Electronics and Telecommunications Research Institute

Other

Lg Chem, Ltd

Korea Advanced Institute of Science and Technology

Korea Atomic Energy Research Institute

Korea Institute of Machinery & Materials

Center for Advanced Meta-materials

Shanjin Optoelectronics (nanjing) Co., Ltd.

Chan Soo Kim

Tae Whan Yoo

Hyeong Ho Lee

Eung Jin Jang

Beom Seok Lee

Yu Jin Lim

Nam Ho Hur

Jong-Hyun Lee

Young-Su Kwon

Won Sik Cheong

Chun-Gi Lyuh

Hong Chang Shin

Jong Arm Jun

Jeongmin Yang

Hyup Jong Kim

Hyun Mi Kim

Kwangseop Kim

Tae-Whan Yoo

Yong Cheol Peter Cho

Kyou Ho Lee

Jaehoon Chung

Yeong Ho Park

Jae Geun Kim

Sung Hyun Baek

Bumjin Kim

Saerom Kim

Hyun Jae Lee

Jong Hyun Lee

Sang Ho Lee

Jae Hyun Kim

Kug Jin Yun

Je Soo Ko

Gwang Soon Lee

Bong Tae Kim

Jae-Hyun Kim

Hak Joo Lee

Seung Hwan Kim

Yool Kwon

Ik Kyun Kim

Gou Young Koh

Gi Mun Um

Bong-Tae Kim

Bongkyun Jang

In San Jeon

Eung Don Lee

Hyun June Jung

Seung-Mo Lee

Yun Hwangbo

Kyeong Soo Han

Hong Ju Kim

Seung-Hwan Kim

Ho Sook Lee

Gyun Min Lee

Yong Wan Kim

Ho Min Kim

Minhwan Kim

Young Wook Cha

Min Ah Yoon

Jae Keun Kim

Jung-Eun Lee

Doo Sub Eom

Byung Ha Park

Sung-Deok Hong

Disclosed is a neural network accelerator including a maximum value determiner outputting a maximum value based on a first magnitude component corresponding to first input data and a second magnitude component corresponding to second input data, a sign determiner outputting a sign component corresponding to the maximum value among a first sign component corresponding to the first input data and a second sign component corresponding to the second input data, as an output sign component, an offset operator quantizing a difference between the first magnitude component and the second magnitude component and outputting an output offset based on the first sign component, the second sign component, and the quantization result, and a magnitude operator calculating an output magnitude component of an output data based on the maximum value and the output offset. Each of the first input data and the second input data is data on a logarithm domain.

Neural network accelerator configured to perform operation on logarithm domain

A reorganizable neural network computing device is provided. The computing device includes a data processing array unit including a plurality of operators disposed at locations corresponding to a row and a column. One or more chaining paths which transfer the first input data from the operator of the first row of the data processing array to the operator of the second row are optionally formed. The plurality of first data input processors of the computing device transfer the first input data for a layer of the neural network to the operators along rows of the data processing array unit, and the plurality of second data input processors of the computing device transfer the second input data to the operators along the columns of the data processing array.

Reorganizable data processing array for neural network computing

A system for manufacturing an optical display element by laminating an optical film on one surface of a rectangular organic light emitting display panel, the system including: a panel supply unit to which an organic light emitting display panel is supplied; a panel conveying unit which conveys the supplied organic light emitting display panel; an optical film supply unit which supplies an optical film by unwinding the optical film from a raw optical film roll; a laminating unit which laminates the optical film on one surface of the organic light emitting display panel; and a turning unit which rotates the organic light emitting display panel by 90° before the optical film is laminated on one surface of the organic light emitting display panel when a width-directional edge of the supplied organic light emitting display panel does not correspond to a width of the raw optical film roll.

System and method for manufacturing optical display element

According to an embodiment, a wick structure of a heat pipe that is capable of bending while increasing a heat transferring operation limit value is provided by improving the wick structure provided inside the heat pipe. The wick structure of the heat pipe includes a plurality of wicks provided inside a heat pipe, wherein the plurality of wicks include: a first wick provided at one side of a length direction to correspond to a condenser section of a heat pipe; a second wick having one side elongated to be connected to the first wick and provided at an adiabatic section of the heat pipe; and a third wick having one side connected to the other side of the second wick to correspond to an evaporator section of the heat pipe and provided at the other side in the length direction, and the first wick, the second wick and the third wick have effective pore radiuses and pore structures that are different from each other and maintain a movement and a capillary force of an working fluid supplied to the inside of the heat pipe when bending the heat pipe.

Wick structure of heat pipe

An exemplary embodiment of the present invention provides a foreign substance inspection system for a display unit, including: a lighting unit configured to provide incident light to a display unit having an organic light emitting display panel; and a foreign substance detecting unit configured to receive incident light that is reflected by the display unit and detect whether a foreign substance is introduced into the display unit.

Foreign material inspection system of display unit

Provided is a neural network accelerator which performs a calculation of a neural network provided with layers, the neural network accelerator including a kernel memory configured to store kernel data related to a filter, a feature map memory configured to store feature map data which are outputs of the layers, and a Processing Element (PE) array including PEs arranged along first and second directions, wherein each of the PEs performs a calculation using the feature map data transmitted in the first direction from the feature map memory and the kernel data transmitted in the second direction from the kernel memory, and transmits a calculation result to the feature map memory in a third direction opposite to the first direction.

Neural network accelerator including bidirectional processing element array

The present invention relates to a carrier film for transferring a microelement, wherein, when a microelement is pressurized, a uniform pressurizing force is provided to the entire microelement, and a damage to the microelement can be prevented. The carrier film for transferring a microelement comprises: a load control layer having a first elastic modulus and having a space formed in a part of the interior thereof; and an adhesive layer having a second elastic modulus that is smaller than the first elastic modulus, the adhesive layer being provided on the upper portion of the load control layer and configured such that a microelement supposed to be transferred to a target substrate is attached thereto. The load control layer has a first zero-rigidity area that maintains a first stress in a first deformation range from a first strain to a second strain.

Carrier film for transferring microelement

The present invention relates to an optical film laminating system for laminating an optical film on a panel in order to manufacture a display unit wherein a control unit which sets, in accordance with attachment conditions, an optical film laminating mode in respect to the panel to at least one of a first mode in which the panel and the optical film are inputted together between a first roller unit and a second roller unit so that the optical film is laminated on the panel, and a second mode in which the second panel conveying unit holds the panel by suction and conveys the panel and the optical film is inputted between the panel and an attachment roller unit, which is any one of the first roller unit and the second roller unit, so that the optical film is laminated on the panel.

System and apparatus for laminating optical film

A selective transfer roll stamp is provided which rotates on a source substrate to selectively transfer elements arranged on the source substrate. The selective transfer roller stamp includes: a roller unit rotating about an axis of rotation and including projecting portions protruding in a radial direction and recessed portions formed between neighboring projecting portions; and an adhesive layer formed on the outer circumferential surface of the roller unit and coming into contact with the elements, wherein the outer circumferential surface of the adhesive layer is made flat in a circumferential direction.

Selective transfer roll stamp

The present application describes an isolated nucleic acid molecule encoding a polypeptide capable of synchronously binding VEGF polypeptide and placenta growth factor (PIGF) polypeptide comprising a nucleotide sequence encoding a VEGFR1 component.

Glycosylated VEGF decoy receptor fusion protein

A hair dye mixing assembly comprises a bowl comprising at least one arm and an opening and at least one support member arranged about said bowl, and attached to the bowl through the bowl arm. The support members can be immovably fixed to the bowl, movably connected to the bowl to move or rotate in different directions, or a combination thereof.

Hair dye mixing assembly

There are provided a method and apparatus for controlling an electronic device. The apparatus for controlling an electronic device includes a marker recognition unit configured to recognize the marker of the electronic device and a control unit configured to perform communication with the electronic device based on the marker and control the electronic device using a Graphic User Interface (GUI) program received from the electronic device.

Method and apparatus for controlling electronic device

A method of receiving stereoscopic video according to the present invention includes receiving a bit stream including image information, extracting a base image stream corresponding to a base image and an additional image stream corresponding to an additional image from the bit stream, generating the base image and the additional image by decoding the base image stream and the additional image stream, respectively, and generating a left image and a right image by using at least one of the base image and the additional image. According to the present invention, 2D/3D broadcasting service efficiency may be improved.

Method and apparatus for transmitting and receiving stereoscopic video

Provided are method and apparatus for dynamically allocating unlink bandwidth in Ethernet PON. In the method and apparatus, report value, which are received for each logical link identifications (LLID), is controlled to be less than a maximum gate value set for each LLID in sequence for all of LLIDs without a predetermined interval. A gate having a length corresponding to the restricted report value is generated. LLIDs are grouped into classes and processing is done class by class. For each class processing, if maximum processing time is reached or maximum total allocation is reached, the processing is switched to a lower class. Also each class processing, if any higher class has its maximum yield processing time or maximum yield allocated gate reached, the processing is switched to that higher class.

Method and apparatus for dynamically allocating bandwidth by hierarchical weight round robin

Growing community of inventors

Daejeon, South Korea

Chan Soo Kim