Abstract: A method of manufacturing an aerosol generating rod includes providing a sheet of a first non-tobacco material by using at least one first conveying roller; crimping the sheet of the first non-tobacco material by using a first crimping device; applying an aerosol generating material to at least one surface of the sheet of the first non-tobacco material by using a first spray nozzle; applying a liquid containing a tobacco component to at least one surface of the sheet of the first non-tobacco material by using a slit nozzle; drying the sheet of the first non-tobacco material by using a drying device; and forming the aerosol generating rod by forming the sheet of the first non-tobacco material into a rod by using a rod forming device.

Abstract: Provided are a smoking article having improved flavor persistence and a manufacturing method thereof, the smoking article including: a smoking material portion: a filter portion: and a wrapper wrapping at least a portion of the smoking material portion, wherein flavor persistence of the smoking article is improved by applying a flavoring sheet including a hydrocolloid material and flavoring to the wrapper.

Abstract: A flavoring sheet with an enhanced flavor holding amount and flavor retention and a smoking article including the same are provided. The flavoring sheet according to some embodiments of the present disclosure may include a hydrocolloid material configured to form a sheet, an emulsifier, and a fat-soluble flavoring. The emulsifier may serve as a cross-link between the water-soluble hydrocolloid material and the fat-soluble flavoring and increase a flavor holding amount and flavor retention of the flavoring sheet.

Abstract: A method and apparatus for improving the quality of an attention-based sequence-to-sequence model. The method includes determining an output sequence corresponding to an input sequence based on an attention-based sequence-to-sequence model, selecting at least one target attention head from among a plurality of attention heads, detecting at least one error output token among output tokens constituting the output sequence based on the target attention head, and correcting the output sequence based on the error output token.

Abstract: A flavoring sheet with improved physical properties, a smoking article including the same, and methods of producing the flavoring sheet and the smoking article are provided. The flavoring sheet according to some embodiments of the present disclosure may include a hydrocolloid material configured to form a sheet, a flavoring, and a plasticizer. The plasticizer can improve physical properties of the flavoring sheet and thus enhance workability of a process of cutting the flavoring sheet.

Abstract: An automated gas supply system includes a gas cylinder transfer unit configured to transfer a cradle in which one or more gas cylinders storing a gas therein are stored; a gas cylinder inspection unit configured to check properties of the gas stored in the gas cylinder transferred from the gas cylinder transfer unit and check whether the gas leaks from the gas cylinder; a storage queue configured to receive the gas cylinder from the gas cylinder inspection unit by a mobile robot and configured to classify and store the transferred gas cylinders according to the properties of the gas stored in the gas cylinder; and a gas cabinet configured to receive the gas cylinder from the storage queue by the mobile robot and fasten a gas pipe, which is connected to a semiconductor manufacturing process line, to a gas spray nozzle, which is disposed at one side of the received gas cylinder, to supply the gas stored in the gas cylinder to the semiconductor manufacturing process line, wherein the gas cabinet includes a resid

Abstract: A real-time processor-implemented translation method and apparatus is provided. The real-time translation method includes receiving a content, determining a delay time for real-time translation based on a silence interval of the received content and an utterance interval of the received content, generating a translation result by translating a language used in the received content, and synthesizing the translation result and the received content.

Abstract: Disclosed are a separator for fuel cells capable of minimizing the volume of a system and the use of sealants, and a stack for fuel cells, more particularly, a stack for solid oxide fuel cells, including the same. Specifically, by adding a metal sheet having a specific shape, position and size to the separator, the stress applied to the sealant can be uniformized, and thus the oxidizing agent and fuel can be separated and electrically isolated using only a piece of sealant. Therefore, the stack for fuel cells is characterized in that there is no variation in temperature, reactant concentration, power, or the like between respective unit cells, so delamination and microcracks do not occur, the volume is minimized, and the power density per unit volume is very high.

Abstract: A decoding method and apparatus in an artificial neural network for speech recognition. The decoding method in the artificial neural network for speech recognition includes performing a first decoding task of decoding a feature including speech information and at least one token recognized up to current time, using a shared decoding layer included in the artificial neural network, performing a second decoding task of decoding the at least one token, using the shared decoding layer, and determining an output token to be recognized subsequent to the at least one token based on a result of the first decoding task and a result of the second decoding task.

Abstract: A decoding method and apparatus in an artificial neural network for speech recognition. The decoding method in the artificial neural network for speech recognition includes performing a first decoding task of decoding a feature including speech information and at least one token recognized up to current time, using a shared decoding layer included in the artificial neural network, performing a second decoding task of decoding the at least one token, using the shared decoding layer, and determining an output token to be recognized subsequent to the at least one token based on a result of the first decoding task and a result of the second decoding task.

Abstract: A speech recognition method includes adding a preset special sequence to a front end of an input sequence that corresponds to an input utterance of a speaker, recognizing the preset special sequence and the input sequence, and recognizing the input sequence based on the preset special sequence and a speech recognition result obtained by recognizing the preset special sequence and the input sequence.

Abstract: A method for determining an output token includes predicting a first probability of each of candidate output tokens of a first model, predicting a second probability of each of the candidate output tokens of a second model interworking with the first model, adjusting the second probability of each of the candidate output tokens based on the first probability, and determining the output token among the candidate output tokens based on the first probability and the adjusted second probability.

Abstract: A speech recognition method includes adding a preset special sequence to a front end of an input sequence that corresponds to an input utterance of a speaker, recognizing the preset special sequence and the input sequence, and recognizing the input sequence based on the preset special sequence and a speech recognition result obtained by recognizing the preset special sequence and the input sequence.

Abstract: A decoding method, the method including: receiving an input sequence corresponding to an input speech at a current time; and in a neural network (NN) for speech recognition, generating an encoded vector sequence by encoding the input sequence, determining reuse tokens from candidate beams of two or more previous times by comparing the candidate beams of the previous times, and decoding one or more tokens subsequent to the reuse tokens based on the reuse tokens and the encoded vector sequence.

Abstract: A processor-implemented decoding method in a first neural network is provided. The method predicts probabilities of candidates of an output token based on at least one previously input token, determines the output token among the candidates based on the predicted probabilities; and determines a next input token by selecting one of the output token and a pre-defined special token based on a determined probability of the output token.

Abstract: A processor implemented natural language processing method and apparatus are provided. The natural language processing method includes converting a natural language phrase into a token vector, calculating a repetition count of the token vector, and generating an input vector by encoding the token vector based on the calculated repetition count and a position of the token vector.

Abstract: A method and apparatus for improving the quality of an attention-based sequence-to-sequence model. The method includes determining an output sequence corresponding to an input sequence based on an attention-based sequence-to-sequence model, selecting at least one target attention head from among a plurality of attention heads, detecting at least one error output token among output tokens constituting the output sequence based on the target attention head, and correcting the output sequence based on the error output token.

Abstract: A sequence processing method and apparatus are provided. The sequence processing method includes determining a word of a first R-node corresponding to a root node based on an input sequence, generating first I-nodes that are connected to the first R-node and include relative position information with respect to the word of the first R-node, determining a word of a second R-node to correspond to each of the first I-nodes, and determining an output sequence corresponding to the input sequence based on the determined words.

Abstract: A processor-implemented utterance time estimation method includes: determining a plurality of attention weight matrices using an attention-based sequence-to-sequence model; selecting an attention weight matrix from the plurality of attention weight matrices; and estimating an utterance time corresponding to an output sequence based on the selected attention weight matrix.

Abstract: A processor-implemented classification method includes: determining a first probability vector including a first probability, for each of a plurality of classes, resulting from a classification of an input with respect to the classes; determining, based on the determined first probability vector, whether one or more of the classes represented in the first probability vector are confusing classes; adjusting, in response to one or more of the classes being the confusing classes, the determined first probability vector based on a first probability of each of the confusing classes and a maximum value of the first probabilities; determining a second probability vector including a second probability, for each of the classes, resulting from another classification of the input with respect to the classes; and performing classification on the input based on a result of a comparison between the determined second probability vector and the adjusted first probability vector.