Toshiba Advances Deep Learning with Extremely Low Power Neuromorphic Processor
Toshiba Corporation (TOKYO:6502) continues to build on its commitment to promoting the Internet of Things and Big Data analysis with development of a Time Domain Neural Network1 (TDNN) that uses an extremely low power consumption neuromorphic semiconductor circuit to perform processing for Deep Learning. TDNN is composed of a massive number of tiny processing units that use Toshiba’s original analog technique, unlike conventional digital processors. TDNN was reported on November 8 at A-SSCC 2016 (Asian Solid-State Circuits Conference 2016), an IEEE-sponsored international conference on semiconductor circuit technology held in Japan.
This Smart News Release features multimedia. View the full release here: http://www.businesswire.com/news/home/20161107005586/en/
Time Domain Neural Network (Graphic: Business Wire)
Deep learning requires massive numbers of calculations, typically executed on high performance processors that consume a lot of power. However, bringing the power of deep learning to IoT edge devices, such as sensors and smart phones requires highly energy efficient ICs that can perform the large number of required operations while consuming extremely little energy.
In von Neumann type2 computer architecture, most energy is consumed moving data from on-chip or off-chip memory devices to the processing unit. The most effective way to reduce movement of a datum is to have massive numbers of processing units, each dedicated to handling only one datum that is located close by. These datum points are given a weight during conversion of an input signal (e.g. an image of a cat) to an output signal (e.g. the recognition of the image as a cat). The closer the datum point is to the desired output, the higher the weight it is given. The weight provides a parameter that automatically guides the deep learning process.
The brain has similar architecture, in that the strength of coupling between neurons (weight data) is built into synapses (processing units). In this case, synapses are connections between neurons and each has a different strength. That strength (weight) determines the signal that passes the connection. In this way, a synapse executes a kind of processing. This architecture, which can be called as fully spatially unrolled architecture, is attractive, but it has an obvious drawback—replicating it on a chip requires a massive number of arithmetic circuits that quickly becomes too large.
Toshiba’s TDNN, which employs time-domain analog and digital mixed signal processing (TDAMS3) techniques developed in 2013 allow miniaturization of the processing unit. In TDAMS, arithmetic operations, such as addition, are performed efficiently by using the delay time of the digital signal passing the logic gate as an analog signal. Using this technique, the processing unit for deep learning can be composed of only three logic gates and a 1-bit memory with the fully spatially unrolled architecture. Toshiba has fabricated a proof-of-concept chip that uses SRAM (static random access memory) cell as the memory and that has demonstrated recognition of handwritten figures. The energy consumption per operation is 20.6 fJ4, which is 1/6x better than previously reported at a leading conference before5.
Toshiba plans to develop TDNN as a resistive random access memory (ReRAM) in order to further improve energy and area efficiencies. The goal is an IC that realizes high performance deep learning technology on edge devices.
- TDNN: Neural network using time-domain analog and digital mixed signal processing technique
- Von Neumann type: The standard and most widely used computer architecture. The architecture loads data from memory devices to the processing unit for processing.
- TDAMS: An analog signal processing technique that uses the delay time of a digital signal passing logic gates as an analog signal. Developed by Toshiba. (reference: http://www.toshiba.co.jp/about/press/2013_02/pr2101.htm)
- 20.6fJ: Equivalent to the performance of 48.6 trillion operations in 1 second with 1 W of power consumption.
- ISSCC 2016 (International Solid-State Circuits Conference 2016) paper number 24.2
About Toshiba Corporation
Toshiba Corporation, a Fortune Global 500 company, channels world-class capabilities in advanced electronic and electrical product and systems into three focus business fields: Energy that sustains everyday life, that is cleaner and safer; Infrastructure that sustains quality of life; and Storage that sustains the advanced information society. Guided by the principles of The Basic Commitment of the Toshiba Group, “Committed to People, Committed to the Future”, Toshiba promotes global operations and is contributing to the realization of a world where generations to come can live better lives.
Founded in Tokyo in 1875, today’s Toshiba is at the heart of a global network of 550 consolidated companies employing 188,000 people worldwide, with annual sales surpassing 5.6 trillion yen (US$50 billion). (As of March 31, 2016.)
To find out more about Toshiba, visit www.toshiba.co.jp/index.htm
Storage & Electronic Devices Solutions Company
Koichi Tanaka / Kota Yamaji
Public Relations & Investor Relations Group
Business Planning Division
Om Business Wire
(c) 2018 Business Wire, Inc., All rights reserved.
Business Wire, a Berkshire Hathaway company, is the global leader in multiplatform press release distribution.
Følg saker fra Business Wire
Registrer deg med din epostadresse under for å få de nyeste sakene fra Business Wire på epost fortløpende. Du kan melde deg av når som helst.
Siste saker fra Business Wire
Trump International Hotel Washington, D.C. Announced as First Forbes Five Star Hotel in Downtown Washington, D.C.22.2.2018 19:21 | Pressemelding
Trump International Hotel Washington, D.C. announces today that it has been awarded the 2018 Forbes Five-Star recognition, making it the first hotel to receive the prestigious distinction in Downtown Washington D.C. Forbes Travel Guide, the global authority on luxury travel, sends anonymous professional inspectors around the world to assess hotels, restaurants and spas against up to 800 objective standards focused on service and facilities. Just opened a short fifteen months ago, the property is located on Pennsylvania Avenue in the center of the city at the site of the Old Post Office. This press release features multimedia. View the full release here: http://www.businesswire.com/news/home/20180222006277/en/ Trump International Hotel Washington, D.C. (Photo: Business Wire) Mickael Damelincourt, Managing Director, oversaw the recruitment and selection of the executives whose teams would lead the hotel through the construction and opening phases, instilling a commitment to the highest l
Entrust Datacard and Infineon Partner to Reduce Complexities with Securing IoT Deployments22.2.2018 18:30 | Pressemelding
Entrust Datacard, a leading provider of trusted identity and secure transaction technology solutions, today announced it has joined the Infineon Security Partner Network (ISPN). The ISPN is part of Infineon Technologies AG, a leading provider of semiconductor solutions used in automotive, industrial, computing and Internet of Things applications. The ISPN is composed of associated and preferred security partners who are proven experts in security consulting, security solution provisioning, electronic design, systems integration and trust services management. The new partnership will apply the Entrust Datacard™ ioTrust™ Security Solution that uses identity-based technology to help securely connect the people, applications and devices that power the connected world. The tiered software uses sophisticated cryptographic technology to provide identity issuance, management and data security of connected things within industrial environments. As a global leader in Trusted Platform Modules (TP
Blue Danube Systems Joins xRAN Forum as a Massive MIMO System Supplier Fully Committed to Open Interfaces22.2.2018 18:10 | Pressemelding
Blue Danube Systems, a provider of intelligent wireless access solutions that help mobile operators address the challenge of explosive data growth, today announced that it has joined the xRAN Forum (www.xran.org). The xRAN Forum was formed by several leading global mobile operators to develop, standardize and promote an open alternative to the traditionally closed, hardware-based RAN (Radio Access Network) architecture. Massive MIMO systems are widely recognized to be a critical element of future 5G wireless networks. Blue Danube’s unique approach to Massive MIMO, based on the company’s underlying High Definition Active Antenna Systems (HDAAS™) technology, not only delivers capacity gains in existing LTE networks with today’s smartphones but also enables intelligent analytics-based RF virtualization for multi-site network performance optimization. Blue Danube’s first Massive MIMO product, the BeamCraft™ 500, has been deployed at four FDD-LTE operators to date, interfacing to four diffe
Blue Danube Systems First to Demonstrate Benefits of FDD Massive MIMO in Clustered Multi-Sector Deployments22.2.2018 18:07 | Pressemelding
Blue Danube Systems, a provider of intelligent wireless access solutions that help mobile operators address the challenge of explosive data growth, today announced it has successfully expanded the commercial deployment of its massive MIMO solution to multiple clustered cell sites, demonstrating total capacity improvements for directly adjoining sectors on existing LTE networks using current smartphones. This press release features multimedia. View the full release here: http://www.businesswire.com/news/home/20180222006158/en/ Blue Danube Systems First to Demonstrate Benefits of FDD Massive MIMO in Clustered Multi-Sector Deployments (Photo: Business Wire) Despite the prevalent view that Massive MIMO systems will be central to 5G networks, it has not been clearly proven that the technology is deployable beyond isolated high-demand sectors where effects such as adjacent cell interference must be addressed. In fact, common digital-only Massive MIMO implementations typically operate as clos
Avnet and IDEMIA Develop Highly Flexible ‘Plug & Play’ Cellular Connectivity Solution for IoT and Industry 4.0 Applications22.2.2018 16:08 | Pressemelding
Avnet (NYSE: AVT), a leading global technology distributor and IDEMIA (formerly known as OT-Morpho), the global leader in Augmented Identity, have joined forces to collaborate on the development of a new future-proof cellular connectivity solution that meets the growing communication requirements for Internet of Things (IoT) and Industry 4.0 applications. This press release features multimedia. View the full release here: http://www.businesswire.com/news/home/20180222005966/en/ (Photo: IDEMIA) The new ‘plug & play’ eUICC (embedded Universal Integrated Circuit Card) and SM-SR (Subscription Management Secure Routing) delivers product future-proofing capability to IoT application solution providers and other businesses. It provides them with the ability to easily choose a mobile network operator (MNO) without eventually having to manually change the embedded SIM card in their cellular connectivity-based devices, machines and sensors that are already deployed in the field. The product bund
TINE Boosts Logistics Performance with Quintiq’s Route Optimization Technology22.2.2018 15:41 | Pressemelding
Quintiq, a Dassault Systèmes company and global leader in supply chain planning and optimization (SCP&O), announced that its tactical planning solution for TINE AS has resulted in more cost-effective routes. Given the success of this project, the leading Norwegian dairy producer is now embarking on a second phase to further enhance its tactical planning and expand the solution’s scope to include operational planning. This press release features multimedia. View the full release here: http://www.businesswire.com/news/home/20180222005936/en/ TINE goes further with optimized logistics. Credit: Bo Mathisen Last year, TINE went live with the first phase of the Quintiq solution, designed to optimize the company’s last-mile delivery of dairy products on a tactical level. TINE’s planners used Quintiq to plan shipment routes from central distribution centers all over Norway to customers. They were able to manage the complexity of trailer swapping, various types of resources, dynamic time window