Samsung Electronics Co., Ltd.

Samsung Sds Co., Ltd.

Young-Joon Choi

Jong-Hwa Kim

Dong-Gi Lee

Ji-Hoon Kim

Dong-Hyun Song

Jong-Won Choi

Bum-Soo Kim

Tae-Gyun Kim

Seok-cheon Kwon

In-Young Kim

Jin-Yub Lee

Jae Young Lee

Seok-heon Lee

Sang-Wook Kim

Tae-Sung Kim

Yeo-jun Yoon

Joon-kyu Seo

Hyun-Jin Kim

Chan-Ik Park

Eun-Jin Yun

Nipun Kumar

Shea-Yun Lee

Seok-Cheon Kwon

Hye-Young Kim

Kyeong-Han Lee

Seok-Heon Lee

Byoung-Jip Kim

Myung-Jin Jung

Seong-Won Bak

Jung-Been Im

Hye-jeong Nam

Jin-Hyuk Kim

Sung-Pack Hong

Soon-Young Kim

Byung-Yong Kim

Jong-hee Han

Cheol-Ung Jang

Eui-Bum Han

Hyun-Duk Cho

Yeo-Jun Yoon

Yong-Yeon Lee

Yeon-Ho Kim

Jong-Min Park

Moon-Sang Kwon

Ka-won Cheon

Chae-Hag Yi

Su-Hwan Kim

Dae-Sik Park

Yang-sup Lee

Sung-bin Kuk

Dong-jin Eun

Sung-Hwan Hyun

Jun-Kyu Cho

Rasmus Abildgren

Kook-Hyung Lee

Jae-Hyun Hwang

Jae-hyun Hwang

Yong-Nam Koh

Kang-Doeg Suh

Jung-Beem Im

Kyoong-Han Lee

Jun-Kyo Cho

Seung-Taek Song

Jae-Young 102-803 Jugong 1 Dani Lee

Seok-Heon Samsung Electronics Taiwan Co Ltd Lee

Ji-Yeon Kwak

A method for training a deep neural network according to an embodiment includes training a deep neural network model using a first data set including a plurality of labeled data and a second data set including a plurality of unlabeled data, assigning a ground-truth label value to some of the plurality of unlabeled data, updating the first data set and the second data set such that the data to which the ground-truth label value is assigned is included in the first data set, and further training the deep neural network model using the updated first data set and the updated second data set.

Apparatus and method for training deep neural network

A machine learning apparatus according to an embodiment includes a feature extractor configured to extract features from an object region of an image, a label processor configured to create sentence label embeddings from a sentence label corresponding to the object region, a first training data creator to extract first sub-features from a plurality of first sub-regions created by partitioning the object region, add the sentence label embeddings to the extracted first sub-features, and add the first sub-features added with the sentence label embeddings to the features of the object region, a second training data creator to extract a plurality of second sub-regions along a bounding surface of the object region, create an attention matrix from the second sub-regions, and create a training data by applying the attention matrix to the features of the object region, and a trainer to train an object detection model using the training data.

Machine learning apparatus and method for object detection

Provided are an optimization system and method of a deep learning model. According to example embodiments, by optimizing the structure of the deep learning model appropriately for a target dataset without fixing the structure of the deep learning model, it is possible to generate a model structure capable of having high performance on the target dataset and also saving resources.

System and method for optimization of deep learning model

An image processing apparatus includes a superpixel extractor configured to extract a plurality of superpixels from an input original image, a backbone network including N feature extracting layers (here, N is a natural number of two or more) which divide the input original image into grids including a plurality of regions and generate an output value including a feature value for each of the divided regions, and a superpixel pooling layer configured to generate a superpixel feature value corresponding to each of the plurality of superpixels using a first output value to an Noutput value output from each of the N feature extracting layers.

Apparatus and method for image processing for machine learning

A method of training a model for object classification and detection includes training a first classification model including a shared feature extractor shared by classification models and a first classifier for outputting a result of an object in a first input image based on feature values of the first input image, training a second classification model including the shared feature extractor and a second classifier for outputting a result about authenticity of a second input image based on feature values of the second input image, and training a third classification model including the shared feature extractor and a third classifier for outputting a classification result about a rotation angle of a third input image on the basis of feature values of the third input image extracted by the shared feature extractor, using a third training image set including images rotated at one or more angles.

Method and apparatus for training model for object classification and detection

A method and apparatus for operating an electronic device may include: acquiring the battery status information of two or more element devices, and determining whether the battery status information satisfies a reference condition. A function of at least one device among the two or more element devices is controlled based on the information such as the battery status information.

Method for controlling function based on battery information and electronic device therefor

The present disclosure relates to a method and an electronic device for processing data on the basis of a Bluetooth Low Energy (BLE) protocol. The electronic device includes a Bluetooth control module and a processor. The Bluetooth control module stores a modified BLE protocol stack including at least two Attribute Protocols (ATTs) for use with an application, the modified BLE protocol stack includes a first path and a second path. The first path includes a first ATT protocol to process an ATT command of the application, and the second path includes a second ATT protocol to process an ATT command of the application. The processor is coupled to the Bluetooth control module. The processor processes data of the application including the ATT command, using the modified BLE protocol stack including the at least two ATTs of the Bluetooth control module.

Method for processing data based on bluetooth protocol and electronic device thereof

A vibration control device and method, wherein the vibration control device includes a first driving unit for vibrating the vibration control device up and down, a second driving unit for moving the vibration control device left or right, and a control unit for controlling the first driving unit and the second driving unit, upon an occurrence of an event. The controller controls the second driving unit to move the vibration control device at a time when the first driving unit vibrates the vibration control device off of a surface.

Vibration control device and method

Integrated circuit systems include a non-volatile memory device (e.g, flash EEPROM device) and a memory processing circuit. The memory processing circuit is electrically coupled to the non-volatile memory device. The memory processing circuit is configured to reallocate addressable space within the non-volatile memory device. This reallocation is performed by increasing a number of physical addresses within the non-volatile memory device that are reserved as redundant memory addresses, in response to a capacity adjust command received by the memory processing circuit.

Methods and apparatus for reallocating addressable spaces within memory devices

A non-volatile memory may include a plurality of map blocks for storing a plurality of map units, the map units representing mapping information between physical addresses and logical addresses. A storage device may include such a non-volatile memory. A method of mapping such a non-volatile memory may include writing historical information regarding locations of valid map units among the map units included in map blocks previously allocated among the map blocks when a new map block among the map blocks is allocated, the valid map units representing valid mapping information, and constructing a map table including all of the valid mapping information based on the historical information and a result of searching a map block recently allocated among the map blocks.

Storage device and method of mapping a nonvolatile memory based on a map history

A flash memory device includes a memory cell array, a clock signal input, an input for receiving a signal designating a writing operating mode, a plurality of data input/output pads, and a data input/output buffer circuit that is electrically connected to the clock signal input and to the plurality of data input/output pads. The data input/output buffer circuit is configured to receive data that is to be written to the memory cell array through the data input/output pads in synchronization with a clock signal that is applied to the clock signal input in response to activation of the signal designating the writing operating mode.

Flash memory devices with high data transmission rates and memory systems including such flash memory devices

The invention provides an operation method of a memory system including a flash memory device. The method includes programming at least one page included in a selected memory block of the flash memory device; and determining the selected memory block or the flash memory device to be invalid, according to whether a loop number of the programmed page is out of a reference loop range.

Flash memory device and system including the same

A data output buffer switches it operating mode according to its operating frequency. The data output buffer includes a delay control unit, and a buffer unit. The buffer unit provides data of an internal buffer input line to an external buffer output line. The delay control unit generates a buffer enable signal corresponding to a received reference control signal. The buffer unit blocks the provision of the data to the buffer output line in response to a deactivation of the buffer enable signal. The buffer enable signal remains in an activated state when the period of the reference control signal is shorter than a reference period. The data output buffer may be included in a semiconductor memory device.

Data output buffer whose mode switches according to operation frequency and semiconductor memory device having the same

Growing community of inventors

Seongnam-si, South Korea

Young-Joon Choi