Page 10
www.us-tech.com
TechWaTch
“LEGO-Like” Stackable Artificial Intelligence Chips
By Jennifer Chu, MIT News Office
wearable devices don’t have to be shelved or discarded for a newer model. Instead, they could be up- graded with the latest sensors and processors that would snap onto a device’s internal chip, like LEGO bricks incorporated into an existing build. Such reconfig- urable chipware could keep de- vices up to date while reducing our electronic waste. Now MIT engineers have
I
taken a step toward that modular vision with a LEGO-like design for a stackable, reconfigurable ar- tificial intelligence chip.
Lightspeed Communication The design comprises alter-
nating layers of sensing and pro- cessing elements, along with light-emitting diodes (LED) that allow for the chip’s layers to com- municate optically. Other modu- lar chip designs employ conven- tional wiring to relay signals be- tween layers. Such intricate con- nections are difficult if not im- possible to sever and rewire, making such stackable designs
magine a more sustainable future, where cellphones, smartwatches, and other
not reconfigurable. The MIT design uses light,
rather than physical wires, to transmit information through the chip. The chip can therefore be reconfigured, with layers that can be swapped out or stacked on, for instance to add new sen- sors or updated processors. “You can add as many com-
puting layers and sensors as you want, such as for light, pressure, and even smell,” says MIT post- doc Jihoon Kang. “We call this a LEGO-like reconfigurable AI chip because it has unlimited ex- pandability depending on the combination of layers.” The researchers are eager to
apply the design to edge comput- ing devices.
Memristors The team’s design is cur-
rently configured to carry out ba- sic image-recognition tasks. It does so via a layering of image sensors, LEDs, and processors made from artificial synapses — arrays of memory resistors, or “memristors,” that the team pre- viously developed, which togeth- er function as a physical neural
network, or “brain-on-a-chip.” Each array can be trained to process and classify signals di- rectly on a chip, without the need for external software or an Inter- net connection. In their new chip design, the
researchers paired image sen- sors with artificial synapse ar- rays, each of which they trained to recognize certain letters. The team’s optical communi-
cation system consists of paired photodetectors and LEDs, each patterned with tiny pixels. Pho- todetectors constitute an image sensor for re- ceiving data, and LEDs to transmit data to the next layer. As a signal (for in-
Chip by Chip The team fabricated a single
chip, with a computing core measuring about 4 square mil- limeters, or about the size of a piece of confetti. The chip is stacked with three image recog- nition “blocks,” each comprising an image sensor, optical commu- nication layer, and artificial synapse array for classifying one of three letters, M, I, or T. They
stance an image of a letter) reaches the image sensor, the image’s light pattern encodes a certain configuration of LED pixels, which in turn stimulates another layer of photodetectors, along with an artificial synapse array, which classifies the signal based on the pattern and strength of the incoming LED light.
The MIT design uses light, rather than physical wires, to transmit in- formation through the chip. The chip can therefore be reconfigured, with layers that can be swapped out or stacked on, for instance to add new sensors or updated processors.
then shone a pixelated image of random letters onto the chip and measured the electrical current that each neural network array produced in response. (The larg- er the current, the larger the chance that the image is indeed the letter that the particular ar- ray is trained to recognize.) The team found that the
Display, Touch & Embedded solutions for smart charging & e-mobility
We combine state-of-the-art technologies, efficient components, appealing design and professional advice. As a leading expert in the field of industrial display, touch, embedded and system solutions, we offer everything from a single source.
› Wide product portfolio › Suitable for indoor and outdoor use › Extended temperature ranges
› Protection against vandalism, vibrations, moisture and harmful environmental influences
› Easy integration into existing systems › Remote diagnosis and maintenance › 24/7 operation
DATA MODUL Inc. US Headquarters | 275 Marcus Blvd. | Hauppauge, NY 11788 |
info.us@
data-modul.com |
www.data-modul.com
chip correctly classified clear im- ages of each letter, but it was less able to distinguish between blurry images, for instance be- tween I and T. However, the re- searchers were able to quickly swap out the chip’s processing layer for a better “denoising” processor, and found the chip then accurately identified the images. “We showed stackability, re-
placeability, and the ability to in- sert a new function into the chip,” notes MIT postdoc Min- Kyu Song. The researchers plan to add
more sensing and processing ca- pabilities to the chip, and they envision the applications to be boundless, including modular consumer-level devices in which components can be swapped in
and out easily. Contact: Massachusetts
Institute of Technology, 77 Mas- sachusetts Avenue, Cambridge, MA 02139 % 617-253-1000 E-mail:
abby.a@mit.edu Web:
www.mit.edu r
July, 2022
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76
