Nippon Telegraph and Telephone Corporation

Mitsubishi Kasei Corporation

Hitachi-ge Nuclear Energy, Ltd.

Mitsubishi Chemical Industries Limited

Sony Corporation

Dowa Metaltech Co., Ltd.

Sumitomo Metal Mining Company, Ltd.

Joled Inc.

Tohoku Electric Power Company, Inc.

Manac Incorporated

Acrodea, Inc.

Tsuyoshi Ito

Takeshi Sakamoto

Yuki Arikawa

Hideyuki Hosokawa

Kenji Tanaka

Kenji Kawai

Makoto Nagase

Junichi Kato

Kazuhiko Terada

Huycu Ngo

Takaharu Itagaki

Yukio Hirama

Hiroto Narieda

Junya Watanabe

Naoshi Usui

Hiroko Sukeda

Hiroshi Kusano

Tsutomu Takeya

Motoaki Sakashita

Hitoshi Kubota

Shintaro Sawada

Takayuki Tashiro

Mayu Sasaki

Etsuko Nakamura

Hiroaki Takayanagi

Yoichiro Sako

Akira Tange

Akira Sugawara

Masashi Takeda

Ken Tsuzuki

Makoto Tachibana

Kazushige Ishida

Naoki Miura

Yoshiharu Morita

Mikihiro Yokozeki

Keisuke Wada

Tsuyoshi Nagata

Yoichi Wada

Yuta Ukon

Motohiro Aizawa

Manabu Kodate

Masaru Katayama

Takashi Ushikubo

Hirohisa Kubota

Ryuzo Iga

Tomotsugu Aoyama

Yuta Sonoda

Masao Tomoi

Hiromi Kochi

Kazuhiko Onuma

Eiji Miyata

Yukiya Kato

Junya Tsutsumi

Shigenori Wada

Kuniaki Miyagi

Teruo Onozuka

Kiyoto Ando

Hiromi Teshima

Kenji Fukunaga

Shuichi Shima

Shigeo Fujisaki

Hiroki Jinno

Go Honda

Tetsurou Gotou

Atsuo Kiyokawa

Tooru Kitazaki

Hiromasa Ogawa

An embodiment is a computer including a plurality of accelerators, a computer for distributed processing includes a plurality of accelerators to each of which a part of a neural network is assigned and each of which is configured to derive a learning result based on input data and update each parameter value included in the part of the neural network by using the learning result; a plurality of network interface circuits each of which is configured to transmit and receive information on learning including the learning result via a network, and an arithmetic processing unit that is configured to control the plurality of accelerators and the plurality of network interface circuits to cause each of the plurality of accelerators to derive a learning result based on input data and to cause the plurality of network interface circuits to transmit and receive information on learning including the learning result.

Distributed processing computer and distributed deep learning system

A data transfer device transfers a received packet. If the transfer destination of the received packet is a data processing device under control, the data transfer device (confirms the state of a shared bus, and if the shared bus is usable, transfers the packet to the data processing device of the transfer destination via the shared bus, gives the use right of the shared bus to one of the data processing devices in response to communication requirements from the data processing devices, and receives, via the shared bus, a packet from the data processing device given the use right.

Data transfer device, data processing device and data processing method

A transfer processing device includes an arithmetic instruction number acquisition circuit, a buffer circuit, a transfer information acquisition circuit, and a software processing unit. The arithmetic instruction number acquisition circuit acquires a transfer instruction number corresponding to transfer information which is information related to the next transfer destination of an arithmetic instruction. The buffer circuit is arranged between the arithmetic instruction number acquisition circuit and the transfer information acquisition circuit, and temporarily stores and relays the arithmetic instruction and the arithmetic instruction number supplied from the arithmetic instruction number acquisition circuit to the transfer information acquisition circuit. The transfer information acquisition circuit acquires transfer information on the basis of the arithmetic instruction number, and gives the acquired transfer information to the arithmetic instruction. The buffer circuit holds the arithmetic instruction number and the arithmetic instruction until an output request from the transfer information acquisition circuit is received.

Transfer processing device

A distributed deep learning system that can achieve speeding-up by processing learning in parallel at a large number of learning nodes connected with a communication network and perform faster cooperative processing among the learning nodes connected through the communication network is provided. The distributed deep learning system includes: a plurality of computing interconnect devicesconnected with each other through a ring communication networkthrough which communication is possible in one direction; and a plurality of learning nodesconnected with the respective computing interconnect devicesin a one-to-one relation, and each computing interconnect deviceexecutes communication packet transmission-reception processing between the learning nodesand All-reduce processing simultaneously in parallel.

Distributed deep learning system, distributed deep learning method, and computing interconnect device

Each NIC performs an aggregation calculation of data output from each processor in a normal order including a head NIC located at a head position of a first pipeline connection, an intermediate NIC located at an intermediate position, and a tail NIC located at a tail position, and when the aggregation calculation in the tail NIC is completed, each NIC starts distribution of an obtained aggregation result, distributes the aggregation result in a reverse order including the tail NIC, the intermediate NIC, and the head NIC, and outputs the aggregation result to the processor of the communication interface.

Distributed processing system and method

A distributed processing node includes a computing device that calculates gradient data of a loss function from an output result obtained by inputting learning data to a learning target model, an interconnect device that aggregates gradient data between the distributed processing node and other distributed processing nodes, a computing function unit that is provided in a bus device and performs processing of gradient data from the computing device, and a DMA controller that controls DMA transfer of gradient data between the computing device and the bus device and DMA transfer of gradient data between the bus device and the interconnect device.

Distributed processing node and distributed processing system

Each of learning nodes calculates gradients of a loss function from an output result obtained by inputting learning data to a learning target neural network, converts a calculation result into a packet, and transmits the packet to a computing interconnect device. The computing interconnect device receives the packet transmitted from each of the learning nodes, acquires a value of the gradients stored in the packet, calculates a sum of the gradients, converts a calculation result into a packet, and transmits the packet to each of the learning nodes. Each of the learning nodes receives the packet transmitted from the computing interconnect device and updates a constituent parameter of a neural network based on a value stored in the packet.

Distributed deep learning system using a communication network for stochastic gradient descent calculations

More accurate prediction of reactor water quality of a nuclear reactor is implemented. A device for prediction of reactor water quality of a nuclear reactor in a nuclear power plant is disclosed. The device stores a reactor water quality prediction model which is learned using learning data, and with which future reactor water quality is predicted. An explanatory variable of the reactor water quality prediction model includes a value in a predetermined period unit that is generated from data acquired in an operating nuclear power plant. The device generates the value in a predetermined period unit from data acquired in a target operating nuclear power plant, and acquires a predicted value of the reactor water quality in the target nuclear power plant based on the reactor water quality prediction model and the value in a predetermined period unit.

Device for prediction of reactor water quality of nuclear reactor

Individual distributed processing nodes packetize distributed data for each weight of a neural network of a learning object in an order of a number of the weight, transmit the distributed data to an aggregation processing node, acquire aggregation data transmitted from the node in order, and update the weight of the neural network. The node acquires the transmitted distributed data, packetizes the aggregation data for which the distributed data of all the distributed processing nodes is aggregated for each weight, and transmits the aggregation data to the individual nodes. The individual nodes monitor an unreceived data amount which is a difference between data amounts of the transmitted distributed data and the acquired aggregation data, and when the unreceived data amount becomes equal to or larger than a threshold Ma, stops transmission of the distributed data until the unreceived data amount becomes equal to or smaller than a threshold Mb (Mb&#x3c;Ma).

Distributed processing system and distributed processing method

A reactor water radioactivity concentration of a nuclear power plant can be predicted with high accuracy. First, a plant state quantity prediction value is calculated by using a physical model that describes plant state quantities of the power plant including a flow rate of feedwater and a metal corrosion product concentration in feedwater of the reactor water is calculated. Next, data for supervised learning is created, and the data for supervised learning includes the previously calculated plant state quantity prediction value and a plant state quantity such as the flow rate of feedwater, the metal corrosion product concentration in feedwater, a metal corrosion product concentration in reactor water, and a radioactive metal corrosion concentration of the reactor water in the reactor as input data and includes a radioactive metal corrosion concentration in the reactor water which is an actual measured value as output data, and a predictive model is trained.

Predictive model construction and prediction method for radioactive metal corrosion concentration in nuclear reactor water

A first distributed processing node transmits distributed data to a second distributed processing node as intermediate consolidated data. A third distributed processing node generates intermediate consolidated data after update from received intermediate consolidated data and distributed data, and transmits the intermediate consolidated data to a fourth distributed processing node. The first distributed processing node transmits the received intermediate consolidated data to fifth distributed processing node as consolidated data. The third distributed processing node transmits the received consolidated data to a sixth distributed processing node. When an aggregation communication time period required by each distributed processing node to consolidate the distributed data or an aggregation dispatch communication time period being a total time period of the aggregation communication time period and a time period required by each distributed processing node to dispatch the consolidated data exceeds a predetermined time period, the first distributed processing node issues a warning.

There are provided a composite plated product, which has little uneven appearance, a low contact resistance and good wear resistance, and a method for producing the same without the need of any silver-plating solutions containing cyanides and any silver-plating solutions containing silver nitrate as a silver salt. After carbon particles are caused to be suspended in water, an oxidizing agent is added thereto for carrying out a wet oxidation treatment of the carbon particles, and a silver-plating solution, which contains at least one sulfonic acid and the carbon particles treated by the wet oxidation treatment, is used for electroplating a base material to form a coating film of a composite material, which contains the carbon particles in a silver layer, on the base material to produce a composite plated product.

Composite plated product and method for producing same

There is provided a system () including a ball (), which incorporates a first sensor that functions as at least a triaxial acceleration sensor, a triaxial gyro sensor, and a triaxial geomagnetic sensor and also includes a first communication unit () that wirelessly transmits sensor data detected by the first sensor; and a mobile terminal () including a second communication unit () that is paired with the first communication unit. The mobile terminal includes: a unit () that acquires external information indicating an environment in which the paired ball moves independently; and a unit () that generates ball movement data () of the paired ball in which the sensor data of the paired ball obtained via the first communication unit and the second communication unit is associated with the external information.

System comprising ball with embedded sensor

A copper alloy sheet material has a copper alloy component system that has a high conductivity of 75.0% IACS or more and has both high strength and good stress relaxation resistance characteristics. A copper alloy sheet material has a composition containing, by mass %, from 0.01 to 0.50% of Zr, from 0.01 to 0.50% of Sn, a total content of from 0 to 0.50% of Mg, Al, Si, P, Ti, Cr, Mn, Co, Ni, Zn, Fe, Ag, Ca, and B, with the balance Cu, and unavoidable impurities, and a metal structure having a number density NOf fine second phase particles having a particle diameter of approximately from 5 to 50 nm of 10.0 per 0.12 mmor more and a ratio N/Nof a number density N(per 0.012 mm) of coarse second phase particles having a particle diameter exceeding approximately 0.2 mm and the Nof 0.50 or less.

Copper alloy sheet material

A cone valve of the present invention enables to extend a lifespan more than a conventional cone valve even if it is used as a check valve when feeding slurry containing highly abrasive coarse particles. A cone valve () used as a check valve when feeding slurry, comprising at least a valve body (), a valve seat (), and a spring () incorporated to make the valve body () contact the valve seat (), wherein an entire length of the spring () is at least shorter than a stroke length of the valve body ().

Cone valve

Growing community of inventors

Tokyo, Japan

Tsuyoshi Ito