Abstract:
A method and a user terminal for performing a call using voice recognition are disclosed. The method and the user terminal include receiving a call request, obtaining a voice signal from a user in response to the call request, and determining whether the obtained voice signal is a call reception command to receive the call. The method and the user terminal perform the call in a handsfree mode based on the call request in response to the voice signal being the call reception command.
Abstract:
A semiconductor package includes a wiring substrate, a solder ball on a lower surface of the wiring substrate, a ball land between the lower surface of the wiring substrate and the solder ball and having an upper surface having a circular shape, and a mask layer covering the lower surface of the wiring substrate and including an opening through which a portion of the ball land is exposed. The ball land includes a first land region which is exposed via the opening and has an upper surface having a semicircular shape with a first radius and a second land region which is integrated with a flat side surface of the first land region and has an upper surface having a semicircular shape with a second radius that is less than the first radius.
Abstract:
A semiconductor package includes a first structure that includes an upper connection pattern; a second structure that includes a lower connection pattern; and a connection structure that connects the upper connection pattern of the first structure to the lower connection pattern of the second structure, wherein the connection structure includes a lower conductor connected to the upper connection pattern of the first structure; an upper conductor connected to the lower conductor and the lower connection pattern of the second structure; and a dielectric pattern that at least partially surrounds the upper conductor, and the dielectric pattern includes a first surface in contact with the upper conductor; and a second surface in contact with the lower conductor.
Abstract:
Provided are a silicon precursor having a heterocyclic group, a composition for depositing a silicon-containing layer including the same, and a method of depositing a silicon-containing layer using the same. The silicon precursor is represented by Formula 1.
In Formula 1, A1 is a heterocyclic group including one or more nitrogen, and R1 is hydrogen or an alkyl group of 1-6 carbon atoms. R2 and R3 may be each independently an alkyl group of 1-6 carbon atoms.
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 speech recognition method includes receiving speech data, obtaining candidate texts corresponding to the speech data and respective scores of the candidate texts using a speech recognition model, adjusting the score of a current candidate text, from among the obtained candidate texts, in response to a text length of the current candidate text satisfying a condition determined based on text lengths of the obtained candidate texts, and determining a target text corresponding to the speech data, from among the obtained candidate texts and the current candidate texts.
Abstract:
A sentence mapping method includes obtaining a source language document in a source language and a target language document in a target language, wherein the target language document is a translation of the source language document, generating a translated document by translating the target language document into the source language, and mapping source language sentences in the source language document and target language sentences with the target language document by comparing the source language document and the translated document.
Abstract:
A method and apparatus for training a language model, include generating a first training feature vector sequence and a second training feature vector sequence from training data. The method is configured to perform forward estimation of a neural network based on the first training feature vector sequence, and perform backward estimation of the neural network based on the second training feature vector sequence. The method is further configured to train a language model based on a result of the forward estimation and a result of the backward estimation.