Abstract: The present invention relates to an apparatus and method for automatically generating machine reading comprehension training data, and more particularly, to an apparatus and method for automatically generating and managing machine reading comprehension training data based on text semantic analysis. The apparatus for automatically generating machine reading comprehension training data according to the present invention includes a domain selection text collection unit configured to collect pieces of text data according to domains and subjects, a paragraph selection unit configured to select a paragraph using the pieces of collected text data and determine whether questions and correct answers are generatable, and a question and correct answer generation unit configured to generate questions and correct answers from the selected paragraph.

Abstract: Disclosed herein are an apparatus and method for retrieving on-chain data based on an index. The apparatus includes memory in which at least one program is recorded and a processor for executing the program. The program may include a blockchain connection manager for recording data in at least one node or retrieving data from at least one node and outputting the same by being connected to a blockchain network formed of multiple nodes, a blockchain index manager for creating and managing an index for data stored in at least one node, and a query engine for requesting storage of data resulting from a transaction in a corresponding node through the blockchain connection manager, creating an index on the stored data by interworking with the blockchain index manager, and acquiring and outputting data that is retrieved based on the index through the blockchain connection manger in response to a retrieval request.

Abstract: A motor-driving apparatus for driving a motor having a plurality of windings respectively corresponding to a plurality of phases is provided. The motor-driving apparatus includes a first inverter having a plurality of first switching devices and connected to first ends of the plurality of windings and a second inverter having a plurality of second switching devices and connected to second ends of the plurality of windings. A third switching device is configured to selectively connect and disconnect points at which a number of turns of each of the windings is divided in a preset ratio. A controller is configured to adjust an on/off state of the first to third switching devices based on required output of the motor.

Abstract: Provided is a method for predicting perovskite synthesizability using a graph convolutional neural network and positive unlabeled learning, capable of predicting perovskite synthesizability by using a graph convolutional neutral network and positive unlabeled learning which is semi-supervised learning based on a labeled model using positive data and positive unlabeled data.

Abstract: An optical film for a display device, includes: a first refractive layer having an upper surface and a lower surface including first projections and second projections extending away from the lower surface in a first direction, the second projections having different heights than the first projections, the first projections having lateral sides with different angles of inclination that decrease in the first direction; and a second refractive layer disposed directly on the upper surface of the first refractive layer, the second refractive layer having a refractive index different from that of the first refractive layer.

Abstract: The present invention relates to a novel electrolyte and a secondary battery including the same. The present invention has an effect of providing a secondary battery having improved charging efficiency and output due to reduced charging resistance and having excellent long-term lifespan and high-temperature capacity retention rate.

Abstract: The present invention relates to an electrolyte solution and a secondary battery including the same. According to the present invention, the present invention has an effect of providing a secondary battery having improved charging efficiency and output due to low discharge resistance and having a long lifespan and excellent high-temperature capacity retention by suppressing gas generation and increase in thickness.

Abstract: The present invention relates to an electrolyte solution and a secondary battery including the same. According to the present invention, the present invention has an effect of providing a secondary battery having improved charging efficiency and output due to low discharge resistance and having a long lifespan and excellent high-temperature capacity retention by suppressing gas generation and increase in thickness.

Abstract: The present invention relates to an electrolyte solution and a secondary battery including the same. According to the present invention, the present invention has an effect of providing a secondary battery having improved charging efficiency and output due to low discharge resistance and having a long lifespan and excellent high-temperature capacity retention by suppressing gas generation and increase in thickness.

Abstract: The present invention relates to an electrolyte solution and a secondary battery including the same. According to the present invention, the present invention has an effect of providing a secondary battery having improved charging efficiency and output due to low discharge resistance and having a long lifespan and excellent high-temperature capacity retention by suppressing gas generation and increase in thickness.

Abstract: Disclosed are novel compounds of Chemical Formula 1, optical isomers of the compounds, and pharmaceutically acceptable salts of the compounds or the optical isomers. The compounds, isomers, and salts exhibit excellent activity as GLP-1 receptor agonists. In particular, they, as GLP-1 receptor agonists, exhibit excellent glucose tolerance, thus having a great potential to be used as therapeutic agents for metabolic diseases. Moreover, they exhibit excellent pharmacological safety for cardiovascular systems.

Abstract: The present specification relates to a conductive structure body, a method for manufacturing the same, and an electrode and an electronic device including the conductive structure body.

Abstract: Disclosed are a spiral curve type music sheet in which different notes are displayed at different positions on a spiral curve based on the pitches of notes, and an apparatus and method for providing a spiral curve type music sheet. The apparatus for providing a spiral curve type music sheet may include a memory configured to store a spiral curve type music sheet in which different notes are displayed at different positions on a spiral curve based on the pitch of the note and note data, and a processor configured to determine the note symbol position related to the note data on the spiral curve in the spiral curve type music sheet based on the frequency of the note data.

Abstract: The present invention relates to a method of accumulating and decoding a PBCH signal received by a 5G NR receiver, the method including: generating an inversion vector for negating at least one bit of a system frame number (SFN) included the received PBCH signal; performing accumulation over at least one frame by performing modulo addition on the generated inversion vector and the received PBCH signal; decoding the accumulated PBCH signals; and checking validity of the decode PBCH and reconstructing the SFN.

Abstract: The present invention relates to a CSI extraction method in a MIMO receiver used in a wireless communication system, the method including: obtaining an effective channel matrix by matrix multiplication of a precoding matrix and a channel estimation value obtained through a CSI-RS; calculating an upper and a lower bound of a minimum distance for each layer through QR decomposition for the effective channel matrix; and mapping the upper and the lower bound of the minimum distance for each layer to each codeword, and extracting a mean mutual information per bit (MMIB) metric that is a transmission capacity per unit frequency for each codeword. According to the present invention, QR decomposition and a MMIB metric are used to obtain the minimum distance for each layer without a multidimensional search process, whereby the CSI is extracted with fewer operations than the conventional method when the MIMO transmission order is high.

Abstract: The present invention relate to a method of extracting an optimum rank without considering a precoding matrix indicator (PMI) in a multiple-input multiple-output (MIMO) receiver, the method including calculating a correlation matrix from a MIMO channel matrix between a base station antenna and an MIMO receiver antenna; obtaining an eigenvalue from the correlation matrix; calculating channel capacity for each rank on the basis of the obtained eigenvalue; and selecting a rank corresponding to a channel capacity having a maximum value among the calculated channel capacities for each rank as an optimum rank.

Abstract: The present invention relates to a CSI extraction method in a MIMO receiver used in a wireless communication system, the method including: obtaining an effective channel matrix by matrix multiplication of a precoding matrix and a channel estimation value obtained through a CSI-RS; calculating an upper and a lower bound of a minimum distance for each layer through QR decomposition for the effective channel matrix; and mapping the upper and the lower bound of the minimum distance for each layer to each codeword, and extracting a mean mutual information per bit (MMIB) metric that is a transmission capacity per unit frequency for each codeword. According to the present invention, QR decomposition and a MMIB metric are used to obtain the minimum distance for each layer without a multidimensional search process, whereby the CSI is extracted with fewer operations than the conventional method when the MIMO transmission order is high.

Abstract: The present invention relates to a method of accumulating and decoding a PBCH signal received by a 5G NR receiver, the method including: generating an inversion vector for negating at least one bit of a system frame number (SFN) included the received PBCH signal; performing accumulation over at least one frame by performing modulo addition on the generated inversion vector and the received PBCH signal; decoding the accumulated PBCH signals; and checking validity of the decode PBCH and reconstructing the SFN.

Abstract: The present invention relate to a method of extracting an optimum rank without considering a precoding matrix indicator (PMI) in a multiple-input multiple-output (MIMO) receiver, the method including calculating a correlation matrix from a MIMO channel matrix between a base station antenna and an MIMO receiver antenna; obtaining an eigenvalue from the correlation matrix; calculating channel capacity for each rank on the basis of the obtained eigenvalue; and selecting a rank corresponding to a channel capacity having a maximum value among the calculated channel capacities for each rank as an optimum rank.

Abstract: An interference canceller (500) includes an interference cancellation coefficient calculator (510) configured to calculate an interference cancellation coefficient using a received value located in a location of a reference signal (RS) from a channel estimate H calculated by a channel estimator and a received signal r delivered from an FFT, an interference cancellation filter (520) configured to cancel interference in the received signal and the channel estimate using the calculated interference cancellation coefficient, an increment or decrement operator (530) configured to increase or decrease the received signal and the channel estimate by ? times; an interference cancellation output controller (540) configured to receive a channel estimate before interference cancellation and a channel estimate after interference cancellation to compare a signal before the interference cancellation with a signal after the interference cancellation and then determine whether to deliver the signal before applying the interf