal; and a control circuit configure to generate the clock control signal and the plurality of threshold voltages according to a target value and an output of the sampling circuit. The control circuit operates to control a ratio of a magnitude of a main cursor of the input signal and a magnitude of a pre-cursor intersymbol interference to be the target value.

Abstract: A method of operating a deep learning system configured to perform private inferences, including performing a convolution operation with respect to input values; and outputting result values from the convolution operation using an activation function, wherein the activation function includes a Hermitic expansion using a Hermite polynomial as an eigenfunction to perform a Fourier transform.

Abstract: There is provided a clock and data recovery circuit for a high-speed PAM-4 receiver through statistical learning. A clock and data recovery device according to an embodiment includes: an input unit through which data is inputted; a clock input unit through which a clock is inputted; a sampling unit configured to sample the inputted data by using the inputted clock; a controller configured to combine results of sampling at a plurality of sampling points, to determine a state of the clock based on the combined results, and to generate a control value for controlling the clock; and an adjustment unit configured to adjust the clock applied to the sampling unit, based on the control value generated by the controller. Accordingly, a hardware structure is simplified and energy efficiency is enhanced compared to an exiting oversampling clock and data recovery circuit for a PAM-4 receiver.

Abstract: A fingerprint extraction apparatus includes a fingerprint generation module configured to generate first and second virtual fingerprint images, perform primary image processing on the first virtual fingerprint image and primary and secondary image processing on the second virtual fingerprint images, and generate virtual overlapped fingerprint images by combining the first and second virtual fingerprint images on which image processing is performed; a machine learning module configured to generate a learning model by performing machine learning using the virtual overlapped fingerprint images as input data; and a fingerprint extraction module configured to extract a fingerprint located vertically on a center of a real image by inputting the real image to a target fingerprint extraction learning model, and the primary image processing comprises image processing on a curve forming a fingerprint, and the secondary image processing comprises image processing on a location of the fingerprint in the image.

Abstract: A photoresist composition, a method of forming a pattern, and a method of synthesizing 6-nitrochrysene, the photoresist composition includes a polymer resin; a photo acid generator; a quencher; an organic solvent; and an etching resistance enhancer, wherein the etching resistance enhancer is represented by the following Chemical Formula 1,

Abstract: The present invention relates to a peptide targeting a toxin-antitoxin system of Mycobacterium tuberculosis and a use thereof. Specifically, the antibiotic peptide of the present invention inhibits the formation of a toxin-antitoxin complex of Mycobacterium tuberculosis without affecting an active site of the toxin, thereby inducing the death of Mycobacterium tuberculosis by means of a separated toxin. Therefore, the antibiotic peptide can be usefully used as an antibiotic composition against Mycobacterium tuberculosis.

Abstract: A processor-implemented accelerator method includes: reading, from a memory, an instruction to be executed in an accelerator; reading, from the memory, input data based on the instruction; and performing, on the input data and a parameter value included in the instruction, an inference task corresponding to the instruction.

Abstract: The present invention relates to a method for predicting physical properties of an amorphous porous material which may predict an acoustic physical property value and an absorption coefficient from parameters of an amorphous porous material, and may estimate the acoustic characteristics with the amorphous porous material which is an amorphous specimen even without separately producing a formalized specimen such as a cylindrical specimen or a flat specimen. Further, the method for predicting physical properties of the amorphous porous material may estimate the acoustic physical properties through the analysis of a three-dimensional pore connection structure, which is a microstructure of an amorphous specimen, even without acoustic impedance, thereby estimating the acoustic physical properties which accurately reflect the characteristics of the actual specimen.

Abstract: A receiver includes an equalization circuit configured to output a data sample signal and an edge sample signal by sampling a data input signal according to clock signal, and to perform an equalization operation according to the data sample signal and the edge sample signal; and a clock gate circuit configured to select the clock signals from among a plurality of multi-phase clock signals according to a selection signal.

Abstract: Strain sensors are attached to a workpiece, and resistance values of the strain sensors are measured with time while cutting the workpiece with a milling cutting machine. A power spectrum in a frequency domain is obtained by Fourier-transforming the resistance values measured with time. A power peak is found in the power spectrum to be outputted as a magnitude of vibration occurring during machining the workpiece. By attaching a strain sensor to a stamper of a plastic working machine, a resistance value of the strain sensor is measured while pressing a workpiece on a die with the stamper. A magnitude of the strain corresponding to the measured resistance value is used to obtain a corresponding plastic working force. By attaching the strain sensor directly to the workpiece, it is possible to monitor the state of the plastic working and/or cutting process of the workpiece with high precision at low cost.

Abstract: Disclosed are modularized garment manufacturing method and system for small quantity batch production of garments. Basic components constituting an ordered garment are generated by analyzing shape graphic data of the ordered garment. For each basic component, basic patterns are generated for each dimension of the ordered garment. The basic patterns are compared with reference garment patterns stored in a garment pattern database to determine a pattern category of each basic pattern. Based on the pattern category of the basic patterns and information on garment production lines currently in operation, subdivided work modules are determined for each basic pattern.

Abstract: A drug injection control device generates a policy model by learning a change in anesthetic state information due to a drug injection rate set so that anesthetic state information of a patient follows target anesthetic state information, generates a prediction model by learning a change in the anesthetic state information according to a change in the drug injection rate, sets the drug injection rate from the anesthetic state information based on the policy model, and predicts expected anesthetic state information from the set drug injection rate and a previously set drug injection rate based on the prediction model.

Abstract: A semiconductor device includes a first switch coupling a first switch coupling a first power source and a first node according to a first control signal; a sense amplifier coupled between the first node and a second node and performing a sensing operation; a second switch coupling a second power source and the second node according to a second control signal; and a sense amplifier control circuit providing the first control signal and the second control signal. The sense amplifier control circuit controls the second control signal so that a voltage of the second node reaches a shift voltage higher than a voltage of the second power source during a first sensing period of the sensing operation and a bias current flows through the second node during a second sensing period of the sensing operation. The sensing period is subsequent to the first sensing period.

Abstract: A processor-implemented neural network method includes: generating a bit vector based on whether each of a plurality of input activations within a neural network is 0; merging the bit vector into the input activations such that bit values within the neural network included in the bit vector are most significant bits (MSBs) of multi bit expressions of the input activations; merging the bit vector into weights such that the bit values included in the bit vector are MSBs of multi bit expressions of the weights; sorting the input activations and the weights based on bits corresponding to the MSBs; and implementing the neural network, including performing operations between the sorted input activations and the sorted weights.

Abstract: Embodiments provide a light emitting device that includes a first electrode, a second electrode facing the first electrode, and an emission layer disposed between the first electrode and the second electrode, wherein the emission layer includes a first compound represented by Formula 1, which is explained in the specification:

Abstract: A receiver includes a plurality of linear equalizers receiving an input signal; and a plurality of samplers configured to sample a plurality of equalization signals output from the plurality of linear equalizers according to a clock signal. Each of the plurality of linear equalizers compares the input signal with a reference voltage among a plurality of reference voltages to determine a level of the input signal.

Abstract: A wind tunnel test equipment using wind force integration includes: a structural model whose interior is empty; an internal support disposed inside the structural model; and load sensors disposed outside the internal support and coming into close contact with the inner surfaces of the structural model to measure aerodynamic forces applied to the structural model.

Abstract: The present invention relates to a cathode material for a lithium-air battery and a method of manufacturing a cathode using the same. The cathode material of the present invention includes a solvent component and thus includes an electrolyte in a small amount compared to a conventional cathode material, thereby reducing the weight of a cathode manufactured using the cathode material, ultimately increasing the energy density of a lithium-air battery including the cathode.