Tag Archives: Hyung-jun Kim

Reconfiguring a LEGO-like AI chip with light

MIT engineers have created a reconfigurable AI chip that comprises alternating layers of sensing and processing elements that can communicate with each other. Credit: Figure courtesy of the researchers and edited by MIT News

This image certainly challenges any ideas I have about what Lego looks like. It seems they see things differently at the Massachusetts Institute of Technology (MIT). From a June 13, 2022 MIT news release (also on EurekAlert),

Imagine a more sustainable future, where cellphones, smartwatches, and other wearable devices don’t have to be shelved or discarded for a newer model. Instead, they could be upgraded with the latest sensors and processors that would snap onto a device’s internal chip — like LEGO bricks incorporated into an existing build. Such reconfigurable chipware could keep devices up to date while reducing our electronic waste. 

Now MIT engineers have taken a step toward that modular vision with a LEGO-like design for a stackable, reconfigurable artificial intelligence chip.

The design comprises alternating layers of sensing and processing elements, along with light-emitting diodes (LED) that allow for the chip’s layers to communicate optically. Other modular chip designs employ conventional wiring to relay signals between layers. Such intricate connections are difficult if not impossible to sever and rewire, making such stackable designs 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 sensors or updated processors.

“You can add as many computing layers and sensors as you want, such as for light, pressure, and even smell,” says MIT postdoc Jihoon Kang. “We call this a LEGO-like reconfigurable AI chip because it has unlimited expandability depending on the combination of layers.”

The researchers are eager to apply the design to edge computing devices — self-sufficient sensors and other electronics that work independently from any central or distributed resources such as supercomputers or cloud-based computing.

“As we enter the era of the internet of things based on sensor networks, demand for multifunctioning edge-computing devices will expand dramatically,” says Jeehwan Kim, associate professor of mechanical engineering at MIT. “Our proposed hardware architecture will provide high versatility of edge computing in the future.”

The team’s results are published today in Nature Electronics. In addition to Kim and Kang, MIT authors include co-first authors Chanyeol Choi, Hyunseok Kim, and Min-Kyu Song, and contributing authors Hanwool Yeon, Celesta Chang, Jun Min Suh, Jiho Shin, Kuangye Lu, Bo-In Park, Yeongin Kim, Han Eol Lee, Doyoon Lee, Subeen Pang, Sang-Hoon Bae, Hun S. Kum, and Peng Lin, along with collaborators from Harvard University, Tsinghua University, Zhejiang University, and elsewhere.

Lighting the way

The team’s design is currently configured to carry out basic 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 previously developed, which together function as a physical neural network, or “brain-on-a-chip.” Each array can be trained to process and classify signals directly on a chip, without the need for external software or an Internet connection.

In their new chip design, the researchers paired image sensors with artificial synapse arrays, each of which they trained to recognize certain letters — in this case, M, I, and T. While a conventional approach would be to relay a sensor’s signals to a processor via physical wires, the team instead fabricated an optical system between each sensor and artificial synapse array to enable communication between the layers, without requiring a physical connection. 

“Other chips are physically wired through metal, which makes them hard to rewire and redesign, so you’d need to make a new chip if you wanted to add any new function,” says MIT postdoc Hyunseok Kim. “We replaced that physical wire connection with an optical communication system, which gives us the freedom to stack and add chips the way we want.”

The team’s optical communication system consists of paired photodetectors and LEDs, each patterned with tiny pixels. Photodetectors constitute an image sensor for receiving data, and LEDs to transmit data to the next layer. As a signal (for instance 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.

Stacking up

The team fabricated a single chip, with a computing core measuring about 4 square millimeters, or about the size of a piece of confetti. The chip is stacked with three image recognition “blocks,” each comprising an image sensor, optical communication layer, and artificial synapse array for classifying one of three letters, M, I, or T. They then shone a pixellated image of random letters onto the chip and measured the electrical current that each neural network array produced in response. (The larger the current, the larger the chance that the image is indeed the letter that the particular array is trained to recognize.)

The team found that the chip correctly classified clear images of each letter, but it was less able to distinguish between blurry images, for instance between I and T. However, the researchers 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, replaceability, and the ability to insert a new function into the chip,” notes MIT postdoc Min-Kyu Song.

The researchers plan to add more sensing and processing capabilities to the chip, and they envision the applications to be boundless.

“We can add layers to a cellphone’s camera so it could recognize more complex images, or makes these into healthcare monitors that can be embedded in wearable electronic skin,” offers Choi, who along with Kim previously developed a “smart” skin for monitoring vital signs.

Another idea, he adds, is for modular chips, built into electronics, that consumers can choose to build up with the latest sensor and processor “bricks.”

“We can make a general chip platform, and each layer could be sold separately like a video game,” Jeehwan Kim says. “We could make different types of neural networks, like for image or voice recognition, and let the customer choose what they want, and add to an existing chip like a LEGO.”

This research was supported, in part, by the Ministry of Trade, Industry, and Energy (MOTIE) from South Korea; the Korea Institute of Science and Technology (KIST); and the Samsung Global Research Outreach Program.

Here’s a link to and a citation for the paper,

Reconfigurable heterogeneous integration using stackable chips with embedded artificial intelligence by Chanyeol Choi, Hyunseok Kim, Ji-Hoon Kang, Min-Kyu Song, Hanwool Yeon, Celesta S. Chang, Jun Min Suh, Jiho Shin, Kuangye Lu, Bo-In Park, Yeongin Kim, Han Eol Lee, Doyoon Lee, Jaeyong Lee, Ikbeom Jang, Subeen Pang, Kanghyun Ryu, Sang-Hoon Bae, Yifan Nie, Hyun S. Kum, Min-Chul Park, Suyoun Lee, Hyung-Jun Kim, Huaqiang Wu, Peng Lin & Jeehwan Kim. Nature Electronics volume 5, pages 386–393 (2022) 05 May 2022 Issue Date: June 2022 Published: 13 June 2022 DOI: https://doi.org/10.1038/s41928-022-00778-y

This paper is behind a paywall.

Inspiration from the sea for titanium implants (mussels) and adhesive panels for flexible sensors (octopuses/octopi/octopodes)

I have two sea-inspired news bits both of which concern adhesion.

Mussels and titanium implants

A July 8, 2016 news item on ScienceDaily features some mussel-inspired research from Japan into how to make better titanium implants,

Titanium is used medically in applications such as artificial joints and dental implants. While it is strong and is not harmful to tissues, the metal lacks some of the beneficial biological properties of natural tissues such as bones and natural teeth. Now, based on insights from mussels–which are able to attach themselves very tightly to even metallic surfaces due to special proteins found in their byssal threads–scientists from RIKEN have successfully attached a biologically active molecule to a titanium surface, paving the way for implants that can be more biologically beneficial.

A July 11, 2016 RIKEN press release (also on EurekAlert but dated July 8, 2016), which originated the news item, provides more information,

The work began from earlier discoveries that mussels can attach to smooth surfaces so effectively thanks to a protein, L-DOPA, which is known to be able to bind very strongly to smooth surfaces such as rocks, ceramics, or metals (…). Interestingly, the same protein functions in humans as a precursor to dopamine, and is used as a treatment for Parkinson’s disease.

According to Chen Zhang of the RIKEN Nano Medical Engineering Laboratory, the first author of the paper published in Angewandte Chemie, “We thought it would be interesting to try to use various techniques to attach a biologically active protein—in our case we chose insulin-like growth factor-1, a promoter of cell proliferation—to a titanium surface like those used in implants” (…).

Using a combination of recombinant DNA technology and treatment with tyrosinase, they were able to create a hybrid protein that contained active parts of both the growth factor and L-DOPA. Tests showed that the proteins were able to fold normally, and further experiments in cell cultures demonstrated that the IGF-1 was still functioning normally. Thanks to the incorporation of the L-DOPA, the team was able to confirm that the proteins bound strongly to the titanium surface, and remained attached even when the metal was washed with phosphate-buffered saline, a water-based solution. Zhang says, “This is similar to the powerful properties of mussel adhesive, which can remain fixed to metallic materials even underwater.”

According to Yoshihiro Ito, Team Leader of the Emergent Bioengineering Research Team of the RIKEN Center for Emergent Matter Science, “We are very excited by this finding, because the modification process is a universal one that could be used with other proteins. It could allow us to prepare new cell-growth enhancing materials, with potential applications in cell culture systems and regenerative medicine. And it is particularly interesting that this is an example of biomimetics, where nature can teach us new ways to do things. The mussel has given us insights that could be used to allow us to live healthier lives.”

The work was done by RIKEN researchers in collaboration with Professor Peibiao Zhang of the Changchun Institute of Applied Chemistry, Chinese Academy of Sciences, and Professor Yi Wang of the School of Pharmaceutical Sciences, Jilin University. The work was partially supported by the Japan Society for the Promotion of Science KAKENHI (Grant Number 15H01810 and 22220009), CAS-JSPS joint fund (GJHZ1519), and RIKEN MOST joint project.

Here’s a link to and a citation for the paper,

A Bioorthogonal Approach for the Preparation of a Titanium-Binding Insulin-like Growth-Factor-1 Derivative by using Tyrosinase by Chen Zhang, Hideyuki Miyatake, Yu Wang, Takehiko Inaba, Yi Wang, Peibiao Zhang, and Prof. Yoshihiro Ito. Angewandte Chemie International Edition DOI: 10.1002/anie.201603155 Version of Record online: 6 JUL 2016

© 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

This paper is behind a paywall.

Octopuses/octopi/octopodes and adhesive panels

Before launching into the science part of this news bit, here’s some grammar (from the Octopus Wikipedia entry; Note: Links have been removed),

The standard pluralized form of “octopus” in the English language is “octopuses” /ˈɒktəpʊsɪz/,[10] although the Ancient Greek plural “octopodes” /ɒkˈtɒpədiːz/, has also been used historically.[9] The alternative plural “octopi” — which misguidedly assumes it is a Latin “-us”-word — is considered grammatically incorrect.[11][12][13][14] It is nevertheless used enough to make it notable, and was formally acknowledged by the descriptivist Merriam-Webster 11th Collegiate Dictionary and Webster’s New World College Dictionary. The Oxford English Dictionary (2008 Draft Revision)[15] lists “octopuses”, “octopi”, and “octopodes”, in that order, labelling “octopodes” as rare and noting that “octopi” derives from the apprehension that octōpus comes from Latin.[16] In contrast, New Oxford American Dictionary (3rd Edition 2010) lists “octopuses” as the only acceptable pluralization, with a usage note indicating “octopodes” as being still occasionally used but “octopi” as being incorrect.[17]

Now the news. A July 12, 2016 news item on Nanowerk highlights some research into adhesives and octopuses,

With increased study of bio-adhesives, a significant effort has been made in search for novel adhesives that will combine reversibility, repeated usage, stronger bonds and faster bonding time, non-toxic, and more importantly be effective in wet and other extreme conditions.

A team of Korean scientists-made up of scientists from Korea Institute of Science and Technology (KIST) and UNIST has recently found a way to make building flexible pressure sensors easier–by mimicking the suction cups on octopus’s tentacles.

A July 5, 2016 UNIST (Ulsan National Institute of Science and Technology) press release, which originated the news item, provides more information,

According to the research team, “Although flexible pressure sensors might give future prosthetics and robots a better sense of touch, building them requires a lot of laborious transferring of nano- and microribbons of inorganic semiconductor materials onto polymer sheets.”

In search of an easier way to process this transfer printing, Prof. Hyunhyub Ko (School of Energy and Chemical Engineering, UNIST) and his colleagues turned to the octopus suction cups for inspiration.

An octopus uses its tentacles to move to a new location and uses suction cups underneath each tentacle to grab onto something. Each suction cup contains a cavity whose pressure is controlled by surrounding muscles. These can be made thinner or thicker on demand, increasing or decreasing air pressure inside the cup, allowing for sucking and releasing as desired.

By mimicking muscle actuation to control cavity-pressure-induced adhesion of octopus suckers, Prof. Ko and his team engineered octopus-inspired smart adhesive pads. They used the rubbery material polydimethylsiloxane (PDMS) to create an array of microscale suckers, which included pores that are coated with a thermally responsive polymer to create sucker-like walls.

The team discovered that the best way to replicate organic nature of muscle contractions would be through applied heat. Indeed, at room temperature, the walls of each pit sit in an ‘open’ state, but when the mat is heated to 32°C, the walls contract, creating suction, therby allowing the entire mate to adhere to a material (mimicking the suction function of an octopus). The adhesive strength also spiked from .32 kilopascals to 94 kilopascals at high temperature.

The team reports that the mat worked as envisioned—they made some indium gallium arsenide transistors that sat on a flexible substrate and also used it to move some nanomaterials to a different type of flexible material.

Prof. Ko and his team expect that their smart adhesive pads can be used as the substrate for wearable health sensors, such as Band-Aids or sensors that stick to the skin at normal body temperatures but fall off when rinsed under cold water.

Here’s a link to and a citation for the paper,

Octopus-Inspired Smart Adhesive Pads for Transfer Printing of Semiconducting Nanomembranes by Hochan Lee, Doo-Seung Um, Youngsu Lee, Seongdong Lim, Hyung-jun Kim,  and Hyunhyub Ko. Advanced Materials DOI: 10.1002/adma.201601407 Version of Record online: 20 JUN 2016

© 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

This paper is behind a paywall.