Abstract: The present invention relates to a context-based QA generation architecture, and an object of the present invention is to generate diverse QA pairs from a single context. To achieve the object, the present invention includes a latent variable generating network including at least one encoder and an artificial neural network (Multi-Layer Perceptron: MLP) and configured to train the artificial neural network using a first context, a first question, and a first answer, and generate a second question latent variable and a second answer latent variable by applying the trained artificial neural network to a second context, an answer generating network configured to generate a second answer by decoding the second answer latent variable, and a question generating network configured to generate a second question based on a second context and the second answer.

Abstract: Aspects of the subject disclosure may include, systems and methods, for example, including receiving a user question data in a speech format or a text format, analyzing the user question data, selecting a plurality of documents from a plurality of domains corresponding to the user question data, searching for a plurality of passages including candidates for an answer value determined as being suitable for the user question data, in the plurality of documents, obtaining candidates by inputting the user question data and the plurality of passages into a plurality of MRC question and answer units, determining the answer value based on whether a reliability value of each of the candidates exceeds a threshold value, and providing the determined answer value to a user. Other embodiments are disclosed.

Abstract: A memory controller is configured with a plurality of processors to be operated in parallel so that overhead of firmware may be reduced. The memory controller includes a first processor and a second processor. The first processor is configured to generate a command corresponding to a request received from a host and to translate a logical address included in the request into a physical address of a memory device. The second processor is configured to operate on data to be output to the memory device or on data received from the memory device. If the request is received from the host when the second processor is in an idle state, the first processor may control the second processor to release the idle state of the second processor and to perform an operation of logging command information corresponding to the request.

Abstract: Disclosed are a rare-earth permanent magnet having improved magnetic properties and a method of manufacturing the same. A method of manufacturing a rare-earth permanent magnet may include: preparing a mixed powder including i) a first alloy represented by R1aR2bBcMdFebal and ii) a second alloy represented by R2bBcMdFebal where R1 is one or two or more of La, Ce, and Y; R2 is a rare-earth element except for La, Ce, and Y; and M is a metal element; press-forming and sintering the prepared mixed powder in a magnetic field to prepare a sintered body; and performing a heat treatment based on diffusion temperature conditions of an R1 component and an R2 component contained in the prepared sintered body.

Abstract: An electronic apparatus includes: a circuit board; a driving chip mounted on the circuit board; a shield-can including a top surface and a side surface extending in a direction from the top surface to the circuit board, wherein the top surface covers the driving chip; and a first film disposed on the shield-can and including a first opening that exposes a part of the top surface of the shield-can.

Abstract: Provided is an artificial intelligence overtopping prediction device including first and second pressure measurement parts 103 and 105 configured to measure a collision pressure of waves colliding with a marine structure 10 and a pressure of waves introduced into a road surface 20, an overtopping alert part 109 configured to collect images around the overtopping prediction device 100 and output an alert sound, an information collection and control part 107 including a drone storage 170 configured to fly the drone 202, and a collection controller 160 configured to control the overtopping prediction device 100, receive and store the wave pressure information and the overtopping amount information, and transmit the information to an overtopping prediction system 500 to predict and generate overtopping prediction information, and a frame part 101 including a support frame 110. Also, disclosed herein is an overtopping prediction system using the artificial intelligence overtopping prediction device 100.

Abstract: A rotary kiln comprises: a sintering device provided with a cylindrical rotation tube for mixing an input powder raw material while rotating in a horizontal direction; an external heating device for heating the powder raw material input into the rotation tube by heating the outside of the rotation tube; and an internal heating device for stirring and heating the powder raw material input into the rotation tube simultaneously. The internal heating device includes a microwave generation part for generating microwaves; a guide part for guiding the microwaves generated from the microwave generation part into the rotation tube; and a stirring heat generation part coupled to an inner circumferential surface of the rotation tube for stirring the powder raw material input into the rotation tube and simultaneously generating heat when absorbing the microwaves so as to heat the powder raw material.

Abstract: Provided is an artificial intelligence (AI) answering system including a user question receiver configured to receive a user question from a user terminal; a first question extender configured to generate a question template by analyzing the user question and determine whether the user question and the generated question template match; a second question extender configured to generate a similar question template by using a natural language processing and a deep learning model when the user question and the generated question template do not match; a training data builder configured to generate training data for training the second question extender by using an neural machine translation (NMT) engine; and a question answering unit configured to transmit a user question result derived through the first question extender or the second question extender to the user terminal.

Abstract: A machine reading comprehension (MRC) question and answer providing method includes receiving a user question; analyzing the user question; selecting at least one document from at least one domain corresponding to an analyzed user question and searching for a passage, which is a candidate answer determined as being suitable for the user question, in the selected at least one document; obtaining at least one correct answer candidate value by inputting the user question and a corresponding passage into each of at least one MRC question and answer unit; and determining whether the at least one correct answer candidate value is a best answer.

Abstract: A method of generating a question-answer learning model through adversarial learning may include: sampling a latent variable based on constraints in an input passage; generating an answer based on the latent variable; generating a question based on the answer; and machine-learning the question-answer learning model using a dataset of the generated question and answer, wherein the constraints are controlled so that the latent variable is present in a data manifold while increasing a loss of the question-answer learning model.

Abstract: A mark printing device includes a driving unit driving along a driving rail of an overhead hoist transport, a printing unit being configured to print on the driving rail a mark for guiding a motion of a vehicle, a data processing unit receiving design data including design information of the driving rail and first information on a position and type of the mark from a server, an encoder unit being configured to calculate a rotation amount of a servo motor provided in the driving unit to detect a current position of the driving unit and a driving distance of the driving unit, and a control unit being configured to control the driving unit and the printing unit to print the mark on the driving rail by using the design data and second information on the current position and the driving distance of the driving unit.

Abstract: A display device includes: a substrate including a plurality of pixels; and a light-emitting element layer on the substrate.

Abstract: The invention relates to a method and a system for improving performance of text summarization and has an object of improving performance of a technique for generating a summary from a given paragraph. According to the invention to achieve the object, a method for improving performance of text summarization includes: an a step of generating an embedding vector by vectorizing a natural language-based context; a b step of generating a graph using the embedding vector and calculating a first likelihood of each of at least one node included in the graph; a c step of generating a second likelihood by assigning a weight to the first likelihood according to a result of comparing at least one node included in the graph with the context; and a d step of calculating a third likelihood for all candidate paths present in the graph based on the second likelihood, selecting a path having a highest third likelihood, and generating a summary based on the path.

Abstract: A CNT X-ray tube control system according to an embodiment of the present disclosure includes a CNT X-ray tube including an anode electrode, a gate electrode, and a cathode electrode, and configured to generate X-rays from electric field emission; a dummy load including an anode electrode, a gate electrode, and a cathode electrode, and configured not to generate X-rays, the dummy load being connected to the CNT X-ray tube in parallel; a high-voltage part configured to control a voltage applied to the CNT X-ray tube; and a controller configured to control the CNT X-ray tube, the dummy load, and the high-voltage part. Particularly, a portable X-ray generation apparatus reduced in size can be provided, the apparatus including the dummy load having the same structure as a CNT to minimize the voltage variation of the high-voltage part during the ON/OFF operation of the CNT.

Abstract: An electronic apparatus includes: a circuit board; a driving chip mounted on the circuit board; a shield-can including a top surface and a side surface extending in a direction from the top surface to the circuit board, wherein the top surface covers the driving chip; and a first film disposed on the shield-can and including a first opening that exposes a part of the top surface of the shield-can.

Abstract: A touch sensor may include a substrate and may include electrode units, first demultiplexers, second demultiplexers, and driving pads all located on the substrate. The electrode units each may include a plurality of electrode groups, the electrode groups each including a plurality of touch electrodes. The first demultiplexers each may include a plurality of sub-demultiplexers and each may be electrically connected to a corresponding one of the electrode units. Each of the sub-demultiplexers of a first demultiplexer may be electrically connected to a corresponding one of the electrode groups of a corresponding electrode unit. The second demultiplexers may be connected between the first demultiplexers and the driving pads.

Abstract: An apparatus for removing a bipolar element for a display module includes: a remover that removes the bipolar element; and a solvent storage that supplies solvent to the remover, wherein the remover includes: a partition wall defining a solvent channel which the solvent is injected into and discharged from; and a vibrator disposed inside the solvent channel.

Abstract: The invention relates to a method and a system for improving performance of text summarization and has an object of improving performance of a technique for generating a summary from a given paragraph. According to the invention to achieve the object, a method for improving performance of text summarization includes: an a step of generating an embedding vector by vectorizing a natural language-based context; a b step of generating a graph by using the embedding vector; a c step of assigning a weight depending on whether or not a keyword corresponding to at least one node included in the graph is present in the context; and a d step of selecting a path having a highest likelihood in the graph and generating a summary based on the path.

Abstract: Disclosed herein is a search method performed by a server, including: receiving a user question from a user terminal; generating a user question vector for the user question; selecting similar question candidates based on a similarity to the user question vector; generating an answer to the user question based on the similar question candidates; and transmitting the answer to the user question to the user terminal.

Abstract: A method of generating a question-answer learning model through adversarial learning may include: sampling a latent variable based on constraints in an input passage; generating an answer based on the latent variable; generating a question based on the answer; and machine-learning the question-answer learning model using a dataset of the generated question and answer, wherein the constraints are controlled so that the latent variable is present in a data manifold while increasing a loss of the question-answer learning model.