Abstract: Disclosed are methods of encoding and decoding an audio signal using side information, and an encoder and a decoder for performing the methods. The method of encoding an audio signal using side information includes identifying an input signal, the input signal being an original audio signal, extracting side information from the input signal using a learning model trained to extract side information from a feature vector of the input signal, encoding the input signal, and generating a bitstream by combining the encoded input signal and the side information.

Abstract: Techniques for speech processing using a deep neural network (DNN) based acoustic model front-end are described. A new modeling approach directly models multi-channel audio data received from a microphone array using a first model (e.g., multi-geometry/multi-channel DNN) that includes a frequency aligned network (FAN) architecture. Thus, the first model may perform spatial filtering to generate a first feature vector by processing individual frequency bins separately, such that multiple frequency bins are not combined. The first feature vector may be used similarly to beamformed features generated by an acoustic beamformer. A second model (e.g., feature extraction DNN) processes the first feature vector and transforms it to a second feature vector having a lower dimensional representation. A third model (e.g., classification DNN) processes the second feature vector to perform acoustic unit classification and generate text data. The DNN front-end enables improved performance despite a reduction in microphones.

Abstract: Disclosed are a method for coding a residual signal of LPC coefficients based on collaborative quantization and a computing device for performing the method. The residual signal coding method includes: generating encoded LPC coefficients and LPC residual signals by performing LPC analysis and quantization on an input speech; Determining a predicted LPC residual signal by applying the LPC residual signal to cross module residual learning; Performing LPC synthesis using the coded LPC coefficients and the predicted LPC residual signal; It may include the step of determining an output speech that is a synthesized output according to a result of performing the LPC synthesis.

Abstract: An embodiment comprises: a lens module; a first optical path change part including an incident surface and an emitting surface, and reflecting light, which is incident onto the incident surface, to emit the light via the emitting surface; and a second optical path change part including a first surface which faces the emitting surface in an optical axial direction of the lens module and a second surface which is an opposite surface to the first surface, wherein, in order to change the path of light emitted via the emitting surface, the first optical path change part is tilted or rotated about a first axis perpendicular to the optical axial direction, and, in order to change the path of light emitted from the second surface, the second optical path change part is tilted or rotated about a second axis perpendicular to the optical axial direction and the first axis.

Abstract: Provided is an electrolyte solution for a lithium secondary battery, the electrolyte solution comprising a non-aqueous organic solvent, a lithium salt, and an additive, wherein the additive includes a compound represented by Chemical Formula 1. The contents of Chemical Formula 1 are the same as described in the specification.

Abstract: A communication device of a vehicle transmits a charging permission request message to a server and receives charging permission information from the server. A charging requester is configured to transmit a charging request message including the charging permission information to a charging device and a power receiver is configured to receive a charging response message from the charging device and to wirelessly receive power from the charging device.

Abstract: A storage device includes a memory device including a plurality of memory areas respectively corresponding to a plurality of address groups each of which includes consecutive logical addresses provided by a host, a buffer memory including a common buffer and a plurality of zone buffers, the plurality of zone buffers respectively corresponding to the plurality of memory areas, and a memory controller for controlling the buffer memory to temporarily store write data corresponding to a logical address provided from the host in one of a zone buffer corresponding to the logical address and the common buffer according to whether a first size of the write data exceeds an available storage capacity of the zone buffer, the available storage capacity representing a size of an empty storage space of the zone buffer in which no data is stored.

Abstract: A lens moving apparatus includes a housing for supporting a first magnet, a bobbin, which includes a first coil provided on an outer surface thereof and which moves in a first direction, a support member, which is disposed over one side surface of the housing, and which supports the bobbin and the housing such that the bobbin and the housing are movable in second and/or third directions, which are perpendicular to the first direction, a second coil, which is disposed over the housing, and which is spaced apart from the support member by a predetermined distance and which generates an electromagnetic force to move the support member in the second and/or third directions, and a printed circuit board disposed over the second coil.

Abstract: An audio signal encoding and decoding method using a neural network model, and an encoder and decoder for performing the same are disclosed. A method of encoding an audio signal using a neural network model, the method may include identifying an input signal, generating a quantized latent vector by inputting the input signal into a neural network model encoding the input signal, and generating a bitstream corresponding to the quantized latent vector, wherein the neural network model may include i) a feature extraction layer generating a latent vector by extracting a feature of the input signal, ii) a plurality of downsampling blocks downsampling the latent vector, and iii) a plurality of quantization blocks performing quantization of a downsampled latent vector.

Abstract: Embodiments of the present disclosure relate to a memory system and an operating method of the memory system. According to embodiments of the present disclosure, a memory system may store, for each of the plurality of memory dies, a program fail count indicating a cumulative number of occurrences of a program fail during a program operation for each memory die, and may change, for a target memory die among the plurality of memory dies, a first operation parameter among operation parameters applied to the target memory die when executing a program operation on the target memory die, based on a target program fail count which is the program fail count for the target memory die.

Abstract: A method of training a malicious code detection model and a method of detecting a malicious code. The method includes acquiring application programming interface (API) call information of called functions from a result log of performing dynamic analysis of a malicious code, calculating time intervals between timestamps using the timestamps which indicate API call times extracted from the API call information, determining a feature value of the malicious code on the basis of the time intervals, and training the malicious code detection model using an API call sequence included in the API call information and the feature value.

Abstract: While operating a vehicle on a travel path, a width of the travel path is determined based on a map. Upon detecting a presence or an absence of a target vehicle in the travel path, a lateral position for the vehicle on the travel path is determined based on the width of the travel path and the presence or absence of the target vehicle. The vehicle is controlled to operate according to the determined lateral position on the travel path and a speed of the vehicle that is based on the determined lateral position on the travel path.

Abstract: An object of the present invention is to provide a return service method for enabling quick return processing by using a waybill of a special form in which return information is included and return boxes scattered in multiple regions.

Abstract: An embodiment comprises: a base which comprises a body and a first column part, a second column part, a third column part, and a fourth column part arranged at corner parts of the body, respectively; a bobbin disposed on the body; a first roller part disposed in a first groove provided on the first column part; a second roller part disposed in a second groove provided on the second column part adjacent to the first column part; a magnet disposed on the bobbin; and a coil corresponding to the magnet and disposed between the first and the second column part, wherein the first roller part contacts at least two areas of the first groove, the second roller part contacts at least two areas of the second groove, the first groove comprises a first opening through which a part of the first roller part is exposed, the second groove comprises a second opening through which a part of the second roller part is exposed, and the bobbin comprises a first support part disposed in the first opening and contacting the first rol

Abstract: A semiconductor device is provided. The semiconductor device includes a substrate a bit line structure disposed on the substrate, a trench adjacent to at least one side of the bit line structure, a storage contact structure disposed within the trench, and comprising a storage contact, a silicide layer, and a storage pad which are stacked sequentially. A spacer structure is disposed between the bit line structure and the storage contact structure.

Abstract: The present invention relates to a chest expander, wherein the chest expander includes a pair of operating arms that are connected by a rotating shaft to approach or be spaced apart from each other, a plurality of elastic force adjusting holes formed as through holes in a longitudinal direction of the operating arm, elastic members that are disposed between the operating arms to cross and configured to provide an elastic force in a direction in which the operating arms are spaced apart from each other when the operating arms approach each other, and elastic force adjusting parts that allow one end portions of the elastic members to move along the elastic force adjusting holes and to be fixed so that lengths of the elastic members between the operating arms are changed so as to adjust strength levels of the elastic forces of the elastic members.

Abstract: Disclosed herein are a gate driving circuit and a flexible display using the same. In gate driving circuit, a control block including an nth controller configured to generate one among a gate-off voltage, an EM start signal, and a carry signal from a (n-1)th EM signal transfer part as an nth EM output control signal to apply the nth EM output control signal to a start terminal of an nth EM signal transfer part on the basis of an nth scan signal (n is a natural number) from the first shift register, a first control signal designating an off-driving signal transfer parts and an on-driving signal transfer parts among the EM signal transfer parts, a second control signal designating a first on-driving signal transfer part, and a third control signal designating second to last on-driving signal transfer parts.

Abstract: Disclosed are a method of encoding a high band of an audio, a method of decoding a high band of an audio, and an encoder and a decoder for performing the methods. The method of decoding a high band of an audio, the method performed by a decoder, includes identifying a parameter extracted through a first neural network, identifying side information extracted through a second neural network, and restoring a high band of an audio by applying the parameter and the side information to a third neural network.

Abstract: Provided are an apparatus and a method for integrally encoding and decoding a speech signal and a audio signal. The encoding apparatus may include: an input signal analyzer to analyze a characteristic of an input signal; a first conversion encoder to convert the input signal to a frequency domain signal, and to encode the input signal when the input signal is a audio characteristic signal; a Linear Predictive Coding (LPC) encoder to perform LPC encoding of the input signal when the input signal is a speech characteristic signal; and a bitstream generator to generate a bitstream using an output.

Abstract: A user trained map for providing driver assistance in a vehicle is obtained by entering a training mode, acquiring and at least temporarily storing vehicle sensor data related to the vehicle's position during training mode while the vehicle is operating in training mode, generating and storing a trained digital map of a parking route at least partially based on the acquired vehicle sensor data, acquiring and storing position information related to a drop-off location and to a pickup location, the drop-off location and the pickup location being part of the parking route, and exiting the training mode.