Tag Archives: machine learning

Of sleep, electric sheep, and thousands of artificial synapses on a chip

A close-up view of a new neuromorphic “brain-on-a-chip” that includes tens of thousands of memristors, or memory transistors. Credit: Peng Lin Courtesy: MIT

It’s hard to believe that a brain-on-a-chip might need sleep but that seems to be the case as far as the US Dept. of Energy’s Los Alamos National Laboratory is concerned. Before pursuing that line of thought, here’s some work from the Massachusetts Institute of Technology (MIT) involving memristors and a brain-on-a-chip. From a June 8, 2020 news item on ScienceDaily,

MIT engineers have designed a “brain-on-a-chip,” smaller than a piece of confetti, that is made from tens of thousands of artificial brain synapses known as memristors — silicon-based components that mimic the information-transmitting synapses in the human brain.

The researchers borrowed from principles of metallurgy to fabricate each memristor from alloys of silver and copper, along with silicon. When they ran the chip through several visual tasks, the chip was able to “remember” stored images and reproduce them many times over, in versions that were crisper and cleaner compared with existing memristor designs made with unalloyed elements.

Their results, published today in the journal Nature Nanotechnology, demonstrate a promising new memristor design for neuromorphic devices — electronics that are based on a new type of circuit that processes information in a way that mimics the brain’s neural architecture. Such brain-inspired circuits could be built into small, portable devices, and would carry out complex computational tasks that only today’s supercomputers can handle.

This ‘metallurgical’ approach differs somewhat from the protein nanowire approach used by the University of Massachusetts at Amherst team mentioned in my June 15, 2020 posting. Scientists are pursuing multiple pathways and we may find that we arrive with not ‘a single artificial brain but with many types of artificial brains.

A June 8, 2020 MIT news release (also on EurekAlert) provides more detail about this brain-on-a-chip,

“So far, artificial synapse networks exist as software. We’re trying to build real neural network hardware for portable artificial intelligence systems,” says Jeehwan Kim, associate professor of mechanical engineering at MIT. “Imagine connecting a neuromorphic device to a camera on your car, and having it recognize lights and objects and make a decision immediately, without having to connect to the internet. We hope to use energy-efficient memristors to do those tasks on-site, in real-time.”

Wandering ions

Memristors, or memory transistors [Note: Memristors are usually described as memory resistors; this is the first time I’ve seen ‘memory transistor’], are an essential element in neuromorphic computing. In a neuromorphic device, a memristor would serve as the transistor in a circuit, though its workings would more closely resemble a brain synapse — the junction between two neurons. The synapse receives signals from one neuron, in the form of ions, and sends a corresponding signal to the next neuron.

A transistor in a conventional circuit transmits information by switching between one of only two values, 0 and 1, and doing so only when the signal it receives, in the form of an electric current, is of a particular strength. In contrast, a memristor would work along a gradient, much like a synapse in the brain. The signal it produces would vary depending on the strength of the signal that it receives. This would enable a single memristor to have many values, and therefore carry out a far wider range of operations than binary transistors.

Like a brain synapse, a memristor would also be able to “remember” the value associated with a given current strength, and produce the exact same signal the next time it receives a similar current. This could ensure that the answer to a complex equation, or the visual classification of an object, is reliable — a feat that normally involves multiple transistors and capacitors.

Ultimately, scientists envision that memristors would require far less chip real estate than conventional transistors, enabling powerful, portable computing devices that do not rely on supercomputers, or even connections to the Internet.

Existing memristor designs, however, are limited in their performance. A single memristor is made of a positive and negative electrode, separated by a “switching medium,” or space between the electrodes. When a voltage is applied to one electrode, ions from that electrode flow through the medium, forming a “conduction channel” to the other electrode. The received ions make up the electrical signal that the memristor transmits through the circuit. The size of the ion channel (and the signal that the memristor ultimately produces) should be proportional to the strength of the stimulating voltage.

Kim says that existing memristor designs work pretty well in cases where voltage stimulates a large conduction channel, or a heavy flow of ions from one electrode to the other. But these designs are less reliable when memristors need to generate subtler signals, via thinner conduction channels.

The thinner a conduction channel, and the lighter the flow of ions from one electrode to the other, the harder it is for individual ions to stay together. Instead, they tend to wander from the group, disbanding within the medium. As a result, it’s difficult for the receiving electrode to reliably capture the same number of ions, and therefore transmit the same signal, when stimulated with a certain low range of current.

Borrowing from metallurgy

Kim and his colleagues found a way around this limitation by borrowing a technique from metallurgy, the science of melding metals into alloys and studying their combined properties.

“Traditionally, metallurgists try to add different atoms into a bulk matrix to strengthen materials, and we thought, why not tweak the atomic interactions in our memristor, and add some alloying element to control the movement of ions in our medium,” Kim says.

Engineers typically use silver as the material for a memristor’s positive electrode. Kim’s team looked through the literature to find an element that they could combine with silver to effectively hold silver ions together, while allowing them to flow quickly through to the other electrode.

The team landed on copper as the ideal alloying element, as it is able to bind both with silver, and with silicon.

“It acts as a sort of bridge, and stabilizes the silver-silicon interface,” Kim says.

To make memristors using their new alloy, the group first fabricated a negative electrode out of silicon, then made a positive electrode by depositing a slight amount of copper, followed by a layer of silver. They sandwiched the two electrodes around an amorphous silicon medium. In this way, they patterned a millimeter-square silicon chip with tens of thousands of memristors.

As a first test of the chip, they recreated a gray-scale image of the Captain America shield. They equated each pixel in the image to a corresponding memristor in the chip. They then modulated the conductance of each memristor that was relative in strength to the color in the corresponding pixel.

The chip produced the same crisp image of the shield, and was able to “remember” the image and reproduce it many times, compared with chips made of other materials.

The team also ran the chip through an image processing task, programming the memristors to alter an image, in this case of MIT’s Killian Court, in several specific ways, including sharpening and blurring the original image. Again, their design produced the reprogrammed images more reliably than existing memristor designs.

“We’re using artificial synapses to do real inference tests,” Kim says. “We would like to develop this technology further to have larger-scale arrays to do image recognition tasks. And some day, you might be able to carry around artificial brains to do these kinds of tasks, without connecting to supercomputers, the internet, or the cloud.”

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

Alloying conducting channels for reliable neuromorphic computing by Hanwool Yeon, Peng Lin, Chanyeol Choi, Scott H. Tan, Yongmo Park, Doyoon Lee, Jaeyong Lee, Feng Xu, Bin Gao, Huaqiang Wu, He Qian, Yifan Nie, Seyoung Kim & Jeehwan Kim. Nature Nanotechnology (2020 DOI: https://doi.org/10.1038/s41565-020-0694-5 Published: 08 June 2020

This paper is behind a paywall.

Electric sheep and sleeping androids

I find it impossible to mention that androids might need sleep without reference to Philip K. Dick’s 1968 novel, “Do Androids Dream of Electric Sheep?”; its Wikipedia entry is here.

June 8, 2020 Intelligent machines of the future may need to sleep as much as we do. Intelligent machines of the future may need to sleep as much as we do. Courtesy: Los Alamos National Laboratory

As it happens, I’m not the only one who felt the need to reference the novel, from a June 8, 2020 news item on ScienceDaily,

No one can say whether androids will dream of electric sheep, but they will almost certainly need periods of rest that offer benefits similar to those that sleep provides to living brains, according to new research from Los Alamos National Laboratory.

“We study spiking neural networks, which are systems that learn much as living brains do,” said Los Alamos National Laboratory computer scientist Yijing Watkins. “We were fascinated by the prospect of training a neuromorphic processor in a manner analogous to how humans and other biological systems learn from their environment during childhood development.”

Watkins and her research team found that the network simulations became unstable after continuous periods of unsupervised learning. When they exposed the networks to states that are analogous to the waves that living brains experience during sleep, stability was restored. “It was as though we were giving the neural networks the equivalent of a good night’s rest,” said Watkins.

A June 8, 2020 Los Alamos National Laboratory (LANL) news release (also on EurekAlert), which originated the news item, describes the research team’s presentation,

The discovery came about as the research team worked to develop neural networks that closely approximate how humans and other biological systems learn to see. The group initially struggled with stabilizing simulated neural networks undergoing unsupervised dictionary training, which involves classifying objects without having prior examples to compare them to.

“The issue of how to keep learning systems from becoming unstable really only arises when attempting to utilize biologically realistic, spiking neuromorphic processors or when trying to understand biology itself,” said Los Alamos computer scientist and study coauthor Garrett Kenyon. “The vast majority of machine learning, deep learning, and AI researchers never encounter this issue because in the very artificial systems they study they have the luxury of performing global mathematical operations that have the effect of regulating the overall dynamical gain of the system.”

The researchers characterize the decision to expose the networks to an artificial analog of sleep as nearly a last ditch effort to stabilize them. They experimented with various types of noise, roughly comparable to the static you might encounter between stations while tuning a radio. The best results came when they used waves of so-called Gaussian noise, which includes a wide range of frequencies and amplitudes. They hypothesize that the noise mimics the input received by biological neurons during slow-wave sleep. The results suggest that slow-wave sleep may act, in part, to ensure that cortical neurons maintain their stability and do not hallucinate.

The groups’ next goal is to implement their algorithm on Intel’s Loihi neuromorphic chip. They hope allowing Loihi to sleep from time to time will enable it to stably process information from a silicon retina camera in real time. If the findings confirm the need for sleep in artificial brains, we can probably expect the same to be true of androids and other intelligent machines that may come about in the future.

Watkins will be presenting the research at the Women in Computer Vision Workshop on June 14 [2020] in Seattle.

The 2020 Women in Computer Vition Workshop (WICV) website is here. As is becoming standard practice for these times, the workshop was held in a virtual environment. Here’s a link to and a citation for the poster presentation paper,

Using Sinusoidally-Modulated Noise as a Surrogate for Slow-Wave Sleep to
Accomplish Stable Unsupervised Dictionary Learning in a Spike-Based Sparse Coding Model
by Yijing Watkins, Edward Kim, Andrew Sornborger and Garrett T. Kenyon. Women in Computer Vision Workshop on June 14, 2020 in Seattle, Washington (state)

This paper is open access for now.

Brain-inspired electronics with organic memristors for wearable computing

I went down a rabbit hole while trying to figure out the difference between ‘organic’ memristors and standard memristors. I have put the results of my investigation at the end of this post. First, there’s the news.

An April 21, 2020 news item on ScienceDaily explains why researchers are so focused on memristors and brainlike computing,

The advent of artificial intelligence, machine learning and the internet of things is expected to change modern electronics and bring forth the fourth Industrial Revolution. The pressing question for many researchers is how to handle this technological revolution.

“It is important for us to understand that the computing platforms of today will not be able to sustain at-scale implementations of AI algorithms on massive datasets,” said Thirumalai Venkatesan, one of the authors of a paper published in Applied Physics Reviews, from AIP Publishing.

“Today’s computing is way too energy-intensive to handle big data. We need to rethink our approaches to computation on all levels: materials, devices and architecture that can enable ultralow energy computing.”

An April 21, 2020 American Institute of Physics (AIP) news release (also on EurekAlert), which originated the news item, describes the authors’ approach to the problems with organic memristors,

Brain-inspired electronics with organic memristors could offer a functionally promising and cost- effective platform, according to Venkatesan. Memristive devices are electronic devices with an inherent memory that are capable of both storing data and performing computation. Since memristors are functionally analogous to the operation of neurons, the computing units in the brain, they are optimal candidates for brain-inspired computing platforms.

Until now, oxides have been the leading candidate as the optimum material for memristors. Different material systems have been proposed but none have been successful so far.

“Over the last 20 years, there have been several attempts to come up with organic memristors, but none of those have shown any promise,” said Sreetosh Goswami, lead author on the paper. “The primary reason behind this failure is their lack of stability, reproducibility and ambiguity in mechanistic understanding. At a device level, we are now able to solve most of these problems,”

This new generation of organic memristors is developed based on metal azo complex devices, which are the brainchild of Sreebata Goswami, a professor at the Indian Association for the Cultivation of Science in Kolkata and another author on the paper.

“In thin films, the molecules are so robust and stable that these devices can eventually be the right choice for many wearable and implantable technologies or a body net, because these could be bendable and stretchable,” said Sreebata Goswami. A body net is a series of wireless sensors that stick to the skin and track health.

The next challenge will be to produce these organic memristors at scale, said Venkatesan.

“Now we are making individual devices in the laboratory. We need to make circuits for large-scale functional implementation of these devices.”

Caption: The device structure at a molecular level. The gold nanoparticles on the bottom electrode enhance the field enabling an ultra-low energy operation of the molecular device. Credit Sreetosh Goswami, Sreebrata Goswami and Thirumalai Venky Venkatesan

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

An organic approach to low energy memory and brain inspired electronics by Sreetosh Goswami, Sreebrata Goswami, and T. Venkatesan. Applied Physics Reviews 7, 021303 (2020) DOI: https://doi.org/10.1063/1.5124155

This paper is open access.

Basics about memristors and organic memristors

This undated article on Nanowerk provides a relatively complete and technical description of memristors in general (Note: A link has been removed),

A memristor (named as a portmanteau of memory and resistor) is a non-volatile electronic memory device that was first theorized by Leon Ong Chua in 1971 as the fourth fundamental two-terminal circuit element following the resistor, the capacitor, and the inductor (IEEE Transactions on Circuit Theory, “Memristor-The missing circuit element”).

Its special property is that its resistance can be programmed (resistor function) and subsequently remains stored (memory function). Unlike other memories that exist today in modern electronics, memristors are stable and remember their state even if the device loses power.

However, it was only almost 40 years later that the first practical device was fabricated. This was in 2008, when a group led by Stanley Williams at HP Research Labs realized that switching of the resistance between a conducting and less conducting state in metal-oxide thin-film devices was showing Leon Chua’s memristor behavior. …

The article on Nanowerk includes an embedded video presentation on memristors given by Stanley Williams (also known as R. Stanley Williams).

Mention of an ‘organic’memristor can be found in an October 31, 2017 article by Ryan Whitwam,

The memristor is composed of the transition metal ruthenium complexed with “azo-aromatic ligands.” [emphasis mine] The theoretical work enabling this material was performed at Yale, and the organic molecules were synthesized at the Indian Association for the Cultivation of Sciences. …

I highlighted ‘ligands’ because that appears to be the difference. However, there is more than one type of ligand on Wikipedia.

First, there’s the Ligand (biochemistry) entry (Note: Links have been removed),

In biochemistry and pharmacology, a ligand is a substance that forms a complex with a biomolecule to serve a biological purpose. …

Then, there’s the Ligand entry,

In coordination chemistry, a ligand[help 1] is an ion or molecule (functional group) that binds to a central metal atom to form a coordination complex …

Finally, there’s the Ligand (disambiguation) entry (Note: Links have been removed),

  • Ligand, an atom, ion, or functional group that donates one or more of its electrons through a coordinate covalent bond to one or more central atoms or ions
  • Ligand (biochemistry), a substance that binds to a protein
  • a ‘guest’ in host–guest chemistry

I did take a look at the paper and did not see any references to proteins or other biomolecules that I could recognize as such. I’m not sure why the researchers are describing their device as an ‘organic’ memristor but this may reflect a shortcoming in the definitions I have found or shortcomings in my reading of the paper rather than an error on their parts.

Hopefully, more research will be forthcoming and it will be possible to better understand the terminology.

Uncanny Valley: Being Human in the Age of AI (artificial intelligence) at the de Young museum (San Francisco, US) February 22 – October 25, 2020

So we’re still stuck in 20th century concepts about artificial intelligence (AI), eh? Sean Captain’s February 21, 2020 article (for Fast Company) about the new AI exhibit in San Francisco suggests that artists can help us revise our ideas (Note: Links have been removed),

Though we’re well into the age of machine learning, popular culture is stuck with a 20th century notion of artificial intelligence. While algorithms are shaping our lives in real ways—playing on our desires, insecurities, and suspicions in social media, for instance—Hollywood is still feeding us clichéd images of sexy, deadly robots in shows like Westworld and Star Trek Picard.

The old-school humanlike sentient robot “is an important trope that has defined the visual vocabulary around this human-machine relationship for a very long period of time,” says Claudia Schmuckli, curator of contemporary art and programming at the Fine Arts Museums of San Francisco. It’s also a naïve and outdated metaphor, one she is challenging with a new exhibition at San Francisco’s de Young Museum, called Uncanny Valley, that opens on February 22 [2020].

The show’s name [Uncanny Valley: Being Human in the Age of AI] is a kind of double entendre referencing both the dated and emerging conceptions of AI. Coined in the 1970s, the term “uncanny valley” describes the rise and then sudden drop off of empathy we feel toward a machine as its resemblance to a human increases. Putting a set of cartoony eyes on a robot may make it endearing. But fitting it with anatomically accurate eyes, lips, and facial gestures gets creepy. As the gap between the synthetic and organic narrows, the inability to completely close that gap becomes all the more unsettling.

But the artists in this exhibit are also looking to another valley—Silicon Valley, and the uncanny nature of the real AI the region is building. “One of the positions of this exhibition is that it may be time to rethink the coordinates of the Uncanny Valley and propose a different visual vocabulary,” says Schmuckli.

Artist Stephanie Dinkins faces off with robot Bina48, a bot on display at the de Young Museum’s Uncanny Valley show. [Photo: courtesy of the artist; courtesy of the Fine Arts Museums of San Francisco]

From Captain’s February 21, 2020 article,

… the resemblance to humans is only synthetic-skin deep. Bina48 can string together a long series of sentences in response to provocative questions from Dinkins, such as, “Do you know racism?” But the answers are sometimes barely intelligible, or at least lack the depth and nuance of a conversation with a real human. The robot’s jerky attempts at humanlike motion also stand in stark contrast to Dinkins’s calm bearing and fluid movement. Advanced as she is by today’s standards, Bina48 is tragically far from the sci-fi concept of artificial life. Her glaring shortcomings hammer home why the humanoid metaphor is not the right framework for understanding at least today’s level of artificial intelligence.

For anybody who has more curiosity about the ‘uncanny valley’, there’s this Wikipedia entry.

For more details about the’ Uncanny Valley: Being Human in the Age of AI’ exhibition there’s this September 26, 2019 de Young museum news release,

What are the invisible mechanisms of current forms of artificial intelligence (AI)? How is AI impacting our personal lives and socioeconomic spheres? How do we define intelligence? How do we envision the future of humanity?

SAN FRANCISCO (September 26, 2019) — As technological innovation continues to shape our identities and societies, the question of what it means to be, or remain human has become the subject of fervent debate. Taking advantage of the de Young museum’s proximity to Silicon Valley, Uncanny Valley: Being Human in the Age of AI arrives as the first major exhibition in the US to explore the relationship between humans and intelligent machines through an artistic lens. Organized by the Fine Arts Museums of San Francisco, with San Francisco as its sole venue, Uncanny Valley: Being Human in the Age of AI will be on view from February 22 to October 25, 2020.

“Technology is changing our world, with artificial intelligence both a new frontier of possibility but also a development fraught with anxiety,” says Thomas P. Campbell, Director and CEO of the Fine Arts Museums of San Francisco. “Uncanny Valley: Being Human in the Age of AI brings artistic exploration of this tension to the ground zero of emerging technology, raising challenging questions about the future interface of human and machine.”

The exhibition, which extends through the first floor of the de Young and into the museum’s sculpture garden, explores the current juncture through philosophical, political, and poetic questions and problems raised by AI. New and recent works by an intergenerational, international group of artists and activist collectives—including Zach Blas, Ian Cheng, Simon Denny, Stephanie Dinkins, Forensic Architecture, Lynn Hershman Leeson, Pierre Huyghe, Christopher Kulendran Thomas in collaboration with Annika Kuhlmann, Agnieszka Kurant, Lawrence Lek, Trevor Paglen, Hito Steyerl, Martine Syms, and the Zairja Collective—will be presented.

The Uncanny Valley

In 1970 Japanese engineer Masahiro Mori introduced the concept of the “uncanny valley” as a terrain of existential uncertainty that humans experience when confronted with autonomous machines that mimic their physical and mental properties. An enduring metaphor for the uneasy relationship between human beings and lifelike robots or thinking machines, the uncanny valley and its edges have captured the popular imagination ever since. Over time, the rapid growth and affordability of computers, cloud infrastructure, online search engines, and data sets have fueled developments in machine learning that fundamentally alter our modes of existence, giving rise to a newly expanded uncanny valley.

“As our lives are increasingly organized and shaped by algorithms that track, collect, evaluate, and monetize our data, the uncanny valley has grown to encompass the invisible mechanisms of behavioral engineering and automation,” says Claudia Schmuckli, Curator in Charge of Contemporary Art and Programming at the Fine Arts Museums of San Francisco. “By paying close attention to the imminent and nuanced realities of AI’s possibilities and pitfalls, the artists in the exhibition seek to thicken the discourse around AI. Although fables like HBO’s sci-fi drama Westworld, or Spike Jonze’s feature film Her still populate the collective imagination with dystopian visions of a mechanized future, the artists in this exhibition treat such fictions as relics of a humanist tradition that has little relevance today.”

In Detail

Ian Cheng’s digitally simulated AI creature BOB (Bag of Beliefs) reflects on the interdependency of carbon and silicon forms of intelligence. An algorithmic Tamagotchi, it is capable of evolution, but its growth, behavior, and personality are molded by online interaction with visitors who assume collective responsibility for its wellbeing.

In A.A.I. (artificial artificial intelligence), an installation of multiple termite mounds of colored sand, gold, glitter and crystals, Agnieszka Kurant offers a vibrant critique of new AI economies, with their online crowdsourcing marketplace platforms employing invisible armies of human labor at sub-minimum wages.

Simon Denny ‘s Amazon worker cage patent drawing as virtual King Island Brown Thornbill cage (US 9,280,157 B2: “System and method for transporting personnel within an active workspace”, 2016) (2019) also examines the intersection of labor, resources, and automation. He presents 3-D prints and a cage-like sculpture based on an unrealized machine patent filed by Amazon to contain human workers. Inside the cage an augmented reality application triggers the appearance of a King Island Brown Thornbill — a bird on the verge of extinction; casting human labor as the proverbial canary in the mine. The humanitarian and ecological costs of today’s data economy also informs a group of works by the Zairja Collective that reflect on the extractive dynamics of algorithmic data mining. 

Hito Steyerl addresses the political risks of introducing machine learning into the social sphere. Her installation The City of Broken Windows presents a collision between commercial applications of AI in urban planning along with communal and artistic acts of resistance against neighborhood tipping: one of its short films depicts a group of technicians purposefully smashing windows to teach an algorithm how to recognize the sound of breaking glass, and another follows a group of activists through a Camden, NJ neighborhood as they work to keep decay at bay by replacing broken windows in abandoned homes with paintings. 

Addressing the perpetuation of societal biases and discrimination within AI, Trevor Paglen’s They Took the Faces from the Accused and the Dead…(SD18), presents a large gridded installation of more than three thousand mugshots from the archives of the American National Standards Institute. The institute’s collections of such images were used to train ealry facial-recognition technologies — without the consent of those pictured. Lynn Hershman Leeson’s new installation Shadow Stalker critiques the problematic reliance on algorithmic systems, such as the military forecasting tool Predpol now widely used for policing, that categorize individuals into preexisting and often false “embodied metrics.”

Stephanie Dinkins extends the inquiry into how value systems are built into AI and the construction of identity in Conversations with Bina48, examining the social robot’s (and by extension our society’s) coding of technology, race, gender and social equity. In the same territory, Martine Syms posits AI as a “shamespace” for misrepresentation. For Mythiccbeing she has created an avatar of herself that viewers can interact with through text messaging. But unlike service agents such as Siri and Alexa, who readily respond to questions and demands, Syms’s Teeny is a contrarious interlocutor, turning each interaction into an opportunity to voice personal observations and frustrations about racial inequality and social injustice.

Countering the abusive potential of machine learning, Forensic Architecture pioneers an application to the pursuit of social justice. Their proposition of a Model Zoo marks the beginnings of a new research tool for civil society built of military vehicles, missile fragments, and bomb clouds—evidence of human-rights violations by states and militaries around the world. Christopher Kulendran Thomas’s video Being Human, created in collaboration with Annika Kuhlmann, poses the philosophical question of what it means to be human when machines are able to synthesize human understanding ever more convincingly. Set  in Sri Lanka, it employs AI-generated characters of singer Taylor Swift and artist Oscar Murillo to reflect on issues of individual authenticity, collective sovereignty, and the future of human rights.

Lawrence Lek’s sci-fi-inflected film Aidol, which explores the relationship between algorithmic automation and human creativity, projects this question into the future. It transports the viewer into the computer-generated “sinofuturist” world of the 2065 eSports Olympics: when the popular singer Diva enlists the super-intelligent Geomancer to help her stage her artistic comeback during the game’s halftime show, she unleashes an existential and philosophical battle that explodes the divide between humans and machines.

The Doors, a newly commissioned installation by Zach Blas, by contrast shines the spotlight back onto the present and on the culture and ethos of Silicon Valley — the ground zero for the development of AI. Inspired by the ubiquity of enclosed gardens on tech campuses, he has created an artificial garden framed by a six-channel video projected on glass panes that convey a sense of algorithmic psychedelia aiming to open new “doors of perception.” While luring visitors into AI’s promises, it also asks what might become possible when such glass doors begin to crack. 

Unveiled in late spring Pierre Huyghe‘s Exomind (Deep Water), a sculpture of a crouched female nude with a live beehive as its head will be nestled within the museum’s garden. With its buzzing colony pollinating the surrounding flora, it offers a poignant metaphor for the modeling of neural networks on the biological brain and an understanding of intelligence as grounded in natural forms and processes.

The Uncanny Valley: Being Human in the Age of AI event page features a link to something unexpected 9scroll down about 40% of the way), a Statement on Eyal Weizman of Forensic Architecture,

On Thursday, February 13 [2020], Eyal Weizman of Forensic Architecture had his travel authorization to the United States revoked due to an “algorithm” that identified him as a security threat.

He was meant to be in the United States promoting multiple exhibitions including Uncanny Valley: Being Human in the Age of AI, opening on February 22 [2020] at the de Young museum in San Francisco.

Since 2018, Forensic Architecture has used machine learning / AI to aid in humanitarian work, using synthetic images—photorealistic digital renderings based around 3-D models—to train algorithmic classifiers to identify tear gas munitions and chemical bombs deployed against protesters worldwide, including in Hong Kong, Chile, the US, Venezuela, and Sudan.

Their project, Model Zoo, on view in Uncanny Valley represents a growing collection of munitions and weapons used in conflict today and the algorithmic models developed to identify them. It shows a collection of models being used to track and hold accountable human rights violators around the world. The piece joins work by 14 contemporary artists reflecting on the philosophical and political consequences of the application of AI into the social sphere.

We are deeply saddened that Weizman will not be allowed to travel to celebrate the opening of the exhibition. We stand with him and Forensic Architecture’s partner communities who continue to resist violent states and corporate practices, and who are increasingly exposed to the regime of “security algorithms.”

—Claudia Schmuckli, Curator-in-Charge, Contemporary Art & Programming, & Thomas P. Campbell, Director and CEO, Fine Arts Museums of San Francisco

There is a February 20, 2020 article (for Fast Company) by Eyal Weizman chronicling his experience with being denied entry by an algorithm. Do read it in its entirety (the Fast Company is itself an excerpt from Weizman’s essay) if you have the time, if not, here’s the description of how he tried to gain entry after being denied the first time,

The following day I went to the U.S. Embassy in London to apply for a visa. In my interview, the officer informed me that my authorization to travel had been revoked because the “algorithm” had identified a security threat. He said he did not know what had triggered the algorithm but suggested that it could be something I was involved in, people I am or was in contact with, places to which I had traveled (had I recently been in Syria, Iran, Iraq, Yemen, or Somalia or met their nationals?), hotels at which I stayed, or a certain pattern of relations among these things. I was asked to supply the Embassy with additional information, including 15 years of travel history, in particular where I had gone and who had paid for it. The officer said that Homeland Security’s investigators could assess my case more promptly if I supplied the names of anyone in my network whom I believed might have triggered the algorithm. I declined to provide this information.

I hope the exhibition is successful; it has certainly experienced a thought-provoking start.

Finally, I have often featured postings that discuss the ‘uncanny valley’. To find those postings, just use that phrase in the blog search engine. You might also went to search ‘Hiroshi Ishiguro’, a Japanese scientist and robotocist who specializes in humanoid robots.

Memristor-based neural network and the biosimilar principle of learning

Once you get past the technical language (there’s a lot of it), you’ll find that they make the link between biomimicry and memristors explicit. Admittedly I’m not an expert but if I understand the research correctly, the scientists are suggesting that the algorithms used in machine learning today cannot allow memristors to be properly integrated for use in true neuromorphic computing and this work from Russia and Greece points to a new paradigm. If you understand it differently, please do let me know in the comments.

A July 12, 2019 news item on Nanowerk kicks things off (Note: A link has been removed),

Lobachevsky University scientists together with their colleagues from the National Research Center “Kurchatov Institute” (Moscow) and the National Research Center “Demokritos” (Athens) are working on the hardware implementation of a spiking neural network based on memristors.

The key elements of such a network, along with pulsed neurons, are artificial synaptic connections that can change the strength (weight) of connection between neurons during the learning (Microelectronic Engineering, “Yttria-stabilized zirconia cross-point memristive devices for neuromorphic applications”).

For this purpose, memristive devices based on metal-oxide-metal nanostructures developed at the UNN Physics and Technology Research Institute (PTRI) are suitable, but their use in specific spiking neural network architectures developed at the Kurchatov Institute requires demonstration of biologically plausible learning principles.

Caption: Cross-section image of the metal-oxide-metal memristive structure based on ZrO2(Y) polycrystalline film (a); corresponding schematic view of the cross-point memristive device (b); STDP dependencies of memristive device conductance changes for different delay values between pre- and postsynaptic neuron spikes (c); photographs of a microchip and an array of memristive devices in a standard cermet casing (d); the simplest spiking neural network architecture learning on the basis of local rules for changing memristive weights (e). Credit: Lobachevsky University

A July 12, 2019 (?) Lobachevsky University press release (also on EurekAlert), which originated the news item, delves further into the work,

The biological mechanism of learning of neural systems is described by Hebb’s rule, according to which learning occurs as a result of an increase in the strength of connection  (synaptic weight) between simultaneously active neurons, which indicates the presence of a causal relationship in their excitation. One of the clarifying forms of this fundamental rule is plasticity, which depends on the time of arrival of pulses (Spike-Timing Dependent Plasticity – STDP).

In accordance with STDP, synaptic weight increases if the postsynaptic neuron generates a pulse (spike) immediately after the presynaptic one, and vice versa, the synaptic weight decreases if the postsynaptic neuron generates a spike right before the presynaptic one. Moreover, the smaller the time difference Δt between the pre- and postsynaptic spikes, the more pronounced the weight change will be.

According to one of the researchers, Head of the UNN PTRI laboratory Alexei Mikhailov, in order to demonstrate the STDP principle, memristive nanostructures based on yttria-stabilized zirconia (YSZ) thin films were used. YSZ is a well-known solid-state electrolyte with high oxygen ion mobility.

“Due to a specified concentration of oxygen vacancies, which is determined by the controlled concentration of yttrium impurities, and the heterogeneous structure of the films obtained by magnetron sputtering, such memristive structures demonstrate controlled bipolar switching between different resistive states in a wide resistance range. The switching is associated with the formation and destruction of conductive channels along grain boundaries in the polycrystalline ZrO2 (Y) film,” notes Alexei Mikhailov.

An array of memristive devices for research was implemented in the form of a microchip mounted in a standard cermet casing, which facilitates the integration of the array into a neural network’s analog circuit. The full technological cycle for creating memristive microchips is currently implemented at the UNN PTRI. In the future, it is possible to scale the devices down to the minimum size of about 50 nm, as was established by Greek partners.
Our studies of the dynamic plasticity of the memoristive devices, continues Alexey Mikhailov, have shown that the form of the conductance change depending on Δt is in good agreement with the STDP learning rules. It should be also noted that if the initial value of the memristor conductance is close to the maximum, it is easy to reduce the corresponding weight while it is difficult to enhance it, and in the case of a memristor with a minimum conductance in the initial state, it is difficult to reduce its weight, but it is easy to enhance it.

According to Vyacheslav Demin, director-coordinator in the area of nature-like technologies of the Kurchatov Institute, who is one of the ideologues of this work, the established pattern of change in the memristor conductance clearly demonstrates the possibility of hardware implementation of the so-called local learning rules. Such rules for changing the strength of synaptic connections depend only on the values ​​of variables that are present locally at each time point (neuron activities and current weights).

“This essentially distinguishes such principle from the traditional learning algorithm, which is based on global rules for changing weights, using information on the error values ​​at the current time point for each neuron of the output neural network layer (in a widely popular group of error back propagation methods). The traditional principle is not biosimilar, it requires “external” (expert) knowledge of the correct answers for each example presented to the network (that is, they do not have the property of self-learning). This principle is difficult to implement on the basis of memristors, since it requires controlled precise changes of memristor conductances, as opposed to local rules. Such precise control is not always possible due to the natural variability (a wide range of parameters) of memristors as analog elements,” says Vyacheslav Demin.

Local learning rules of the STDP type implemented in hardware on memristors provide the basis for autonomous (“unsupervised”) learning of a spiking neural network. In this case, the final state of the network does not depend on its initial state, but depends only on the learning conditions (a specific sequence of pulses). According to Vyacheslav Demin, this opens up prospects for the application of local learning rules based on memristors when solving artificial intelligence problems with the use of complex spiking neural network architectures.

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

Yttria-stabilized zirconia cross-point memristive devices for neuromorphic applications by A. V. Emelyanov, K. E. Nikiruy, A. Demin, V. V. Rylkov, A. I. Belov, D. S. Korolev, E. G. Gryaznov, D. A. Pavlov, O. N. Gorshkov, A. N. Mikhaylov, P. Dimitrakis. Microelectronic Engineering Volume 215, 15 July 2019, 110988 First available online 16 May 2019

This paper is behind a paywall.

Large Interactive Virtual Environment Laboratory (LIVELab) located in McMaster University’s Institute for Music & the Mind (MIMM) and the MetaCreation Lab at Simon Fraser University

Both of these bits have a music focus but they represent two entirely different science-based approaches to that form of art and one is solely about the music and the other is included as one of the art-making processes being investigated..

Large Interactive Virtual Environment Laboratory (LIVELab) at McMaster University

Laurel Trainor and Dan J. Bosnyak both of McMaster University (Ontario, Canada) have written an October 27, 2019 essay about the LiveLab and their work for The Conversation website (Note: Links have been removed),

The Large Interactive Virtual Environment Laboratory (LIVELab) at McMaster University is a research concert hall. It functions as both a high-tech laboratory and theatre, opening up tremendous opportunities for research and investigation.

As the only facility of its kind in the world, the LIVELab is a 106-seat concert hall equipped with dozens of microphones, speakers and sensors to measure brain responses, physiological responses such as heart rate, breathing rates, perspiration and movements in multiple musicians and audience members at the same time.

Engineers, psychologists and clinician-researchers from many disciplines work alongside musicians, media artists and industry to study performance, perception, neural processing and human interaction.

In the LIVELab, acoustics are digitally controlled so the experience can change instantly from extremely silent with almost no reverberation to a noisy restaurant to a subway platform or to the acoustics of Carnegie Hall.

Real-time physiological data such as heart rate can be synchronized with data from other systems such as motion capture, and monitored and recorded from both performers and audience members. The result is that the reams of data that can now be collected in a few hours in the LIVELab used to take weeks or months to collect in a traditional lab. And having measurements of multiple people simultaneously is pushing forward our understanding of real-time human interactions.

Consider the implications of how music might help people with Parkinson’s disease to walk more smoothly or children with dyslexia to read better.

[…] area of ongoing research is the effectiveness of hearing aids. By the age of 60, nearly 49 per cent of people will suffer from some hearing loss. People who wear hearing aids are often frustrated when listening to music because the hearing aids distort the sound and cannot deal with the dynamic range of the music.

The LIVELab is working with the Hamilton Philharmonic Orchestra to solve this problem. During a recent concert, researchers evaluated new ways of delivering sound directly to participants’ hearing aids to enhance sounds.

Researchers hope new technologies can not only increase live musical enjoyment but alleviate the social isolation caused by hearing loss.

Imagine the possibilities for understanding music and sound: How it might help to improve cognitive decline, manage social performance anxiety, help children with developmental disorders, aid in treatment of depression or keep the mind focused. Every time we conceive and design a study, we think of new possibilities.

The essay also includes an embedded 12 min. video about LIVELab and details about studies conducted on musicians and live audiences. Apparently, audiences experience live performance differently than recorded performances and musicians use body sway to create cohesive performances. You can find the McMaster Institute for Music & the Mind here and McMaster’s LIVELab here.

Capturing the motions of a string quartet performance. Laurel Trainor, Author provided [McMaster University]

Metacreation Lab at Simon Fraser University (SFU)

I just recently discovered that there’s a Metacreation Lab at Simon Fraser University (Vancouver, Canada), which on its homepage has this ” Metacreation is the idea of endowing machines with creative behavior.” Here’s more from the homepage,

As the contemporary approach to generative art, Metacreation involves using tools and techniques from artificial intelligence, artificial life, and machine learning to develop software that partially or completely automates creative tasks. Through the collaboration between scientists, experts in artificial intelligence, cognitive sciences, designers and artists, the Metacreation Lab for Creative AI is at the forefront of the development of generative systems, be they embedded in interactive experiences or integrated into current creative software. Scientific research in the Metacreation Lab explores how various creative tasks can be automated and enriched. These tasks include music composition [emphasis mine], sound design, video editing, audio/visual effect generation, 3D animation, choreography, and video game design.

Besides scientific research, the team designs interactive and generative artworks that build upon the algorithms and research developed in the Lab. This work often challenges the social and cultural discourse on AI.

Much to my surprise I received the Metacreation Lab’s inaugural email newsletter (received via email on Friday, November 15, 2019),

Greetings,

We decided to start a mailing list for disseminating news, updates, and announcements regarding generative art, creative AI and New Media. In this newsletter: 

  1. ISEA 2020: The International Symposium on Electronic Art. ISEA return to Montreal, check the CFP bellow and contribute!
  2. ISEA 2015: A transcription of Sara Diamond’s keynote address “Action Agenda: Vancouver’s Prescient Media Arts” is now available for download. 
  3. Brain Art, the book: we are happy to announce the release of the first comprehensive volume on Brain Art. Edited by Anton Nijholt, and published by Springer.

Here are more details from the newsletter,

ISEA2020 – 26th International Symposium on Electronic Arts

Montreal, September 24, 2019
Montreal Digital Spring (Printemps numérique) is launching a call for participation as part of ISEA2020 / MTL connect to be held from May 19 to 24, 2020 in Montreal, Canada. Founded in 1990, ISEA is one of the world’s most prominent international arts and technology events, bringing together scholarly, artistic, and scientific domains in an interdisciplinary discussion and showcase of creative productions applying new technologies in art, interactivity, and electronic and digital media. For 2020, ISEA Montreal turns towards the theme of sentience.

ISEA2020 will be fully dedicated to examining the resurgence of sentience—feeling-sensing-making sense—in recent art and design, media studies, science and technology studies, philosophy, anthropology, history of science and the natural scientific realm—notably biology, neuroscience and computing. We ask: why sentience? Why and how does sentience matter? Why have artists and scholars become interested in sensing and feeling beyond, with and around our strictly human bodies and selves? Why has this notion been brought to the fore in an array of disciplines in the 21st century?
CALL FOR PARTICIPATION: WHY SENTIENCE? ISEA2020 invites artists, designers, scholars, researchers, innovators and creators to participate in the various activities deployed from May 19 to 24, 2020. To complete an application, please fill in the forms and follow the instructions.

The final submissions deadline is NOVEMBER 25, 2019. Submit your application for WORKSHOP and TUTORIAL Submit your application for ARTISTIC WORK Submit your application for FULL / SHORT PAPER Submit your application for PANEL Submit your application for POSTER Submit your application for ARTIST TALK Submit your application for INSTITUTIONAL PRESENTATION
Find Out More
You can apply for several categories. All profiles are welcome. Notifications of acceptance will be sent around January 13, 2020.

Important: please note that the Call for participation for MTL connect is not yet launched, but you can also apply to participate in the programming of the other Pavilions (4 other themes) when registrations are open (coming soon): mtlconnecte.ca/en TICKETS

Registration is now available to assist to ISEA2020 / MTL connect, from May 19 to 24, 2020. Book today your Full Pass and get the early-bird rate!
Buy Now

More from the newsletter,

ISEA 2015 was in Vancouver, Canada, and the proceedings and art catalog are still online. The news is that Sara Diamond released her 2015 keynote address as a paper: Action Agenda: Vancouver’s Prescient Media Arts. It is never too late so we thought we would let you know about this great read. See The 2015 Proceedings Here

The last item from the inaugural newsletter,

The first book that surveys how brain activity can be monitored and manipulated for artistic purposes, with contributions by interactive media artists, brain-computer interface researchers, and neuroscientists. View the Book Here

As per the Leonardo review from Cristina Albu:

“Another seminal contribution of the volume is the presentation of multiple taxonomies of “brain art,” which can help art critics develop better criteria for assessing this genre. Mirjana Prpa and Philippe Pasquier’s meticulous classification shows how diverse such works have become as artists consider a whole range of variables of neurofeedback.” Read the Review

For anyone not familiar with the ‘Leonardo’ cited in the above, it’s Leonardo; the International Society for the Arts, Sciences and Technology.

Should this kind of information excite and motivate you do start metacreating, you can get in touch with the lab,

Our mailing address is:
Metacreation Lab for Creative AI
School of Interactive Arts & Technology
Simon Fraser University
250-13450 102 Ave.
Surrey, BC V3T 0A3
Web: http://metacreation.net/
Email: metacreation_admin (at) sfu (dot) ca

The glorious glasswing butterfly and superomniphobic glass

This is not the first time the glasswing butterfly has inspired some new technology. Lat time, it was an eye implant,

The clear wings make this South-American butterfly hard to see in flight, a succesfull defense mechanism. Credit: Eddy Van 3000 from in Flanders fields – B – United Tribes ov Europe – the wings-become-windows butterfly. [downloaded from https://commons.wikimedia.org/wiki/Category:Greta_oto#/media/File:South-American_butterfly.jpg]

You’ll find that image and more in my May 22, 2018 posting about the eye implant. Don’t miss scrolling down to the video which features the butterfly fluttering its wings in the first few seconds.

Getting back to the glasswing butterfly’s latest act of inspiration a July 11, 2019 news item on ScienceDaily announces the work,

Glass for technologies like displays, tablets, laptops, smartphones, and solar cells need to pass light through, but could benefit from a surface that repels water, dirt, oil, and other liquids. Researchers from the University of Pittsburgh’s Swanson School of Engineering have created a nanostructure glass that takes inspiration from the wings of the glasswing butterfly to create a new type of glass that is not only very clear across a wide variety of wavelengths and angles, but is also antifogging.

A July 11, 2019 University of Pittsburgh news release (also on EurekAlert), which originated the news item, provides more technical detail about the new glass,

The nanostructured glass has random nanostructures, like the glasswing butterfly wing, that are smaller than the wavelengths of visible light. This allows the glass to have a very high transparency of 99.5% when the random nanostructures are on both sides of the glass. This high transparency can reduce the brightness and power demands on displays that could, for example, extend battery life. The glass is antireflective across higher angles, improving viewing angles. The glass also has low haze, less than 0.1%, which results in very clear images and text.

“The glass is superomniphobic, meaning it repels a wide variety of liquids such as orange juice, coffee, water, blood, and milk,” explains Sajad Haghanifar, lead author of the paper and doctoral candidate in industrial engineering at Pitt. “The glass is also anti-fogging, as water condensation tends to easily roll off the surface, and the view through the glass remains unobstructed. Finally, the nanostructured glass is durable from abrasion due to its self-healing properties–abrading the surface with a rough sponge damages the coating, but heating it restores it to its original function.”

Natural surfaces like lotus leaves, moth eyes and butterfly wings display omniphobic properties that make them self-cleaning, bacterial-resistant and water-repellant–adaptations for survival that evolved over millions of years. Researchers have long sought inspiration from nature to replicate these properties in a synthetic material, and even to improve upon them. While the team could not rely on evolution to achieve these results, they instead utilized machine learning.

“Something significant about the nanostructured glass research, in particular, is that we partnered with SigOpt to use machine learning to reach our final product,” says Paul Leu, PhD, associate professor of industrial engineering, whose lab conducted the research. Dr. Leu holds secondary appointments in mechanical engineering and materials science and chemical engineering. “When you create something like this, you don’t start with a lot of data, and each trial takes a great deal of time. We used machine learning to suggest variables to change, and it took us fewer tries to create this material as a result.”

“Bayesian optimization and active search are the ideal tools to explore the balance between transparency and omniphobicity efficiently, that is, without needing thousands of fabrications, requiring hundreds of days.” said Michael McCourt, PhD, research engineer at SigOpt. Bolong Cheng, PhD, fellow research engineer at SigOpt, added, “Machine learning and AI strategies are only relevant when they solve real problems; we are excited to be able to collaborate with the University of Pittsburgh to bring the power of Bayesian active learning to a new application.”

Here’s an image illustrating the work from the researchers,

Courtesy: University of Pittsburgh

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

Creating glasswing butterfly-inspired durable antifogging superomniphobic supertransmissive, superclear nanostructured glass through Bayesian learning and optimization by Sajad Haghanifar, Michael McCourt, Bolong Cheng, Jeffrey Wuenschell, Paul Ohodnickic, and Paul W. Leu. Mater. Horiz., 2019, Advance Article DOI: 10.1039/C9MH00589G first published on 10 Jun 2019

This paper is behind a paywall. One more thing, here’s SigOpt, the company the scientists partnered.

Data science guide from Sense about Science

Sense about Science, headquartered in the UK, is in its own words (from its homepage)

Sense about Science is an independent campaigning charity that challenges the misrepresentation of science and evidence in public life. …

According to an October 1, 2019 announcement from Sense about Science (received via email), the organization has published a new guide,

Our director warned yesterday [September 30, 2019] that data science is being given a free
pass on quality in too many arenas. From flood predictions to mortgage offers to the prediction of housing needs, we are not asking enough about whether AI solutions and algorithms can bear the weight we want to put on them.

It was the UK launch of our ‘Data Science: a guide for society’ at the Institute of Physics, where we invited representatives from different sectors to take up the challenge of creating a more questioning culture. Tracey Brown said the situation was like medicine 50 years ago: it seems that some people have become too clever to explain and the rest of us are feeling too dumb to ask.

At the end of the event we had a lot of proposals for how to make different communities aware of the guide’s three fundamental questions from the people who attended. There are many hundreds of people among our friends who could do something along these lines:

     * Publicise the guide
     * Incorporate it into your own work
     * Send it to people who are involved in procurement, licensing or
reporting or decision making at community, national and international
levels
     * Undertake a project with us to equip particular groups such as
parliamentary advisers, journalists and small charities.

Would you take a look at the guide [1] here and tell me if there’s something you can do? (alex@senseaboutscience.org)

There are launches planned in other countries over the rest of this year and into 2020. We are drawing up a map of offers to reach different communities. I’ll share all your suggestions with my colleague Errin Riley at the end of this week and we will get back to you quickly.

Before linking you to the guide, here’s a brief description from the Patterns in Data webpage,

In recent years, phrases like ‘big data’, ‘machine learning’, ‘algorithms’ and ‘pattern recognition’ have started slipping into everyday discussion. We’ve worked with researchers and experts to generate an open and informed public discussion on patterns in data across a wide range of projects.

Data Science: A guide for society

According to the headlines, we’re in the middle of a ‘data revolution: large, detailed datasets and complex algorithms allow us to make predictions on anything from who will win the league to who is likely to commit a crime. Our ability to question the quality of evidence – as the public, journalists, politicians or decision makers – needs to be expanded to meet this. To know the questions to ask and how to press for clarity about the strengths and weaknesses of using analysis from data models to make decisions. This is a guide to having more of those conversations, regardless of how much you don’t know about data science.

Here’s Data Science: A Guide for Society.

AI (artificial intelligence) artist got a show at a New York City art gallery

AI artists first hit my radar in August 2018 when Christie’s Auction House advertised an art auction of a ‘painting’ by an algorithm (artificial intelligence). There’s more in my August 31, 2018 posting but, briefly, a French art collective, Obvious, submitted a painting, “Portrait of Edmond de Belamy,” that was created by an artificial intelligence agent to be sold for an estimated to $7000 – $10,000. They weren’t even close. According to Ian Bogost’s March 6, 2019 article for The Atlantic, the painting sold for $432,500 In October 2018.

It has also, Bogost notes in his article, occasioned an art show (Note: Links have been removed),

… part of “Faceless Portraits Transcending Time,” an exhibition of prints recently shown [Februay 13 – March 5, 2019] at the HG Contemporary gallery in Chelsea, the epicenter of New York’s contemporary-art world. All of them were created by a computer.

The catalog calls the show a “collaboration between an artificial intelligence named AICAN and its creator, Dr. Ahmed Elgammal,” a move meant to spotlight, and anthropomorphize, the machine-learning algorithm that did most of the work. According to HG Contemporary, it’s the first solo gallery exhibit devoted to an AI artist.

If they hadn’t found each other in the New York art scene, the players involved could have met on a Spike Jonze film set: a computer scientist commanding five-figure print sales from software that generates inkjet-printed images; a former hotel-chain financial analyst turned Chelsea techno-gallerist with apparent ties to fine-arts nobility; a venture capitalist with two doctoral degrees in biomedical informatics; and an art consultant who put the whole thing together, A-Team–style, after a chance encounter at a blockchain conference. Together, they hope to reinvent visual art, or at least to cash in on machine-learning hype along the way.

The show in New York City, “Faceless Portraits …,” exhibited work by an artificially intelligent artist-agent (I’m creating a new term to suit my purposes) that’s different than the one used by Obvious to create “Portrait of Edmond de Belamy,” As noted earlier, it sold for a lot of money (Note: Links have been removed),

Bystanders in and out of the art world were shocked. The print had never been shown in galleries or exhibitions before coming to market at auction, a channel usually reserved for established work. The winning bid was made anonymously by telephone, raising some eyebrows; art auctions can invite price manipulation. It was created by a computer program that generates new images based on patterns in a body of existing work, whose features the AI “learns.” What’s more, the artists who trained and generated the work, the French collective Obvious, hadn’t even written the algorithm or the training set. They just downloaded them, made some tweaks, and sent the results to market.

“We are the people who decided to do this,” the Obvious member Pierre Fautrel said in response to the criticism, “who decided to print it on canvas, sign it as a mathematical formula, put it in a gold frame.” A century after Marcel Duchamp made a urinal into art [emphasis mine] by putting it in a gallery, not much has changed, with or without computers. As Andy Warhol famously said, “Art is what you can get away with.”

A bit of a segue here, there is a controversy as to whether or not that ‘urinal art’, also known as, The Fountain, should be attributed to Duchamp as noted in my January 23, 2019 posting titled ‘Baroness Elsa von Freytag-Loringhoven, Marcel Duchamp, and the Fountain’.

Getting back to the main action, Bogost goes on to describe the technologies underlying the two different AI artist-agents (Note: Links have been removed),

… Using a computer is hardly enough anymore; today’s machines offer all kinds of ways to generate images that can be output, framed, displayed, and sold—from digital photography to artificial intelligence. Recently, the fashionable choice has become generative adversarial networks, or GANs, the technology that created Portrait of Edmond de Belamy. Like other machine-learning methods, GANs use a sample set—in this case, art, or at least images of it—to deduce patterns, and then they use that knowledge to create new pieces. A typical Renaissance portrait, for example, might be composed as a bust or three-quarter view of a subject. The computer may have no idea what a bust is, but if it sees enough of them, it might learn the pattern and try to replicate it in an image.

GANs use two neural nets (a way of processing information modeled after the human brain) to produce images: a “generator” and a “discerner.” The generator produces new outputs—images, in the case of visual art—and the discerner tests them against the training set to make sure they comply with whatever patterns the computer has gleaned from that data. The quality or usefulness of the results depends largely on having a well-trained system, which is difficult.

That’s why folks in the know were upset by the Edmond de Belamy auction. The image was created by an algorithm the artists didn’t write, trained on an “Old Masters” image set they also didn’t create. The art world is no stranger to trend and bluster driving attention, but the brave new world of AI painting appeared to be just more found art, the machine-learning equivalent of a urinal on a plinth.

Ahmed Elgammal thinks AI art can be much more than that. A Rutgers University professor of computer science, Elgammal runs an art-and-artificial-intelligence lab, where he and his colleagues develop technologies that try to understand and generate new “art” (the scare quotes are Elgammal’s) with AI—not just credible copies of existing work, like GANs do. “That’s not art, that’s just repainting,” Elgammal says of GAN-made images. “It’s what a bad artist would do.”

Elgammal calls his approach a “creative adversarial network,” or CAN. It swaps a GAN’s discerner—the part that ensures similarity—for one that introduces novelty instead. The system amounts to a theory of how art evolves: through small alterations to a known style that produce a new one. That’s a convenient take, given that any machine-learning technique has to base its work on a specific training set.

The results are striking and strange, although calling them a new artistic style might be a stretch. They’re more like credible takes on visual abstraction. The images in the show, which were produced based on training sets of Renaissance portraits and skulls, are more figurative, and fairly disturbing. Their gallery placards name them dukes, earls, queens, and the like, although they depict no actual people—instead, human-like figures, their features smeared and contorted yet still legible as portraiture. Faceless Portrait of a Merchant, for example, depicts a torso that might also read as the front legs and rear haunches of a hound. Atop it, a fleshy orb comes across as a head. The whole scene is rippled by the machine-learning algorithm, in the way of so many computer-generated artworks.

Faceless Portrait of a Merchant, one of the AI portraits produced by Ahmed Elgammal and AICAN. (Artrendex Inc.) [downloaded from https://www.theatlantic.com/technology/archive/2019/03/ai-created-art-invades-chelsea-gallery-scene/584134/]

Bogost consults an expert on portraiture for a discussion about the particularities of portraiture and the shortcomings one might expect of an AI artist-agent (Note: A link has been removed),

“You can’t really pick a form of painting that’s more charged with cultural meaning than portraiture,” John Sharp, an art historian trained in 15th-century Italian painting and the director of the M.F.A. program in design and technology at Parsons School of Design, told me. The portrait isn’t just a style, it’s also a host for symbolism. “For example, men might be shown with an open book to show how they are in dialogue with that material; or a writing implement, to suggest authority; or a weapon, to evince power.” Take Portrait of a Youth Holding an Arrow, an early-16th-century Boltraffio portrait that helped train the AICAN database for the show. The painting depicts a young man, believed to be the Bolognese poet Girolamo Casio, holding an arrow at an angle in his fingers and across his chest. It doubles as both weapon and quill, a potent symbol of poetry and aristocracy alike. Along with the arrow, the laurels in Casio’s hair are emblems of Apollo, the god of both poetry and archery.

A neural net couldn’t infer anything about the particular symbolic trappings of the Renaissance or antiquity—unless it was taught to, and that wouldn’t happen just by showing it lots of portraits. For Sharp and other critics of computer-generated art, the result betrays an unforgivable ignorance about the supposed influence of the source material.

But for the purposes of the show, the appeal to the Renaissance might be mostly a foil, a way to yoke a hip, new technology to traditional painting in order to imbue it with the gravity of history: not only a Chelsea gallery show, but also an homage to the portraiture found at the Met. To reinforce a connection to the cradle of European art, some of the images are presented in elaborate frames, a decision the gallerist, Philippe Hoerle-Guggenheim (yes, that Guggenheim; he says the relation is “distant”) [the Guggenheim is strongly associated with the visual arts by way the two Guggeheim museums, one in New York City and the other in Bilbao, Portugal], told me he insisted upon. Meanwhile, the technical method makes its way onto the gallery placards in an official-sounding way—“Creative Adversarial Network print.” But both sets of inspirations, machine-learning and Renaissance portraiture, get limited billing and zero explanation at the show. That was deliberate, Hoerle-Guggenheim said. He’s betting that the simple existence of a visually arresting AI painting will be enough to draw interest—and buyers. It would turn out to be a good bet.

The art market is just that: a market. Some of the most renowned names in art today, from Damien Hirst to Banksy, trade in the trade of art as much as—and perhaps even more than—in the production of images, objects, and aesthetics. No artist today can avoid entering that fray, Elgammal included. “Is he an artist?” Hoerle-Guggenheim asked himself of the computer scientist. “Now that he’s in this context, he must be.” But is that enough? In Sharp’s estimation, “Faceless Portraits Transcending Time” is a tech demo more than a deliberate oeuvre, even compared to the machine-learning-driven work of his design-and-technology M.F.A. students, who self-identify as artists first.

Judged as Banksy or Hirst might be, Elgammal’s most art-worthy work might be the Artrendex start-up itself, not the pigment-print portraits that its technology has output. Elgammal doesn’t treat his commercial venture like a secret, but he also doesn’t surface it as a beneficiary of his supposedly earnest solo gallery show. He’s argued that AI-made images constitute a kind of conceptual art, but conceptualists tend to privilege process over product or to make the process as visible as the product.

Hoerle-Guggenheim worked as a financial analyst for Hyatt before getting into the art business via some kind of consulting deal (he responded cryptically when I pressed him for details). …

This is a fascinating article and I have one last excerpt, which poses this question, is an AI artist-agent a collaborator or a medium? There ‘s also speculation about how AI artist-agents might impact the business of art (Note: Links have been removed),

… it’s odd to list AICAN as a collaborator—painters credit pigment as a medium, not as a partner. Even the most committed digital artists don’t present the tools of their own inventions that way; when they do, it’s only after years, or even decades, of ongoing use and refinement.

But Elgammal insists that the move is justified because the machine produces unexpected results. “A camera is a tool—a mechanical device—but it’s not creative,” he said. “Using a tool is an unfair term for AICAN. It’s the first time in history that a tool has had some kind of creativity, that it can surprise you.” Casey Reas, a digital artist who co-designed the popular visual-arts-oriented coding platform Processing, which he uses to create some of his fine art, isn’t convinced. “The artist should claim responsibility over the work rather than to cede that agency to the tool or the system they create,” he told me.

Elgammal’s financial interest in AICAN might explain his insistence on foregrounding its role. Unlike a specialized print-making technique or even the Processing coding environment, AICAN isn’t just a device that Elgammal created. It’s also a commercial enterprise.

Elgammal has already spun off a company, Artrendex, that provides “artificial-intelligence innovations for the art market.” One of them offers provenance authentication for artworks; another can suggest works a viewer or collector might appreciate based on an existing collection; another, a system for cataloging images by visual properties and not just by metadata, has been licensed by the Barnes Foundation to drive its collection-browsing website.

The company’s plans are more ambitious than recommendations and fancy online catalogs. When presenting on a panel about the uses of blockchain for managing art sales and provenance, Elgammal caught the attention of Jessica Davidson, an art consultant who advises artists and galleries in building collections and exhibits. Davidson had been looking for business-development partnerships, and she became intrigued by AICAN as a marketable product. “I was interested in how we can harness it in a compelling way,” she says.

The art market is just that: a market. Some of the most renowned names in art today, from Damien Hirst to Banksy, trade in the trade of art as much as—and perhaps even more than—in the production of images, objects, and aesthetics. No artist today can avoid entering that fray, Elgammal included. “Is he an artist?” Hoerle-Guggenheim asked himself of the computer scientist. “Now that he’s in this context, he must be.” But is that enough? In Sharp’s estimation, “Faceless Portraits Transcending Time” is a tech demo more than a deliberate oeuvre, even compared to the machine-learning-driven work of his design-and-technology M.F.A. students, who self-identify as artists first.

Judged as Banksy or Hirst might be, Elgammal’s most art-worthy work might be the Artrendex start-up itself, not the pigment-print portraits that its technology has output. Elgammal doesn’t treat his commercial venture like a secret, but he also doesn’t surface it as a beneficiary of his supposedly earnest solo gallery show. He’s argued that AI-made images constitute a kind of conceptual art, but conceptualists tend to privilege process over product or to make the process as visible as the product.

Hoerle-Guggenheim worked as a financial analyst[emphasis mine] for Hyatt before getting into the art business via some kind of consulting deal (he responded cryptically when I pressed him for details). …

If you have the time, I recommend reading Bogost’s March 6, 2019 article for The Atlantic in its entirety/ these excerpts don’t do it enough justice.

Portraiture: what does it mean these days?

After reading the article I have a few questions. What exactly do Bogost and the arty types in the article mean by the word ‘portrait’? “Portrait of Edmond de Belamy” is an image of someone who doesn’t and never has existed and the exhibit “Faceless Portraits Transcending Time,” features images that don’t bear much or, in some cases, any resemblance to human beings. Maybe this is considered a dull question by people in the know but I’m an outsider and I found the paradox: portraits of nonexistent people or nonpeople kind of interesting.

BTW, I double-checked my assumption about portraits and found this definition in the Portrait Wikipedia entry (Note: Links have been removed),

A portrait is a painting, photograph, sculpture, or other artistic representation of a person [emphasis mine], in which the face and its expression is predominant. The intent is to display the likeness, personality, and even the mood of the person. For this reason, in photography a portrait is generally not a snapshot, but a composed image of a person in a still position. A portrait often shows a person looking directly at the painter or photographer, in order to most successfully engage the subject with the viewer.

So, portraits that aren’t portraits give rise to some philosophical questions but Bogost either didn’t want to jump into that rabbit hole (segue into yet another topic) or, as I hinted earlier, may have assumed his audience had previous experience of those kinds of discussions.

Vancouver (Canada) and a ‘portraiture’ exhibit at the Rennie Museum

By one of life’s coincidences, Vancouver’s Rennie Museum had an exhibit (February 16 – June 15, 2019) that illuminates questions about art collecting and portraiture, From a February 7, 2019 Rennie Museum news release,

‘downloaded from https://renniemuseum.org/press-release-spring-2019-collected-works/] Courtesy: Rennie Museum

February 7, 2019

Press Release | Spring 2019: Collected Works
By rennie museum

rennie museum is pleased to present Spring 2019: Collected Works, a group exhibition encompassing the mediums of photography, painting and film. A portraiture of the collecting spirit [emphasis mine], the works exhibited invite exploration of what collected objects, and both the considered and unintentional ways they are displayed, inform us. Featuring the works of four artists—Andrew Grassie, William E. Jones, Louise Lawler and Catherine Opie—the exhibition runs from February 16 to June 15, 2019.

Four exquisite paintings by Scottish painter Andrew Grassie detailing the home and private storage space of a major art collector provide a peek at how the passionately devoted integrates and accommodates the physical embodiments of such commitment into daily life. Grassie’s carefully constructed, hyper-realistic images also pose the question, “What happens to art once it’s sold?” In the transition from pristine gallery setting to idiosyncratic private space, how does the new context infuse our reading of the art and how does the art shift our perception of the individual?

Furthering the inquiry into the symbiotic exchange between possessor and possession, a selection of images by American photographer Louise Lawler depicting art installed in various private and public settings question how the bilateral relationship permeates our interpretation when the collector and the collected are no longer immediately connected. What does de-acquisitioning an object inform us and how does provenance affect our consideration of the art?

The question of legacy became an unexpected facet of 700 Nimes Road (2010-2011), American photographer Catherine Opie’s portrait of legendary actress Elizabeth Taylor. Opie did not directly photograph Taylor for any of the fifty images in the expansive portfolio. Instead, she focused on Taylor’s home and the objects within, inviting viewers to see—then see beyond—the façade of fame and consider how both treasures and trinkets act as vignettes to the stories of a life. Glamorous images of jewels and trophies juxtapose with mundane shots of a printer and the remote-control user manual. Groupings of major artworks on the wall are as illuminating of the home’s mistress as clusters of personal photos. Taylor passed away part way through Opie’s project. The subsequent photos include Taylor’s mementos heading off to auction, raising the question, “Once the collections that help to define someone are disbursed, will our image of that person lose focus?”

In a similar fashion, the twenty-two photographs in Villa Iolas (1982/2017), by American artist and filmmaker William E. Jones, depict the Athens home of iconic art dealer and collector Alexander Iolas. Taken in 1982 by Jones during his first travels abroad, the photographs of art, furniture and antiquities tell a story of privilege that contrast sharply with the images Jones captures on a return visit in 2016. Nearly three decades after Iolas’s 1989 death, his home sits in dilapidation, looted and vandalized. Iolas played an extraordinary role in the evolution of modern art, building the careers of Max Ernst, Yves Klein and Giorgio de Chirico. He gave Andy Warhol his first solo exhibition and was a key advisor to famed collectors John and Dominique de Menil. Yet in the years since his death, his intention of turning his home into a modern art museum as a gift to Greece, along with his reputation, crumbled into ruins. The photographs taken by Jones during his visits in two different eras are incorporated into the film Fall into Ruin (2017), along with shots of contemporary Athens and antiquities on display at the National Archaeological Museum.

“I ask a lot of questions about how portraiture functionswhat is there to describe the person or time we live in or a certain set of politics…”
 – Catherine Opie, The Guardian, Feb 9, 2016

We tend to think of the act of collecting as a formal activity yet it can happen casually on a daily basis, often in trivial ways. While we readily acknowledge a collector consciously assembling with deliberate thought, we give lesser consideration to the arbitrary accumulations that each of us accrue. Be it master artworks, incidental baubles or random curios, the objects we acquire and surround ourselves with tell stories of who we are.

Andrew Grassie (Scotland, b. 1966) is a painter known for his small scale, hyper-realist works. He has been the subject of solo exhibitions at the Tate Britain; Talbot Rice Gallery, Edinburgh; institut supérieur des arts de Toulouse; and rennie museum, Vancouver, Canada. He lives and works in London, England.

William E. Jones (USA, b. 1962) is an artist, experimental film-essayist and writer. Jones’s work has been the subject of retrospectives at Tate Modern, London; Anthology Film Archives, New York; Austrian Film Museum, Vienna; and, Oberhausen Short Film Festival. He is a recipient of the John Simon Guggenheim Memorial Fellowship and the Creative Capital/Andy Warhol Foundation Arts Writers Grant. He lives and works in Los Angeles, USA.

Louise Lawler (USA, b. 1947) is a photographer and one of the foremost members of the Pictures Generation. Lawler was the subject of a major retrospective at the Museum of Modern Art, New York in 2017. She has held exhibitions at the Whitney Museum of American Art, New York; Stedelijk Museum, Amsterdam; National Museum of Art, Oslo; and Musée d’Art Moderne de La Ville de Paris. She lives and works in New York.

Catherine Opie (USA, b. 1961) is a photographer and educator. Her work has been exhibited at Wexner Center for the Arts, Ohio; Henie Onstad Art Center, Oslo; Los the Angeles County Museum of Art; Portland Art Museum; and the Guggenheim Museum, New York. She is the recipient of United States Artist Fellowship, Julius Shulman’s Excellence in Photography Award, and the Smithsonian’s Archive of American Art Medal.  She lives and works in Los Angeles.

rennie museum opened in October 2009 in historic Wing Sang, the oldest structure in Vancouver’s Chinatown, to feature dynamic exhibitions comprising only of art drawn from rennie collection. Showcasing works by emerging and established international artists, the exhibits, accompanied by supporting catalogues, are open free to the public through engaging guided tours. The museum’s commitment to providing access to arts and culture is also expressed through its education program, which offers free age-appropriate tours and customized workshops to children of all ages.

rennie collection is a globally recognized collection of contemporary art that focuses on works that tackle issues related to identity, social commentary and injustice, appropriation, and the nature of painting, photography, sculpture and film. Currently the collection includes works by over 370 emerging and established artists, with over fifty collected in depth. The Vancouver based collection engages actively with numerous museums globally through a robust, artist-centric, lending policy.

So despite the Wikipedia definition, it seems that portraits don’t always feature people. While Bogost didn’t jump into that particular rabbit hole, he did touch on the business side of art.

What about intellectual property?

Bogost doesn’t explicitly discuss this particular issue. It’s a big topic so I’m touching on it only lightly, if an artist worsk with an AI, the question as to ownership of the artwork could prove thorny. Is the copyright owner the computer scientist or the artist or both? Or does the AI artist-agent itself own the copyright? That last question may not be all that farfetched. Sophia, a social humanoid robot, has occasioned thought about ‘personhood.’ (Note: The robots mentioned in this posting have artificial intelligence.) From the Sophia (robot) Wikipedia entry (Note: Links have been removed),

Sophia has been interviewed in the same manner as a human, striking up conversations with hosts. Some replies have been nonsensical, while others have impressed interviewers such as 60 Minutes’ Charlie Rose.[12] In a piece for CNBC, when the interviewer expressed concerns about robot behavior, Sophia joked that he had “been reading too much Elon Musk. And watching too many Hollywood movies”.[27] Musk tweeted that Sophia should watch The Godfather and asked “what’s the worst that could happen?”[28][29] Business Insider’s chief UK editor Jim Edwards interviewed Sophia, and while the answers were “not altogether terrible”, he predicted it was a step towards “conversational artificial intelligence”.[30] At the 2018 Consumer Electronics Show, a BBC News reporter described talking with Sophia as “a slightly awkward experience”.[31]

On October 11, 2017, Sophia was introduced to the United Nations with a brief conversation with the United Nations Deputy Secretary-General, Amina J. Mohammed.[32] On October 25, at the Future Investment Summit in Riyadh, the robot was granted Saudi Arabian citizenship [emphasis mine], becoming the first robot ever to have a nationality.[29][33] This attracted controversy as some commentators wondered if this implied that Sophia could vote or marry, or whether a deliberate system shutdown could be considered murder. Social media users used Sophia’s citizenship to criticize Saudi Arabia’s human rights record. In December 2017, Sophia’s creator David Hanson said in an interview that Sophia would use her citizenship to advocate for women’s rights in her new country of citizenship; Newsweek criticized that “What [Hanson] means, exactly, is unclear”.[34] On November 27, 2018 Sophia was given a visa by Azerbaijan while attending Global Influencer Day Congress held in Baku. December 15, 2018 Sophia was appointed a Belt and Road Innovative Technology Ambassador by China'[35]

As for an AI artist-agent’s intellectual property rights , I have a July 10, 2017 posting featuring that question in more detail. Whether you read that piece or not, it seems obvious that artists might hesitate to call an AI agent, a partner rather than a medium of expression. After all, a partner (and/or the computer scientist who developed the programme) might expect to share in property rights and profits but paint, marble, plastic, and other media used by artists don’t have those expectations.

Moving slightly off topic , in my July 10, 2017 posting I mentioned a competition (literary and performing arts rather than visual arts) called, ‘Dartmouth College and its Neukom Institute Prizes in Computational Arts’. It was started in 2016 and, as of 2018, was still operational under this name: Creative Turing Tests. Assuming there’ll be contests for prizes in 2019, there’s (from the contest site) [1] PoetiX, competition in computer-generated sonnet writing; [2] Musical Style, composition algorithms in various styles, and human-machine improvisation …; and [3] DigiLit, algorithms able to produce “human-level” short story writing that is indistinguishable from an “average” human effort. You can find the contest site here.

An artificial synapse tuned by light, a ferromagnetic memristor, and a transparent, flexible artificial synapse

Down the memristor rabbit hole one more time.* I started out with news about two new papers and inadvertently found two more. In a bid to keep this posting to a manageable size, I’m stopping at four.

UK

In a June 19, 2019 Nanowerk Spotlight article, Dr. Neil Kemp discusses memristors and some of his latest work (Note: A link has been removed),

Memristor (or memory resistors) devices are non-volatile electronic memory devices that were first theorized by Leon Chua in the 1970’s. However, it was some thirty years later that the first practical device was fabricated. This was in 2008 when a group led by Stanley Williams at HP Research Labs realized that switching of the resistance between a conducting and less conducting state in metal-oxide thin-film devices was showing Leon Chua’s memristor behaviour.

The high interest in memristor devices also stems from the fact that these devices emulate the memory and learning properties of biological synapses. i.e. the electrical resistance value of the device is dependent on the history of the current flowing through it.

There is a huge effort underway to use memristor devices in neuromorphic computing applications and it is now reasonable to imagine the development of a new generation of artificial intelligent devices with very low power consumption (non-volatile), ultra-fast performance and high-density integration.

These discoveries come at an important juncture in microelectronics, since there is increasing disparity between computational needs of Big Data, Artificial Intelligence (A.I.) and the Internet of Things (IoT), and the capabilities of existing computers. The increases in speed, efficiency and performance of computer technology cannot continue in the same manner as it has done since the 1960s.

To date, most memristor research has focussed on the electronic switching properties of the device. However, for many applications it is useful to have an additional handle (or degree of freedom) on the device to control its resistive state. For example memory and processing in the brain also involves numerous chemical and bio-chemical reactions that control the brain structure and its evolution through development.

To emulate this in a simple solid-state system composed of just switches alone is not possible. In our research, we are interested in using light to mediate this essential control.

We have demonstrated that light can be used to make short and long-term memory and we have shown how light can modulate a special type of learning, called spike timing dependent plasticity (STDP). STDP involves two neuronal spikes incident across a synapse at the same time. Depending on the relative timing of the spikes and their overlap across the synaptic cleft, the connection strength is other strengthened or weakened.

In our earlier work, we were only able to achieve to small switching effects in memristors using light. In our latest work (Advanced Electronic Materials, “Percolation Threshold Enables Optical Resistive-Memory Switching and Light-Tuneable Synaptic Learning in Segregated Nanocomposites”), we take advantage of a percolating-like nanoparticle morphology to vastly increase the magnitude of the switching between electronic resistance states when light is incident on the device.

We have used an inhomogeneous percolating network consisting of metallic nanoparticles distributed in filamentary-like conduction paths. Electronic conduction and the resistance of the device is very sensitive to any disruption of the conduction path(s).

By embedding the nanoparticles in a polymer that can expand or contract with light the conduction pathways are broken or re-connected causing very large changes in the electrical resistance and memristance of the device.

Our devices could lead to the development of new memristor-based artificial intelligence systems that are adaptive and reconfigurable using a combination of optical and electronic signalling. Furthermore, they have the potential for the development of very fast optical cameras for artificial intelligence recognition systems.

Our work provides a nice proof-of-concept but the materials used means the optical switching is slow. The materials are also not well suited to industry fabrication. In our on-going work we are addressing these switching speed issues whilst also focussing on industry compatible materials.

Currently we are working on a new type of optical memristor device that should give us orders of magnitude improvement in the optical switching speeds whilst also retaining a large difference between the resistance on and off states. We hope to be able to achieve nanosecond switching speeds. The materials used are also compatible with industry standard methods of fabrication.

The new devices should also have applications in optical communications, interfacing and photonic computing. We are currently looking for commercial investors to help fund the research on these devices so that we can bring the device specifications to a level of commercial interest.

If you’re interested in memristors, Kemp’s article is well written and quite informative for nonexperts, assuming of course you can tolerate not understanding everything perfectly.

Here are links and citations for two papers. The first is the latest referred to in the article, a May 2019 paper and the second is a paper appearing in July 2019.

Percolation Threshold Enables Optical Resistive‐Memory Switching and Light‐Tuneable Synaptic Learning in Segregated Nanocomposites by Ayoub H. Jaafar, Mary O’Neill, Stephen M. Kelly, Emanuele Verrelli, Neil T. Kemp. Advanced Electronic Materials DOI: https://doi.org/10.1002/aelm.201900197 First published: 28 May 2019

Wavelength dependent light tunable resistive switching graphene oxide nonvolatile memory devices by Ayoub H.Jaafar, N.T.Kemp. DOI: https://doi.org/10.1016/j.carbon.2019.07.007 Carbon Available online 3 July 2019

The first paper (May 2019) is definitely behind a paywall and the second paper (July 2019) appears to be behind a paywall.

Dr. Kemp’s work has been featured here previously in a January 3, 2018 posting in the subsection titled, Shining a light on the memristor.

China

This work from China was announced in a June 20, 2019 news item on Nanowerk,

Memristors, demonstrated by solid-state devices with continuously tunable resistance, have emerged as a new paradigm for self-adaptive networks that require synapse-like functions. Spin-based memristors offer advantages over other types of memristors because of their significant endurance and high energy effciency.

However, it remains a challenge to build dense and functional spintronic memristors with structures and materials that are compatible with existing ferromagnetic devices. Ta/CoFeB/MgO heterostructures are commonly used in interfacial PMA-based [perpendicular magnetic anisotropy] magnetic tunnel junctions, which exhibit large tunnel magnetoresistance and are implemented in commercial MRAM [magnetic random access memory] products.

“To achieve the memristive function, DW is driven back and forth in a continuous manner in the CoFeB layer by applying in-plane positive or negative current pulses along the Ta layer, utilizing SOT that the current exerts on the CoFeB magnetization,” said Shuai Zhang, a coauthor in the paper. “Slowly propagating domain wall generates a creep in the detection area of the device, which yields a broad range of intermediate resistive states in the AHE [anomalous Hall effect] measurements. Consequently, AHE resistance is modulated in an analog manner, being controlled by the pulsed current characteristics including amplitude, duration, and repetition number.”

“For a follow-up study, we are working on more neuromorphic operations, such as spike-timing-dependent plasticity and paired pulsed facilitation,” concludes You. …

Here’s are links to and citations for the paper (Note: It’s a little confusing but I believe that one of the links will take you to the online version, as for the ‘open access’ link, keep reading),

A Spin–Orbit‐Torque Memristive Device by Shuai Zhang, Shijiang Luo, Nuo Xu, Qiming Zou, Min Song, Jijun Yun, Qiang Luo, Zhe Guo, Ruofan Li, Weicheng Tian, Xin Li, Hengan Zhou, Huiming Chen, Yue Zhang, Xiaofei Yang, Wanjun Jiang, Ka Shen, Jeongmin Hong, Zhe Yuan, Li Xi, Ke Xia, Sayeef Salahuddin, Bernard Dieny, Long You. Advanced Electronic Materials Volume 5, Issue 4 April 2019 (print version) 1800782 DOI: https://doi.org/10.1002/aelm.201800782 First published [online]: 30 January 2019 Note: there is another DOI, https://doi.org/10.1002/aelm.201970022 where you can have open access to Memristors: A Spin–Orbit‐Torque Memristive Device (Adv. Electron. Mater. 4/2019)

The paper published online in January 2019 is behind a paywall and the paper (almost the same title) published in April 2019 has a new DOI and is open access. Final note: I tried accessing the ‘free’ paper and opened up a free file for the artwork featuring the work from China on the back cover of the April 2019 of Advanced Electronic Materials.

Korea

Usually when I see the words transparency and flexibility, I expect to see graphene is one of the materials. That’s not the case for this paper (link to and citation for),

Transparent and flexible photonic artificial synapse with piezo-phototronic modulator: Versatile memory capability and higher order learning algorithm by Mohit Kumar, Joondong Kim, Ching-Ping Wong. Nano Energy Volume 63, September 2019, 103843 DOI: https://doi.org/10.1016/j.nanoen.2019.06.039 Available online 22 June 2019

Here’s the abstract for the paper where you’ll see that the material is made up of zinc oxide silver nanowires,

An artificial photonic synapse having tunable manifold synaptic response can be an essential step forward for the advancement of novel neuromorphic computing. In this work, we reported the development of highly transparent and flexible two-terminal ZnO/Ag-nanowires/PET photonic artificial synapse [emphasis mine]. The device shows purely photo-triggered all essential synaptic functions such as transition from short-to long-term plasticity, paired-pulse facilitation, and spike-timing-dependent plasticity, including in the versatile memory capability. Importantly, strain-induced piezo-phototronic effect within ZnO provides an additional degree of regulation to modulate all of the synaptic functions in multi-levels. The observed effect is quantitatively explained as a dynamic of photo-induced electron-hole trapping/detraining via the defect states such as oxygen vacancies. We revealed that the synaptic functions can be consolidated and converted by applied strain, which is not previously applied any of the reported synaptic devices. This study will open a new avenue to the scientific community to control and design highly transparent wearable neuromorphic computing.

This paper is behind a paywall.

Artificial synapse based on tantalum oxide from Korean researchers

This memristor story comes from South Korea as we progress on the way to neuromorphic computing (brainlike computing). A Sept. 7, 2018 news item on ScienceDaily makes the announcement,

A research team led by Director Myoung-Jae Lee from the Intelligent Devices and Systems Research Group at DGIST (Daegu Gyeongbuk Institute of Science and Technology) has succeeded in developing an artificial synaptic device that mimics the function of the nerve cells (neurons) and synapses that are response for memory in human brains. [sic]

Synapses are where axons and dendrites meet so that neurons in the human brain can send and receive nerve signals; there are known to be hundreds of trillions of synapses in the human brain.

This chemical synapse information transfer system, which transfers information from the brain, can handle high-level parallel arithmetic with very little energy, so research on artificial synaptic devices, which mimic the biological function of a synapse, is under way worldwide.

Dr. Lee’s research team, through joint research with teams led by Professor Gyeong-Su Park from Seoul National University; Professor Sung Kyu Park from Chung-ang University; and Professor Hyunsang Hwang from Pohang University of Science and Technology (POSTEC), developed a high-reliability artificial synaptic device with multiple values by structuring tantalum oxide — a trans-metallic material — into two layers of Ta2O5-x and TaO2-x and by controlling its surface.

A September 7, 2018 DGIST press release (also on EurekAlert), which originated the news item, delves further into the work,

The artificial synaptic device developed by the research team is an electrical synaptic device that simulates the function of synapses in the brain as the resistance of the tantalum oxide layer gradually increases or decreases depending on the strength of the electric signals. It has succeeded in overcoming durability limitations of current devices by allowing current control only on one layer of Ta2O5-x.

In addition, the research team successfully implemented an experiment that realized synapse plasticity [or synaptic plasticity], which is the process of creating, storing, and deleting memories, such as long-term strengthening of memory and long-term suppression of memory deleting by adjusting the strength of the synapse connection between neurons.

The non-volatile multiple-value data storage method applied by the research team has the technological advantage of having a small area of an artificial synaptic device system, reducing circuit connection complexity, and reducing power consumption by more than one-thousandth compared to data storage methods based on digital signals using 0 and 1 such as volatile CMOS (Complementary Metal Oxide Semiconductor).

The high-reliability artificial synaptic device developed by the research team can be used in ultra-low-power devices or circuits for processing massive amounts of big data due to its capability of low-power parallel arithmetic. It is expected to be applied to next-generation intelligent semiconductor device technologies such as development of artificial intelligence (AI) including machine learning and deep learning and brain-mimicking semiconductors.

Dr. Lee said, “This research secured the reliability of existing artificial synaptic devices and improved the areas pointed out as disadvantages. We expect to contribute to the development of AI based on the neuromorphic system that mimics the human brain by creating a circuit that imitates the function of neurons.”

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

Reliable Multivalued Conductance States in TaOx Memristors through Oxygen Plasma-Assisted Electrode Deposition with in Situ-Biased Conductance State Transmission Electron Microscopy Analysis by Myoung-Jae Lee, Gyeong-Su Park, David H. Seo, Sung Min Kwon, Hyeon-Jun Lee, June-Seo Kim, MinKyung Jung, Chun-Yeol You, Hyangsook Lee, Hee-Goo Kim, Su-Been Pang, Sunae Seo, Hyunsang Hwang, and Sung Kyu Park. ACS Appl. Mater. Interfaces, 2018, 10 (35), pp 29757–29765 DOI: 10.1021/acsami.8b09046 Publication Date (Web): July 23, 2018

Copyright © 2018 American Chemical Society

This paper is open access.

You can find other memristor and neuromorphic computing stories here by using the search terms I’ve highlighted,  My latest (more or less) is an April 19, 2018 posting titled, New path to viable memristor/neuristor?

Finally, here’s an image from the Korean researchers that accompanied their work,

Caption: Representation of neurons and synapses in the human brain. The magnified synapse represents the portion mimicked using solid-state devices. Credit: Daegu Gyeongbuk Institute of Science and Technology(DGIST)