Back to Table of contents

Primeur weekly 2018-02-19

Exascale supercomputing

Ready for Exascale: researchers find algorithm for large-scale brain simulations on next-generation supercomputers ...

Quantum computing

Researchers demonstrate promising method for improving quantum information processing ...

Fingerprints of quantum entanglement ...

Focus on Europe

Karel Luyben appointed Dutch National Coordinator for Open Science ...

Middleware

An OLCF-developed visualization tool offers customization and faster rendering ...

Hardware

Dell EMC next generation converged infrastructure makes data centre modernization even simpler ...

Cray reports 2017 full year and fourth quarter financial results ...

Michael Levine and Ralph Roskies Day proclaimed in Pittsburgh and Allegheny County ...

MACH-2 put into operation at 77 trillion operations/second ...

Applications

CENIC recognizes technology projects to combat California wildfires ...

3D-e-Chem team develops building blocks and recipes for computer-aided drug discovery ...

Researchers find blood pressure drug holds promise for preventing onset of Type 1 diabetes ...

NCSA allocates over $2.4 million in new Blue Waters supercomputer awards to Illinois researchers ...

NCSA researchers create one of the most reliable tools for long-term crop prediction in the U.S. Corn Belt ...

Embracing complexity in biological systems ...

Particle interactions calculated on Titan support the search for new physics discoveries ...

GM revs up diesel combustion modeling on Titan supercomputer ...

NEC and Tohoku University succeed in AI-based new material development ...

Physics data processing at NERSC dramatically cuts reconstruction time ...

Advanced computing and water management at the AAAS Meeting 2018 ...

Neural networks everywhere ...

Supermassive black hole model predicts characteristic light signals at cusp of collision ...

New turbulent transport modelling shows multiscale fluctuations in heated plasma ...

TACC and DOD engage in four-year transformational design project ...

The Cloud

Atos signs key contract with the European Space Agency to enable new services with satellite data ...

Oracle buys Zenedge ...

NEC contributes to ecosystem for enhancing virtualized network orchestration ...

Neural networks everywhere


MIT researchers have developed a special-purpose chip that increases the speed of neural-network computations by three to seven times over its predecessors, while reducing power consumption 93 to 96 percent. That could make it practical to run neural networks locally on smartphones or even to embed them in household appliances. Image: Chelsea Turner/MIT.
14 Feb 2018 Cambridge - Most recent advances in artificial-intelligence systems such as speech- or face-recognition programmes have come courtesy of neural networks, densely interconnected meshes of simple information processors that learn to perform tasks by analyzing huge sets of training data.

But neural nets are large, and their computations are energy intensive, so they're not very practical for handheld devices. Most smartphone apps that rely on neural nets simply upload data to internet servers, which process it and send the results back to the phone.

Now, MIT researchers have developed a special-purpose chip that increases the speed of neural-network computations by three to seven times over its predecessors, while reducing power consumption 94 to 95 percent. That could make it practical to run neural networks locally on smartphones or even to embed them in household appliances.

"The general processor model is that there is a memory in some part of the chip, and there is a processor in another part of the chip, and you move the data back and forth between them when you do these computations", stated Avishek Biswas, an MIT graduate student in electrical engineering and computer science, who led the new chip's development.

"Since these machine-learning algorithms need so many computations, this transferring back and forth of data is the dominant portion of the energy consumption. But the computation these algorithms do can be simplified to one specific operation, called the dot product. Our approach was, can we implement this dot-product functionality inside the memory so that you don't need to transfer this data back and forth?"

Avishek Biswas and his thesis advisor, Anantha Chandrakasan, dean of MIT's School of Engineering and the Vannevar Bush Professor of Electrical Engineering and Computer Science, describe the new chip in a paper that Avishek Biswas has presented at the International Solid State Circuits Conference.

Neural networks are typically arranged into layers. A single processing node in one layer of the network will generally receive data from several nodes in the layer below and pass data to several nodes in the layer above. Each connection between nodes has its own "weight", which indicates how large a role the output of one node will play in the computation performed by the next. Training the network is a matter of setting those weights.

A node receiving data from multiple nodes in the layer below will multiply each input by the weight of the corresponding connection and sum the results. That operation - the summation of multiplications - is the definition of a dot product. If the dot product exceeds some threshold value, the node will transmit it to nodes in the next layer, over connections with their own weights.

A neural net is an abstraction: The "nodes" are just weights stored in a computer's memory. Calculating a dot product usually involves fetching a weight from memory, fetching the associated data item, multiplying the two, storing the result somewhere, and then repeating the operation for every input to a node. Given that a neural net will have thousands or even millions of nodes, that's a lot of data to move around.

But that sequence of operations is just a digital approximation of what happens in the brain, where signals traveling along multiple neurons meet at a "synapse", or a gap between bundles of neurons. The neurons' firing rates and the electrochemical signals that cross the synapse correspond to the data values and weights. The MIT researchers' new chip improves efficiency by replicating the brain more faithfully.

In the chip, a node's input values are converted into electrical voltages and then multiplied by the appropriate weights. Only the combined voltages are converted back into a digital representation and stored for further processing.

The chip can thus calculate dot products for multiple nodes - 16 at a time, in the prototype - in a single step, instead of shuttling between a processor and memory for every computation.

One of the keys to the system is that all the weights are either 1 or -1. That means that they can be implemented within the memory itself as simple switches that either close a circuit or leave it open. Recent theoretical work suggests that neural nets trained with only two weights should lose little accuracy - somewhere between 1 and 2 percent.

Avishek Biswas and Anantha Chandrakasan's research bears that prediction out. In experiments, they ran the full implementation of a neural network on a conventional computer and the binary-weight equivalent on their chip. Their chip's results were generally within 2 to 3 percent of the conventional network's.

Source: Massachusetts Institute of Technology - MIT

Back to Table of contents

Primeur weekly 2018-02-19

Exascale supercomputing

Ready for Exascale: researchers find algorithm for large-scale brain simulations on next-generation supercomputers ...

Quantum computing

Researchers demonstrate promising method for improving quantum information processing ...

Fingerprints of quantum entanglement ...

Focus on Europe

Karel Luyben appointed Dutch National Coordinator for Open Science ...

Middleware

An OLCF-developed visualization tool offers customization and faster rendering ...

Hardware

Dell EMC next generation converged infrastructure makes data centre modernization even simpler ...

Cray reports 2017 full year and fourth quarter financial results ...

Michael Levine and Ralph Roskies Day proclaimed in Pittsburgh and Allegheny County ...

MACH-2 put into operation at 77 trillion operations/second ...

Applications

CENIC recognizes technology projects to combat California wildfires ...

3D-e-Chem team develops building blocks and recipes for computer-aided drug discovery ...

Researchers find blood pressure drug holds promise for preventing onset of Type 1 diabetes ...

NCSA allocates over $2.4 million in new Blue Waters supercomputer awards to Illinois researchers ...

NCSA researchers create one of the most reliable tools for long-term crop prediction in the U.S. Corn Belt ...

Embracing complexity in biological systems ...

Particle interactions calculated on Titan support the search for new physics discoveries ...

GM revs up diesel combustion modeling on Titan supercomputer ...

NEC and Tohoku University succeed in AI-based new material development ...

Physics data processing at NERSC dramatically cuts reconstruction time ...

Advanced computing and water management at the AAAS Meeting 2018 ...

Neural networks everywhere ...

Supermassive black hole model predicts characteristic light signals at cusp of collision ...

New turbulent transport modelling shows multiscale fluctuations in heated plasma ...

TACC and DOD engage in four-year transformational design project ...

The Cloud

Atos signs key contract with the European Space Agency to enable new services with satellite data ...

Oracle buys Zenedge ...

NEC contributes to ecosystem for enhancing virtualized network orchestration ...