Samsung Electronics Co., Ltd.

Maple Vision Technologies Inc.

Daewoo Electronics Corporation

Pil Sang Yoon

Jun Jin Kong

Nak Young Kim

Kyu Il Jung

Hong Rak Son

Kyoung Lae Cho

Chang Kyu Seol

Chan Ho Yoon

Kwang Hoon Kim

Jin Soo Lim

Eun Chu Oh

Avner Dor

Beom Kyu Shin

Jae Hun Jang

Young Bae Kim

Ju-young Jung

Myung Kyu Lee

Shmuel Dashevsky

Yoav Shereshevski

Han Shin Shin

Hye Jeong So

Ji Youp Kim

Eui Seok Hwang

Ung Hwan Kim

Hak Sun Kim

An artificial intelligence calculation semiconductor device is provided. The artificial intelligence calculation semiconductor device comprising: a control unit; and a MAC (Multiply and Accumulator) calculator which executes a homomorphic encryption calculation through the control unit, wherein the MAC calculator includes an NTT (Numeric Theoretic Transform)/INTT (Inverse NTT) circuit which generates cipher texts by performing a homomorphic multiplication calculation through transformation or inverse transformation of data, a cipher text multiplier which executes a multiplication calculation between the cipher texts, a cipher text adder/subtractor which executes addition and/or subtraction calculations between the cipher texts, and a rotator which performs a cyclic shift of a slot of the cipher texts.

Artificial intelligence calculation semiconductor device and storage device comprising the same

A storage device is provided including an interface circuit configured to receive application information from a host; a field programmable gate array (FPGA); a neural processing unit (NPU); and a central processing unit (CPU) configured to select a hardware image from among a plurality of hardware images stored in a memory using the application information, and reconfigure the FPGA using the selected hardware image. The NPU is configured to perform an operation using the reconfigured FPGA.

Neural processing unit and electronic apparatus including the same

An operating method of a memory controller includes generating a random value using a seed, generating encrypted intermediate data by encrypting plaintext data using the random value, and storing the seed and the intermediate data in a memory device. Ciphertext data is generated using the seed and the intermediate data based on Ring Learning with Error (RLWE).

Operating method of memory controller, cloud computing device and edge computer

A resistive memory system having a plurality of memory cells includes a memory device having a resistive memory cell array and a controller. The controller generates write data to be written to the memory cell array by encoding input data such that the input data corresponds to an erase state and a plurality of programming states that a memory cell may have. The input data is encoded such that at least one of the number of memory cells assigned a first programming state and the number of memory cells assigned a second programming state is smaller than at least one of the numbers of memory cells in the erase state and the other programming states. The first programming state has a highest resistance level among the plurality of programming states, and the second programming state has a second highest resistance level among the plurality of programming states.

Resistive memory system and method of operating the resistive memory system

A method of decoding data of a non-volatile memory device is provided. The method includes a first decoding operation of reading first hard decision data from the non-volatile memory device using a first hard decision read level and performing decoding using the first hard decision data; a second decoding operation of reading first soft decision data from the non-volatile memory device when the decoding fails in the first decoding operation, and performing decoding using the first soft decision; and a third decoding operation of changing from the first hard decision read level to a second hard decision read level when the decoding fails in the second decoding operation, reading second hard decision data using the second hard decision read level, and performing decoding either using the second hard decision data or using both the second hard decision data and the first soft decision data.

Data decoding method of non-volatile memory device and apparatus for performing the method

Provided is a data processing system for recording holographic optical information. The data processing system includes, a data interface constructing a data page by using data transmitted from a host information device, a memory storing data transmitted from the data interface, an encoder ECC-encoding data that is stored in the memory, and a modulator modulating the encoded data so as to record optical information. Accordingly, it is possible to efficiently transmit data when recording and reproducing holographic optical information.

Data processing system and method for recording and reproducing holographic optical information

A method of programming a non-volatile memory device includes; defining a set of verification voltages, setting a maximum verification voltage among verification voltages that are less than or equal to a first target programming voltage to be a target verification voltage, calculating a number of extra pulses based on the target verification voltage and the first target programming voltage, verifying whether a threshold voltage of the memory cell is equal to or greater than the target verification voltage by applying an incremental step pulse program (ISPP) pulse to the memory cell and then applying at least one verification voltage in the set of verification voltages to the memory cell, and further applying the ISPP pulse to the memory cell a number of times equal to the number of extra pulses when the threshold voltage is verified to be equal to or greater than the target verification voltage.

Non-volatile memory device and programming method using fewer verification voltages than programmable data states

A method of operating a data compression device includes analyzing data using an analyzer and generating a result of the analysis, while the data is buffered by an input buffer, and selectively compressing the buffered data according to the result of the analysis. A data compression device includes a data pattern analyzer configured to analyze data transmitted to an input buffer, and generate an analysis code based on the analysis of the data; and a data compression manager configured to selectively compress the data in the input buffer based on the analysis code.

Data compression devices, operating methods thereof, and data processing apparatuses including the same

An optical information reproducing apparatus and an optical information reproducing method using the same are provided. The optical information reproducing apparatus includes: a reference beam providing unit which inputs a reference beam to a reproduction recording area to be reproduced and a peripheral recording area around the reproduction recording area on an optical information recording medium, a peripheral beam detecting unit which transmits a reproduction beam reproduced from the reproduction recording area in response to the reference beam and detects a peripheral beam reproduced from the peripheral recording area, a tracking servo unit which analyzes optical information of the detected peripheral beam and controls a tracking position, and a reproduction beam detecting unit which detects the reproduction beam having pass through the peripheral beam detecting unit. Accordingly, it is possible to judge a current tracking state using the peripheral beam and to perform a tracking servo control process using the judgment result at the time of reproducing optical information.

Apparatus and method for providing a tracking servo control when reproducing optical information

In a method of recording and reading optical information in a recording medium using holography, a first recording operation is performed for recording optical information in a first recording region in an overlapping manner by allowing a reference beam and a signal beam with data loaded to interfere with each other in the first recording region, the reference beam being angularly multiplexed at predetermined intervals. Furthermore, a second recording operation is performed for recording optical information in a second recording region adjacent to the first recording region in an overlapping manner by allowing the reference beam and the signal beam to interfere with each other in the second recording region, the reference beam being angularly multiplexed between the predetermined intervals. Consequently, optical information can be recorded so as to easily detect and determine a selected reproducing beam, thus increasing the reading efficiency of the optical information.

Method of reading optical information and method of recording optical information

Exemplary embodiments are provided for recording and reading optical information using holography in which a first recording operation is performed in a first recording region in an overlapping manner by allowing a reference beam and a signal beam with data loaded to interfere with each other in the first recording region, the reference beam being angularly multiplexed at predetermined intervals. And a second recording operation is performed in a second recording region located adjacent to the first recording region or overlapped with the first recording region by allowing the reference beam and the signal beam to interfere with each other in the second recording region, the reference beam being angularly multiplexed between the predetermined intervals.

Provided are a method for detecting a pattern of an over-sampling image and an optical information processing apparatus and method using the same. A method for detecting a pattern from an image of an over-sampled datapage includes: over-sampling the datapage to detect a detection image of the datapage; comparing the detection image and a reference image of over-sampling for the pattern by a covariance; and calculating a pixel of the pattern over-sampled by values of the reference image and the detection image compared by the covariance. Accordingly, data with a specific pattern such as a mark can be detected from an image of an over-sampled datapage, thereby making it possible to increase the data reproduction accuracy, thus making it possible to greatly increase the reliability of reproduced optical data.

Method for detecting pattern of over-sampling image and an optical information processing apparatus and method using the same

An apparatus for compensating for pixel distortion while reproducing hologram data includes an extraction unit, a determination and calculation unit, a table, and a compensation unit. The extraction unit extracts a reproduced data image from a reproduced image frame including the reproduced data image and borders. The determination and calculation unit determines position values of edges of the extracted reproduced data image, and calculates average magnification error values of pixels within line data from position values of start and end point pixels thereof, which are based on the determined position values of the edges. The table stores misalignment compensation values for the pixels within the line data, wherein the misalignment compensation values correspond to predetermined references for average magnification error values. The compensation unit compensates for pixel positions in the extracted reproduced data image using the misalignment compensation values that correspond to the calculated average magnification error values.

Apparatus and method for compensating for pixel distortion in reproduction of hologram data

For compensating for misalignment of pixels at the time of reproduction in a Holographic Digital Data Storage (HDDS) system, and thereby correcting an image, a stored page-based image is detected using a Charge Coupled Device (CCD) through a one-to-one (1:1) pixel matching between data pixels and CCD detector pixels at the time of reproduction. And then, in case the misalignment of pixels occurs during the 1:1 pixel matching, an amount of the misalignment d (0&#x2266;d&#x3c;1) between a corresponding data pixel and a corresponding CCD detector pixel is calculated. Further, an interference influence of a neighboring pixel on the corresponding CCD detector pixel is calculated based on the calculated amount of the misalignment, to thereby correct image data of the CCD detector pixel. And last, a misalignment influence in accordance with the amount of the misalignment is compensated for, to thereby correct the image data of the CCD detector pixel.

Growing community of inventors

Seoul, South Korea

Pil Sang Yoon