Category Archives: neuromorphic engineering

Brain-inspired navigation technology for robots

Leave a reply

An August 22, 2024 Beijing Institute of Technology Press Co. press release on EurekAlert announces the publication of a paper reviewing the current state of research into brain-inspired (neuromorphic) navigation technology,

In the ever-evolving field of robotics, a groundbreaking approach has emerged, revolutionizing how robots perceive, navigate, and interact with their environments. This new frontier, known as brain-inspired navigation technology, integrates insights from neuroscience into robotics, offering enhanced capabilities and efficiency.

Brain-inspired navigation technologies are not just a mere improvement over traditional methods; they represent a paradigm shift. By mimicking the neural mechanisms of animals, these technologies provide robots with the ability to navigate through complex and unknown terrains with unprecedented accuracy and adaptability.

At the heart of this technology lies the concept of spatial cognition, which is central to how animals, including humans, navigate their environments. Spatial cognition involves the brain’s ability to organize and interpret spatial data for navigation and memory. Robots equipped with brain-inspired navigation systems utilize a multi-layered network model that integrates sensory data from multiple sources. This model allows the robot to create a ‘cognitive map’ of its surroundings, much like the neural maps created by the hippocampus in the human brain.

One of the significant advantages of brain-inspired navigation is its robustness in challenging environments. Traditional navigation systems often struggle with dynamic and unpredictable settings, where the reliance on pre-mapped routes and landmarks can lead to failures. In contrast, brain-inspired systems continuously learn and adapt, improving their navigational strategies over time. This capability is particularly beneficial in environments like disaster zones or extraterrestrial surfaces, where prior mapping is either impossible or impractical.

Moreover, these systems significantly reduce energy consumption and computational needs. By focusing only on essential data and employing efficient neural network models, robots can operate longer and perform more complex tasks without the need for frequent recharging or maintenance.

The technology’s applications are vast and varied. For instance, autonomous vehicles equipped with brain-inspired systems could navigate more safely and efficiently, reacting in real-time to sudden changes in traffic conditions or road layouts. Similarly, drones used for delivery services could plan their routes more effectively, avoiding obstacles and optimizing delivery times.

Despite its promising potential, the development of brain-inspired navigation technology faces several challenges. Integrating biological principles into mechanical systems is inherently complex, requiring multidisciplinary efforts from fields such as neuroscience, cognitive science, robotics, and artificial intelligence. Moreover, these systems must be scalable and versatile enough to be customized for different types of robotic platforms and applications.

As researchers continue to unravel the mysteries of the brain’s navigational capabilities, the future of robotics looks increasingly intertwined with the principles of neuroscience. The collaboration across disciplines promises not only to advance our understanding of the brain but also to pave the way for a new generation of intelligent robots. These robots will not only assist in mundane tasks but also perform critical roles in search and rescue operations, planetary exploration, and much more.

In conclusion, brain-inspired navigation technology represents a significant leap forward in robotics, merging the abstract with the applied, the biological with the mechanical, and the theoretical with the practical. As this technology continues to evolve, it will undoubtedly open new horizons for robotic applications, making machines an even more integral part of our daily lives and work.

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

A Review of Brain-Inspired Cognition and Navigation Technology for Mobile Robots by Yanan Bai, Shiliang Shao, Jin Zhang, Xianzhe Zhao, Chuxi Fang, Ting Wang, Yongliang Wang, and Hai Zhao. Cyborg and Bionic Systems 27 Jun 2024 Vol 5 Article ID: 0128 DOI: 10.34133/cbsystems.0128

This paper is open access.

About the journal publisher, Science Journal Partners

Cyborg and Bionic Systems is published by the American Association for the Advancement of Science (AAAS) and is part of an open access publishing project known as Science Journal Partners, from the Program Overview webpage of science.org,

The Science Partner Journal (SPJ) program was launched in late 2017 by the American Association for the Advancement of Science (AAAS), the nonprofit publisher of the Science family of journals.

The program features high-quality, online-only, Open Access publications produced in collaboration with international research institutions, foundations, funders, and societies. Through these collaborations, AAAS furthers its mission to communicate science broadly and for the benefit of all people by providing top-tier international research organizations with the technology, visibility, and publishing expertise that AAAS is uniquely positioned to offer as the world’s largest general science membership society.

Organizations participating in the SPJ program are editorially independent and responsible for the content published in each journal. To oversee the publication process, each organization appoints editors committed to best practices in peer review and author service. It is the responsibility of the staff of each SPJ title to establish and execute all aspects of the peer review process, including oversight of editorial policy, selection of papers, and editing of content, following best practices advised by AAAS.

…

I’m starting to catch up with changes in the world of science publishing as you can see in my October 1, 2024 posting titled, “Nanomedicine: two stories about wound healing,” which features a subhead near the end of the post, Science Publishing, should you be interested in another science publishing initiative.

Light-based neural networks

Leave a reply

It’s unusual to see the same headline used to highlight research from two different teams released in such proximity, February 2024 and July 2024, respectively. Both of these are neuromorphic (brainlike) computing stories.

February 2024: Neural networks made of light

The first team’s work is announced in a February 21, 2024 Friedrich Schiller University press release, Note: A link has been removed,

Researchers from the Leibniz Institute of Photonic Technology (Leibniz IPHT) and the Friedrich Schiller University in Jena, along with an international team, have developed a new technology that could significantly reduce the high energy demands of future AI systems. This innovation utilizes light for neuronal computing, inspired by the neural networks of the human brain. It promises not only more efficient data processing but also speeds many times faster than current methods, all while consuming considerably less energy. Published in the prestigious journal „Advanced Science,“ their work introduces new avenues for environmentally friendly AI applications, as well as advancements in computerless diagnostics and intelligent microscopy.

Artificial intelligence (AI) is pivotal in advancing biotechnology and medical procedures, ranging from cancer diagnostics to the creation of new antibiotics. However, the ecological footprint of large-scale AI systems is substantial. For instance, training extensive language models like ChatGPT-3 requires several gigawatt-hours of energy—enough to power an average nuclear power plant at full capacity for several hours.

Prof. Mario Chemnitz, new Junior Professor of Intelligent Photonic SystemsExternal link at Friedrich Schiller University Jena, and Dr Bennet Fischer from Leibniz IPHT in Jena, in collaboration with their international team, have devised an innovative method to develop potentially energy-efficient computing systems that forego the need for extensive electronic infrastructure. They harness the unique interactions of light waves within optical fibers to forge an advanced artificial learning system.

A single fiber instead of thousands of components

Unlike traditional systems that rely on computer chips containing thousands of electronic components, their system uses a single optical fiber. This fiber is capable of performing the tasks of various neural networks—at the speed of light. “We utilize a single optical fiber to mimic the computational power of numerous neural networks,“ Mario Chemnitz, who is also leader of the “Smart Photonics“ junior research group at Leibniz IPHT, explains. “By leveraging the unique physical properties of light, this system will enable the rapid and efficient processing of vast amounts of data in the future.“

Delving into the mechanics reveals how information transmission occurs through the mixing of light frequencies: Data—whether pixel values from images or frequency components of an audio track—are encoded onto the color channels of ultrashort light pulses. These pulses carry the information through the fiber, undergoing various combinations, amplifications, or attenuations. The emergence of new color combinations at the fiber’s output enables the prediction of data types or contexts. For example, specific color channels can indicate visible objects in images or signs of illness in a voice.

A prime example of machine learning is identifying different numbers from thousands of handwritten characters. Mario Chemnitz, Bennet Fischer, and their colleagues from the Institut National de la Recherche Scientifique (INRS) in Québec utilized their technique to encode images of handwritten digits onto light signals and classify them via the optical fiber. The alteration in color composition at the fiber’s end forms a unique color spectrum—a „fingerprint“ for each digit. Following training, the system can analyze and recognize new handwriting digits with significantly reduced energy consumption.

System recognizes COVID-19 from voice samples

“In simpler terms, pixel values are converted into varying intensities of primary colors—more red or less blue, for instance,“ Mario Chemnitz details. “Within the fiber, these primary colors blend to create the full spectrum of the rainbow. The shade of our mixed purple, for example, reveals much about the data processed by our system.“

The team has also successfully applied this method in a pilot study to diagnose COVID-19 infections using voice samples, achieving a detection rate that surpasses the best digital systems to date.

“We are the first to demonstrate that such a vibrant interplay of light waves in optical fibers can directly classify complex information without any additional intelligent software,“ Mario Chemnitz states.

Since December 2023, Mario Chemnitz has held the position of Junior Professor of Intelligent Photonic Systems at Friedrich Schiller University Jena. Following his return from INRS in Canada in 2022, where he served as a postdoc, Chemnitz has been leading an international team at Leibniz IPHT in Jena. With Nexus funding support from the Carl Zeiss Foundation, their research focuses on exploring the potentials of non-linear optics. Their goal is to develop computer-free intelligent sensor systems and microscopes, as well as techniques for green computing.

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

Neuromorphic Computing via Fission-based Broadband Frequency Generation by Bennet Fischer, Mario Chemnitz, Yi Zhu, Nicolas Perron, Piotr Roztocki, Benjamin MacLellan, Luigi Di Lauro, A. Aadhi, Cristina Rimoldi, Tiago H. Falk, Roberto Morandotti. Advanced Science Volume 10, Issue 35 December 15, 2023 2303835 DOI: https://doi.org/10.1002/advs.202303835. First published: 02 October 2023

This paper is open access.

July 2024: Neural networks made of light

A July 12, 2024 news item on ScienceDaily announces research from another German team,

Scientists propose a new way of implementing a neural network with an optical system which could make machine learning more sustainable in the future. The researchers at the Max Planck Institute for the Science of Light have published their new method in Nature Physics, demonstrating a method much simpler than previous approaches.

…

A July 12, 2024 Max Planck Institute for the Science of Light press release (also on EurekAlert), which originated the news item, provides more detail about their approach to neuromorphic computiing,

Machine learning and artificial intelligence are becoming increasingly widespread with applications ranging from computer vision to text generation, as demonstrated by ChatGPT. However, these complex tasks require increasingly complex neural networks; some with many billion parameters. This rapid growth of neural network size has put the technologies on an unsustainable path due to their exponentially growing energy consumption and training times. For instance, it is estimated that training GPT-3 consumed more than 1,000 MWh of energy, which amounts to the daily electrical energy consumption of a small town. This trend has created a need for faster, more energy- and cost-efficient alternatives, sparking the rapidly developing field of neuromorphic computing. The aim of this field is to replace the neural networks on our digital computers with physical neural networks. These are engineered to perform the required mathematical operations physically in a potentially faster and more energy-efficient way.

Optics and photonics are particularly promising platforms for neuromorphic computing since energy consumption can be kept to a minimum. Computations can be performed in parallel at very high speeds only limited by the speed of light. However, so far, there have been two significant challenges: Firstly, realizing the necessary complex mathematical computations requires high laser powers. Secondly, the lack of an efficient general training method for such physical neural networks.

Both challenges can be overcome with the new method proposed by Clara Wanjura and Florian Marquardt from the Max Planck Institute for the Science of Light in their new article in Nature Physics. “Normally, the data input is imprinted on the light field. However, in our new methods we propose to imprint the input by changing the light transmission,” explains Florian Marquardt, Director at the Institute. In this way, the input signal can be processed in an arbitrary fashion. This is true even though the light field itself behaves in the simplest way possible in which waves interfere without otherwise influencing each other. Therefore, their approach allows one to avoid complicated physical interactions to realize the required mathematical functions which would otherwise require high-power light fields. Evaluating and training this physical neural network would then become very straightforward: “It would really be as simple as sending light through the system and observing the transmitted light. This lets us evaluate the output of the network. At the same time, this allows one to measure all relevant information for the training”, says Clara Wanjura, the first author of the study. The authors demonstrated in simulations that their approach can be used to perform image classification tasks with the same accuracy as digital neural networks.

In the future, the authors are planning to collaborate with experimental groups to explore the implementation of their method. Since their proposal significantly relaxes the experimental requirements, it can be applied to many physically very different systems. This opens up new possibilities for neuromorphic devices allowing physical training over a broad range of platforms.

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

Fully nonlinear neuromorphic computing with linear wave scattering by Clara C. Wanjura & Florian Marquardt. Nature Physics (2024) DOI: https://doi.org/10.1038/s41567-024-02534-9 Published: 09 July 2024

This paper is open access.

Dual functions—neuromorphic (brainlike) and security—with papertronic devices

Leave a reply

Michael Berger’s June 27, 2024 Nanowerk Spotlight article describes some of the latest work on developing electronic paper devices (yes, paper), Note 1: Links have been removed, Note 2: If you do check out Berger’s article, you will need to click a box confirming you are human,+

Paper-based electronic devices have long been an intriguing prospect for researchers, offering potential advantages in sustainability, cost-effectiveness, and flexibility. However, translating the unique properties of paper into functional electronic components has presented significant challenges. Traditional semiconductor manufacturing processes are incompatible with paper’s thermal sensitivity and porous structure. Previous attempts to create paper-based electronics often resulted in devices with limited functionality or poor durability.

Recent advances in materials science and nanofabrication techniques have opened new avenues for realizing sophisticated electronic devices on paper substrates. Researchers have made progress in developing conductive inks, flexible electrodes, and solution-processable semiconductors that can be applied to paper without compromising its inherent properties. These developments have paved the way for creating paper-based sensors, energy storage devices, and simple circuits.

Despite these advancements, achieving complex electronic functionalities on paper, particularly in areas like neuromorphic computing and security applications, has remained elusive. Neuromorphic devices, which mimic the behavior of biological synapses, typically require precise control of charge transport and storage mechanisms.

Similarly, physically unclonable functions (PUFs) used in security applications depend on the ability to generate random, unique patterns at the nanoscale level. Implementing these sophisticated functionalities on paper substrates has been a persistent challenge due to the material’s inherent variability and limited compatibility with advanced fabrication techniques.

A research team in Korea has now made significant strides in addressing these challenges, developing a versatile paper-based electronic device that demonstrates both neuromorphic and security capabilities. Their work, published in Advanced Materials (“Versatile Papertronics: Photo-Induced Synapse and Security Applications on Papers”), describes a novel approach to creating multifunctional “papertronics” using a combination of solution-processable materials and innovative device architectures.

…

The team showcased the potential of their device by simulating a facial recognition task. Using a simple neural network architecture and the light-responsive properties of their paper-based device, they achieved a recognition accuracy of 91.7% on a standard face database. This impressive performance was achieved with a remarkably low voltage bias of -0.01 V, demonstrating the energy efficiency of the approach. The ability to operate at such low voltages is particularly advantageous for portable and low-power applications.

In addition to its neuromorphic capabilities, the device also showed promise as a physically unclonable function (PUF) for security applications. The researchers leveraged the inherent randomness in the deposition of SnO2 nanoparticles [tin oxide nanoparticles] to create unique electrical characteristics in each device. By fabricating arrays of these devices on paper, they generated security keys that exhibited high levels of randomness and uniqueness.

…

One of the most intriguing aspects of this research is the dual functionality achieved with a single device structure. The ability to serve as both a neuromorphic component and a security element could lead to the development of highly integrated, secure edge computing devices on paper substrates. This convergence of functionalities addresses growing concerns about data privacy and security in Internet of Things (IoT) applications.

…

Berger’s June 27, 2024 Nanowerk Spotlight article offers more detail about the work and it’s written in an accessible fashion. Berger also notes at the end, that there are still a lot of challenges before this work leaves the laboratory.

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

Versatile Papertronics: Photo-Induced Synapse and Security Applications on Papers by Wangmyung Choi, Jihyun Shin, Yeong Jae Kim, Jaehyun Hur, Byung Chul Jang, Hocheon Yoo. Advanced Materials DOI: https://doi.org/10.1002/adma.202312831 First published: 13 June 2024

This paper is behind a paywall.

Proposed platform for brain-inspired computing

Leave a reply

Researchers at the University of California at Santa Barbara (UCSB) have proposed a more energy-efficient architecture for neuromorphic (brainlike or brain-inspored) computing according to a June 25, 2024 news item on ScienceDaily,

Computers have come so far in terms of their power and potential, rivaling and even eclipsing human brains in their ability to store and crunch data, make predictions and communicate. But there is one domain where human brains continue to dominate: energy efficiency.

“The most efficient computers are still approximately four orders of magnitude — that’s 10,000 times — higher in energy requirements compared to the human brain for specific tasks such as image processing and recognition, although they outperform the brain in tasks like mathematical calculations,” said UC Santa Barbara electrical and computer engineering Professor Kaustav Banerjee, a world expert in the realm of nanoelectronics. “Making computers more energy efficient is crucial because the worldwide energy consumption by on-chip electronics stands at #4 in the global rankings of nation-wise energy consumption, and it is increasing exponentially each year, fueled by applications such as artificial intelligence.” Additionally, he said, the problem of energy inefficient computing is particularly pressing in the context of global warming, “highlighting the urgent need to develop more energy-efficient computing technologies.”

….

A June 24, 2024 UCSB news release (also on Eurekalert), which originated the news item, delves further into the subject,

Neuromorphic (NM) computing has emerged as a promising way to bridge the energy efficiency gap. By mimicking the structure and operations of the human brain, where processing occurs in parallel across an array of low power-consuming neurons, it may be possible to approach brain-like energy efficiency. In a paper published in the journal Nature Communications, Banerjee and co-workers Arnab Pal, Zichun Chai, Junkai Jiang and Wei Cao, in collaboration with researchers Vivek De and Mike Davies from Intel Labs propose such an ultra-energy efficient platform, using 2D transition metal dichalcogenide (TMD)-based tunnel-field-effect transistors (TFETs). Their platform, the researchers say, can bring the energy requirements to within two orders of magnitude (about 100 times) with respect to the human brain.

Leakage currents and subthreshold swing

The concept of neuromorphic computing has been around for decades, though the research around it has intensified only relatively recently. Advances in circuitry that enable smaller, denser arrays of transistors, and therefore more processing and functionality for less power consumption are just scratching the surface of what can be done to enable brain-inspired computing. Add to that an appetite generated by its many potential applications, such as AI and the Internet-of-Things, and it’s clear that expanding the options for a hardware platform for neuromorphic computing must be addressed in order to move forward.

Enter the team’s 2D tunnel-transistors. Emerging out of Banerjee’s longstanding research efforts to develop high-performance, low-power consumption transistors to meet the growing hunger for processing without a matching increase in power requirement, these atomically thin, nanoscale transistors are responsive at low voltages, and as the foundation of the researchers’ NM platform, can mimic the highly energy efficient operations of the human brain. In addition to lower off-state currents, the 2D TFETs also have a low subthreshold swing (SS), a parameter that describes how effectively a transistor can switch from off to on. According to Banerjee, a lower SS means a lower operating voltage, and faster and more efficient switching.

“Neuromorphic computing architectures are designed to operate with very sparse firing circuits,” said lead author Arnab Pal, “meaning they mimic how neurons in the brain fire only when necessary.” In contrast to the more conventional von Neumann architecture of today’s computers, in which data is processed sequentially, memory and processing components are separated and which continuously draw power throughout the entire operation, an event-driven system such as a NM computer fires up only when there is input to process, and memory and processing are distributed across an array of transistors. Companies like Intel and IBM have developed brain-inspired platforms, deploying billions of interconnected transistors and generating significant energy savings.

However, there’s still room for energy efficiency improvement, according to the researchers.

“In these systems, most of the energy is lost through leakage currents when the transistors are off, rather than during their active state,” Banerjee explained. A ubiquitous phenomenon in the world of electronics, leakage currents are small amounts of electricity that flow through a circuit even when it is in the off state (but still connected to power). According to the paper, current NM chips use traditional metal-oxide-semiconductor field-effect transistors (MOSFETs) which have a high on-state current, but also high off-state leakage. “Since the power efficiency of these chips is constrained by the off-state leakage, our approach — using tunneling transistors with much lower off-state current — can greatly improve power efficiency,” Banerjee said.

When integrated into a neuromorphic circuit, which emulates the firing and reset of neurons, the TFETs proved themselves more energy efficient than state-of-the-art MOSFETs, particularly the FinFETs (a MOSFET design that incorporates vertical “fins” as a way to provide better control of switching and leakage). TFETs are still in the experimental stage, however the performance and energy efficiency of neuromorphic circuits based on them makes them a promising candidate for the next generation of brain-inspired computing.

According to co-authors Vivek De (Intel Fellow) and Mike Davies (Director of Intel’s Neuromorphic Computing Lab), “Once realized, this platform can bring the energy consumption in chips to within two orders of magnitude with respect to the human brain — not accounting for the interface circuitry and memory storage elements. This represents a significant improvement from what is achievable today.”

Eventually, one can realize three-dimensional versions of these 2D-TFET based neuromorphic circuits to provide even closer emulation of the human brain, added Banerjee, widely recognized as one of the key visionaries behind 3D integrated circuits that are now witnessing wide scale commercial proliferation.

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

An ultra energy-efficient hardware platform for neuromorphic computing enabled by 2D-TMD tunnel-FETs by Arnab Pal, Zichun Chai, Junkai Jiang, Wei Cao, Mike Davies, Vivek De & Kaustav Banerjee. Nature Communications volume 15, Article number: 3392 (2024) DOI: https://doi.org/10.1038/s41467-024-46397-3 Published: 22 April 2024

This paper is open access.

New approach to brain-inspired (neuromorphic) computing: measuring information transfer

Leave a reply

An April 8, 2024 news item on Nanowerk announces a new approach to neuromorphic computing that involves measurement, Note: Links have been removed,

The biological brain, especially the human brain, is a desirable computing system that consumes little energy and runs at high efficiency. To build a computing system just as good, many neuromorphic scientists focus on designing hardware components intended to mimic the elusive learning mechanism of the brain. Recently, a research team has approached the goal from a different angle, focusing on measuring information transfer instead.

Their method went through biological and simulation experiments and then proved effective in an electronic neuromorphic system. It was published in Intelligent Computing (“Information Transfer in Neuronal Circuits: From Biological Neurons to Neuromorphic Electronics”).

…

An April 8, 2024 Intelligent Computing news release on EurekAlert delves further into the topic,

Although electronic systems have not fully replicated the complex information transfer between synapses and neurons, the team has demonstrated that it is possible to transform biological circuits into electronic circuits while maintaining the amount of information transferred. “This represents a key step toward brain-inspired low-power artificial systems,” the authors note.

To evaluate the efficiency of information transfer, the team drew inspiration from information theory. They quantified the amount of information conveyed by synapses in single neurons, then measured the quantity using mutual information, the analysis of which reveals the relationship between input stimuli and neuron responses.

First, the team conducted experiments with biological neurons. They used brain slices from rats, recording and analyzing the biological circuits in cerebellar granule cells. Then they evaluated the information transmitted at the synapses from mossy fiber neurons to the cerebellar granule cells. The mossy fibers were periodically stimulated with electrical spikes to induce synaptic plasticity, a fundamental biological feature where the information transfer at the synapses is constantly strengthened or weakened with repeated neuronal activity.

The results show that the changes in mutual information values are largely consistent with the changes in biological information transfer induced by synaptic plasticity. The findings from simulation and electronic neuromorphic experiments mirrored the biological results.

Second, the team conducted experiments with simulated neurons. They applied a spiking neural network model, which was developed by the same research group. Spiking neural networks were inspired by the functioning of biological neurons and are considered a promising approach for achieving efficient neuromorphic computing.

In the model, four mossy fibers are connected to one cerebellar granule cell, and each connection is given a random weight, which affects the information transfer efficiency like synaptic plasticity does in biological circuits. In the experiments, the team applied eight stimulation patterns to all mossy fibers and recorded the responses to evaluate the information transfer in the artificial neural network.

Third, the team conducted experiments with electronic neurons. A setup similar to those in the biological and simulation experiments was used. A previously developed semiconductor device functioned as a neuron, and four specialized memristors functioned as synapses. The team applied 20 spike sequences to decrease resistance values, then applied another 20 to increase them. The changes in resistance values were investigated to assess the information transfer efficiency within the neuromorphic system.

In addition to verifying the quantity of information transferred in biological, simulated and electronic neurons, the team also highlighted the importance of spike timing, which as they observed is closely related to information transfer. This observation could influence the development of neuromorphic computing, given that most devices are designed with spike-frequency-based algorithms.

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

Information Transfer in Neuronal Circuits: From Biological Neurons to Neuromorphic Electronics by Daniela Gandolfi, Lorenzo Benatti, Tommaso Zanotti, Giulia M. Boiani, Albertino Bigiani, Francesco M. Puglisi, and Jonathan Mapell. Intelligent Computing 1 Feb 2024 Vol 3 Article ID: 0059 DOI: 10.34133/icomputing.0059

This paper is open access.

Brain-inspired (neuromorphic) wireless system for gathering data from sensors the size of a grain of salt

Leave a reply

This is what a sensor the size of a grain of salt looks like,

Caption: The sensor network is designed so the chips can be implanted into the body or integrated into wearable devices. Each submillimeter-sized silicon sensor mimics how neurons in the brain communicate through spikes of electrical activity. Credit: Nick Dentamaro/Brown University

A March 19, 2024 news item on Nanowerk announces this research from Brown University (Rhode Island, US), Note: A link has been removed,

Tiny chips may equal a big breakthrough for a team of scientists led by Brown University engineers.

Writing in Nature Electronics (“An asynchronous wireless network for capturing event-driven data from large populations of autonomous sensors”), the research team describes a novel approach for a wireless communication network that can efficiently transmit, receive and decode data from thousands of microelectronic chips that are each no larger than a grain of salt.

…

One of the potential applications is for brain (neural) implants,

Caption: Writing in Nature Electronics, the research team describes a novel approach for a wireless communication network that can efficiently transmit, receive and decode data from thousands of microelectronic chips that are each no larger than a grain of salt. Credit: Nick Dentamaro/Brown University

A March 19, 2024 Brown University news release (also on EurekAlert), which originated the news item, provides more detail about the research, Note: Links have been removed,

…

The sensor network is designed so the chips can be implanted into the body or integrated into wearable devices. Each submillimeter-sized silicon sensor mimics how neurons in the brain communicate through spikes of electrical activity. The sensors detect specific events as spikes and then transmit that data wirelessly in real time using radio waves, saving both energy and bandwidth.

“Our brain works in a very sparse way,” said Jihun Lee, a postdoctoral researcher at Brown and study lead author. “Neurons do not fire all the time. They compress data and fire sparsely so that they are very efficient. We are mimicking that structure here in our wireless telecommunication approach. The sensors would not be sending out data all the time — they’d just be sending relevant data as needed as short bursts of electrical spikes, and they would be able to do so independently of the other sensors and without coordinating with a central receiver. By doing this, we would manage to save a lot of energy and avoid flooding our central receiver hub with less meaningful data.”

This radiofrequency [sic] transmission scheme also makes the system scalable and tackles a common problem with current sensor communication networks: they all need to be perfectly synced to work well.

The researchers say the work marks a significant step forward in large-scale wireless sensor technology and may one day help shape how scientists collect and interpret information from these little silicon devices, especially since electronic sensors have become ubiquitous as a result of modern technology.

“We live in a world of sensors,” said Arto Nurmikko, a professor in Brown’s School of Engineering and the study’s senior author. “They are all over the place. They’re certainly in our automobiles, they are in so many places of work and increasingly getting into our homes. The most demanding environment for these sensors will always be inside the human body.”

That’s why the researchers believe the system can help lay the foundation for the next generation of implantable and wearable biomedical sensors. There is a growing need in medicine for microdevices that are efficient, unobtrusive and unnoticeable but that also operate as part of a large ensembles to map physiological activity across an entire area of interest.

“This is a milestone in terms of actually developing this type of spike-based wireless microsensor,” Lee said. “If we continue to use conventional methods, we cannot collect the high channel data these applications will require in these kinds of next-generation systems.”

The events the sensors identify and transmit can be specific occurrences such as changes in the environment they are monitoring, including temperature fluctuations or the presence of certain substances.

The sensors are able to use as little energy as they do because external transceivers supply wireless power to the sensors as they transmit their data — meaning they just need to be within range of the energy waves sent out by the transceiver to get a charge. This ability to operate without needing to be plugged into a power source or battery make them convenient and versatile for use in many different situations.

The team designed and simulated the complex electronics on a computer and has worked through several fabrication iterations to create the sensors. The work builds on previous research from Nurmikko’s lab at Brown that introduced a new kind of neural interface system called “neurograins.” This system used a coordinated network of tiny wireless sensors to record and stimulate brain activity.

“These chips are pretty sophisticated as miniature microelectronic devices, and it took us a while to get here,” said Nurmikko, who is also affiliated with Brown’s Carney Institute for Brain Science. “The amount of work and effort that is required in customizing the several different functions in manipulating the electronic nature of these sensors — that being basically squeezed to a fraction of a millimeter space of silicon — is not trivial.”

The researchers demonstrated the efficiency of their system as well as just how much it could potentially be scaled up. They tested the system using 78 sensors in the lab and found they were able to collect and send data with few errors, even when the sensors were transmitting at different times. Through simulations, they were able to show how to decode data collected from the brains of primates using about 8,000 hypothetically implanted sensors.

The researchers say next steps include optimizing the system for reduced power consumption and exploring broader applications beyond neurotechnology.

“The current work provides a methodology we can further build on,” Lee said.

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

An asynchronous wireless network for capturing event-driven data from large populations of autonomous sensors by Jihun Lee, Ah-Hyoung Lee, Vincent Leung, Farah Laiwalla, Miguel Angel Lopez-Gordo, Lawrence Larson & Arto Nurmikko. Nature Electronics volume 7, pages 313–324 (2024) DOI: https://doi.org/10.1038/s41928-024-01134-y Published: 19 March 2024 Issue Date: April 2024

This paper is behind a paywall.

Prior to this, 2021 seems to have been a banner year for Nurmikko’s lab. There’s this August 12, 2021 Brown University news release touting publication of a then new study in Nature Electronics and I have an April 2, 2021 post, “BrainGate demonstrates a high-bandwidth wireless brain-computer interface (BCI),” touting an earlier 2021 published study from the lab.

Hardware policies best way to manage AI safety?

Leave a reply

Regulation of artificial intelligence (AI) has become very topical in the last couple of years. There was an AI safety summit in November 2023 at Bletchley Park in the UK (see my November 2, 2023 posting for more about that international meeting).

A very software approach?

This year (2024) has seen a rise in legislative and proposed legislative activity. I have some articles on a few of these activities. China was the first to enact regulations of any kind on AI according to Matt Sheehan’s February 27, 2024 paper for the Carnegie Endowment for International Peace,

In 2021 and 2022, China became the first country to implement detailed, binding regulations on some of the most common applications of artificial intelligence (AI). These rules formed the foundation of China’s emerging AI governance regime, an evolving policy architecture that will affect everything from frontier AI research to the functioning of the world’s second-largest economy, from large language models in Africa to autonomous vehicles in Europe.

…

The Chinese Communist Party (CCP) and the Chinese government started that process with the 2021 rules on recommendation algorithms, an omnipresent use of the technology that is often overlooked in international AI governance discourse. Those rules imposed new obligations on companies to intervene in content recommendations, granted new rights to users being recommended content, and offered protections to gig workers subject to algorithmic scheduling. The Chinese party-state quickly followed up with a new regulation on “deep synthesis,” the use of AI to generate synthetic media such as deepfakes. Those rules required AI providers to watermark AI-generated content and ensure that content does not violate people’s “likeness rights” or harm the “nation’s image.” Together, these two regulations also created and amended China’s algorithm registry, a regulatory tool that would evolve into a cornerstone of the country’s AI governance regime.

…

The UK has adopted a more generalized approach focused on encouraging innovation according to Valeria Gallo’s and Suchitra Nair’s February 21, 2024 article for Deloitte (a British professional services firm also considered one of the big four accounting firms worldwide),

At a glance

The UK Government has adopted a cross-sector and outcome-based framework for regulating AI, underpinned by five core principles. These are safety, security and robustness, appropriate transparency and explainability, fairness, accountability and governance, and contestability and redress.

Regulators will implement the framework in their sectors/domains by applying existing laws and issuing supplementary regulatory guidance. Selected regulators will publish their AI annual strategic plans by 30th April [2024], providing businesses with much-needed direction.

Voluntary safety and transparency measures for developers of highly capable AI models and systems will also supplement the framework and the activities of individual regulators.

The framework will not be codified into law for now, but the Government anticipates the need for targeted legislative interventions in the future. These interventions will address gaps in the current regulatory framework, particularly regarding the risks posed by complex General Purpose AI and the key players involved in its development.

Organisations must prepare for increased AI regulatory activity over the next year, including guidelines, information gathering, and enforcement. International firms will inevitably have to navigate regulatory divergence.

…

While most of the focus appears to be on the software (e.g., General Purpose AI), the UK framework does not preclude hardware.

The European Union (EU) is preparing to pass its own AI regulation act through the European Parliament in 2024 according to a December 19, 2023 “EU AI Act: first regulation on artificial intelligence” article update, Note: Links have been removed,

As part of its digital strategy, the EU wants to regulate artificial intelligence (AI) to ensure better conditions for the development and use of this innovative technology. AI can create many benefits, such as better healthcare; safer and cleaner transport; more efficient manufacturing; and cheaper and more sustainable energy.

In April 2021, the European Commission proposed the first EU regulatory framework for AI. It says that AI systems that can be used in different applications are analysed and classified according to the risk they pose to users. The different risk levels will mean more or less regulation.

The agreed text is expected to be finally adopted in April 2024. It will be fully applicable 24 months after entry into force, but some parts will be applicable sooner:

*The ban of AI systems posing unacceptable risks will apply six months after the entry into force

*Codes of practice will apply nine months after entry into force

*Rules on general-purpose AI systems that need to comply with transparency requirements will apply 12 months after the entry into force

High-risk systems will have more time to comply with the requirements as the obligations concerning them will become applicable 36 months after the entry into force.

…

This EU initiative, like the UK framework, seems largely focused on AI software and according to the Wikipedia entry “Regulation of artificial intelligence,”

… The AI Act is expected to come into effect in late 2025 or early 2026.[109

I do have a few postings about Canadian regulatory efforts, which also seem to be focused on software but don’t preclude hardware. While the January 20, 2024 posting is titled “Canada’s voluntary code of conduct relating to advanced generative AI (artificial intelligence) systems,” information about legislative efforts is also included although you might find my May 1, 2023 posting titled “Canada, AI regulation, and the second reading of the Digital Charter Implementation Act, 2022 (Bill C-27)” offers more comprehensive information about Canada’s legislative progress or lack thereof.

The US is always to be considered in these matters and I have a November 2023 ‘briefing’ by Müge Fazlioglu on the International Association of Privacy Professionals (IAPP) website where she provides a quick overview of the international scene before diving deeper into US AI governance policy through the Barack Obama, Donald Trump, and Joe Biden administrations. There’s also this January 29, 2024 US White House “Fact Sheet: Biden-⁠Harris Administration Announces Key AI Actions Following President Biden’s Landmark Executive Order.”

What about AI and hardware?

A February 15, 2024 news item on ScienceDaily suggests that regulating hardware may be the most effective way of regulating AI,

Chips and datacentres — the ‘compute’ power driving the AI revolution — may be the most effective targets for risk-reducing AI policies as they have to be physically possessed, according to a new report.

A global registry tracking the flow of chips destined for AI supercomputers is one of the policy options highlighted by a major new report calling for regulation of “compute” — the hardware that underpins all AI — to help prevent artificial intelligence misuse and disasters.

Other technical proposals floated by the report include “compute caps” — built-in limits to the number of chips each AI chip can connect with — and distributing a “start switch” for AI training across multiple parties to allow for a digital veto of risky AI before it feeds on data.

The experts point out that powerful computing chips required to drive generative AI models are constructed via highly concentrated supply chains, dominated by just a handful of companies — making the hardware itself a strong intervention point for risk-reducing AI policies.

The report, published 14 February [2024], is authored by nineteen experts and co-led by three University of Cambridge institutes — the Leverhulme Centre for the Future of Intelligence (LCFI), the Centre for the Study of Existential Risk (CSER) and the Bennett Institute for Public Policy — along with OpenAI and the Centre for the Governance of AI.

…

A February 14, 2024 University of Cambridge press release by Fred Lewsey (also on EurekAlert), which originated the news item, provides more information about the ‘hardware approach to AI regulation’,

“Artificial intelligence has made startling progress in the last decade, much of which has been enabled by the sharp increase in computing power applied to training algorithms,” said Haydn Belfield, a co-lead author of the report from Cambridge’s LCFI. 

“Governments are rightly concerned about the potential consequences of AI, and looking at how to regulate the technology, but data and algorithms are intangible and difficult to control.

“AI supercomputers consist of tens of thousands of networked AI chips hosted in giant data centres often the size of several football fields, consuming dozens of megawatts of power,” said Belfield.

“Computing hardware is visible, quantifiable, and its physical nature means restrictions can be imposed in a way that might soon be nearly impossible with more virtual elements of AI.”

The computing power behind AI has grown exponentially since the “deep learning era” kicked off in earnest, with the amount of “compute” used to train the largest AI models doubling around every six months since 2010. The biggest AI models now use 350 million times more compute than thirteen years ago.

Government efforts across the world over the past year – including the US Executive Order on AI, EU AI Act, China’s Generative AI Regulation, and the UK’s AI Safety Institute – have begun to focus on compute when considering AI governance.

Outside of China, the cloud compute market is dominated by three companies, termed “hyperscalers”: Amazon, Microsoft, and Google. “Monitoring the hardware would greatly help competition authorities in keeping in check the market power of the biggest tech companies, and so opening the space for more innovation and new entrants,” said co-author Prof Diane Coyle from Cambridge’s Bennett Institute. 

The report provides “sketches” of possible directions for compute governance, highlighting the analogy between AI training and uranium enrichment. “International regulation of nuclear supplies focuses on a vital input that has to go through a lengthy, difficult and expensive process,” said Belfield. “A focus on compute would allow AI regulation to do the same.”

Policy ideas are divided into three camps: increasing the global visibility of AI computing; allocating compute resources for the greatest benefit to society; enforcing restrictions on computing power.

For example, a regularly-audited international AI chip registry requiring chip producers, sellers, and resellers to report all transfers would provide precise information on the amount of compute possessed by nations and corporations at any one time.

The report even suggests a unique identifier could be added to each chip to prevent industrial espionage and “chip smuggling”.

“Governments already track many economic transactions, so it makes sense to increase monitoring of a commodity as rare and powerful as an advanced AI chip,” said Belfield. However, the team point out that such approaches could lead to a black market in untraceable “ghost chips”.

Other suggestions to increase visibility – and accountability – include reporting of large-scale AI training by cloud computing providers, and privacy-preserving “workload monitoring” to help prevent an arms race if massive compute investments are made without enough transparency.  

“Users of compute will engage in a mixture of beneficial, benign and harmful activities, and determined groups will find ways to circumvent restrictions,” said Belfield. “Regulators will need to create checks and balances that thwart malicious or misguided uses of AI computing.”

These might include physical limits on chip-to-chip networking, or cryptographic technology that allows for remote disabling of AI chips in extreme circumstances. One suggested approach would require the consent of multiple parties to unlock AI compute for particularly risky training runs, a mechanism familiar from nuclear weapons.

AI risk mitigation policies might see compute prioritised for research most likely to benefit society – from green energy to health and education. This could even take the form of major international AI “megaprojects” that tackle global issues by pooling compute resources.

The report’s authors are clear that their policy suggestions are “exploratory” rather than fully fledged proposals and that they all carry potential downsides, from risks of proprietary data leaks to negative economic impacts and the hampering of positive AI development.

They offer five considerations for regulating AI through compute, including the exclusion of small-scale and non-AI computing, regular revisiting of compute thresholds, and a focus on privacy preservation.

Added Belfield: “Trying to govern AI models as they are deployed could prove futile, like chasing shadows. Those seeking to establish AI regulation should look upstream to compute, the source of the power driving the AI revolution. If compute remains ungoverned it poses severe risks to society.”

You can find the report, “Computing Power and the Governance of Artificial Intelligence” on the University of Cambridge’s Centre for the Study of Existential Risk.

Authors include: Girish Sastry, Lennart Heim, Haydn Belfield, Markus Anderljung, Miles Brundage, Julian Hazell, Cullen O’Keefe, Gillian K. Hadfield, Richard Ngo, Konstantin Pilz, George Gor, Emma Bluemke, Sarah Shoker, Janet Egan, Robert F. Trager, Shahar Avin, Adrian Weller, Yoshua Bengio, and Diane Coyle.

The authors are associated with these companies/agencies: OpenAI, Centre for the Governance of AI (GovAI), Leverhulme Centre for the Future of Intelligence at the Uni. of Cambridge, Oxford Internet Institute, Institute for Law & AI, University of Toronto Vector Institute for AI, Georgetown University, ILINA Program, Harvard Kennedy School (of Government), *AI Governance Institute,* Uni. of Oxford, Centre for the Study of Existential Risk at Uni. of Cambridge, Uni. of Cambridge, Uni. of Montreal / Mila, Bennett Institute for Public Policy at the Uni. of Cambridge.

“The ILINIA program is dedicated to providing an outstanding platform for Africans to learn and work on questions around maximizing wellbeing and responding to global catastrophic risks” according to the organization’s homepage.

*As for the AI Governance Institute, I believe that should be the Centre for the Governance of AI at Oxford University since the associated academic is Robert F. Trager from the University of Oxford.

As the months (years?) fly by, I guess we’ll find out if this hardware approach gains any traction where AI regulation is concerned.

Butterfly mating inspires neuromorphic (brainlike) computing

Leave a reply

Michael Berger writes about a multisensory approach to neuromorphic computing inspired by butterflies in his February 2, 2024 Nanowerk Spotlight article, Note: Links have been removed,

Artificial intelligence systems have historically struggled to integrate and interpret information from multiple senses the way animals intuitively do. Humans and other species rely on combining sight, sound, touch, taste and smell to better understand their surroundings and make decisions. However, the field of neuromorphic computing has largely focused on processing data from individual senses separately.

This unisensory approach stems in part from the lack of miniaturized hardware able to co-locate different sensing modules and enable in-sensor and near-sensor processing. Recent efforts have targeted fusing visual and tactile data. However, visuochemical integration, which merges visual and chemical information to emulate complex sensory processing such as that seen in nature—for instance, butterflies integrating visual signals with chemical cues for mating decisions—remains relatively unexplored. Smell can potentially alter visual perception, yet current AI leans heavily on visual inputs alone, missing a key aspect of biological cognition.

Now, researchers at Penn State University have developed bio-inspired hardware that embraces heterogeneous integration of nanomaterials to allow the co-location of chemical and visual sensors along with computing elements. This facilitates efficient visuochemical information processing and decision-making, taking cues from the courtship behaviors of a species of tropical butterfly.

In the paper published in Advanced Materials (“A Butterfly-Inspired Multisensory Neuromorphic Platform for Integration of Visual and Chemical Cues”), the researchers describe creating their visuochemical integration platform inspired by Heliconius butterflies. During mating, female butterflies rely on integrating visual signals like wing color from males along with chemical pheromones to select partners. Specialized neurons combine these visual and chemical cues to enable informed mate choice.

To emulate this capability, the team constructed hardware encompassing monolayer molybdenum disulfide (MoS2) memtransistors serving as visual capture and processing components. Meanwhile, graphene chemitransistors functioned as artificial olfactory receptors. Together, these nanomaterials provided the sensing, memory and computing elements necessary for visuochemical integration in a compact architecture.

…

While mating butterflies served as inspiration, the developed technology has much wider relevance. It represents a significant step toward overcoming the reliance of artificial intelligence on single data modalities. Enabling integration of multiple senses can greatly improve situational understanding and decision-making for autonomous robots, vehicles, monitoring devices and other systems interacting with complex environments.

The work also helps progress neuromorphic computing approaches seeking to emulate biological brains for next-generation ML acceleration, edge deployment and reduced power consumption. In nature, cross-modal learning underpins animals’ adaptable behavior and intelligence emerging from brains organizing sensory inputs into unified percepts. This research provides a blueprint for hardware co-locating sensors and processors to more closely replicate such capabilities

…

It’s fascinating to me how many times butterflies inspire science,

Butterfly-inspired visuo-chemical integration. a) A simplified abstraction of visual and chemical stimuli from male butterflies and visuo-chemical integration pathway in female butterflies. b) Butterfly-inspired neuromorphic hardware comprising of monolayer MoS2 memtransistor-based visual afferent neuron, graphene-based chemoreceptor neuron, and MoS2 memtransistor-based neuro-mimetic mating circuits. Courtesy: Wiley/Penn State University Researchers

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

A Butterfly-Inspired Multisensory Neuromorphic Platform for Integration of Visual and Chemical Cues by Yikai Zheng, Subir Ghosh, Saptarshi Das. Advanced Materials SOI: https://doi.org/10.1002/adma.202307380 First published: 09 December 2023

This paper is open access.

Brainlike transistor and human intelligence

Leave a reply

This brainlike transistor (not a memristor) is important because it functions at room temperature as opposed to others, which require cryogenic temperatures.

A December 20, 2023 Northwestern University news release (received via email; also on EurekAlert) fills in the details,

  • Researchers develop transistor that simultaneously processes and stores information like the human brain
  • Transistor goes beyond categorization tasks to perform associative learning
  • Transistor identified similar patterns, even when given imperfect input
  • Previous similar devices could only operate at cryogenic temperatures; new transistor operates at room temperature, making it more practical

EVANSTON, Ill. — Taking inspiration from the human brain, researchers have developed a new synaptic transistor capable of higher-level thinking.

Designed by researchers at Northwestern University, Boston College and the Massachusetts Institute of Technology (MIT), the device simultaneously processes and stores information just like the human brain. In new experiments, the researchers demonstrated that the transistor goes beyond simple machine-learning tasks to categorize data and is capable of performing associative learning.

Although previous studies have leveraged similar strategies to develop brain-like computing devices, those transistors cannot function outside cryogenic temperatures. The new device, by contrast, is stable at room temperatures. It also operates at fast speeds, consumes very little energy and retains stored information even when power is removed, making it ideal for real-world applications.

The study was published today (Dec. 20 [2023]) in the journal Nature.

“The brain has a fundamentally different architecture than a digital computer,” said Northwestern’s Mark C. Hersam, who co-led the research. “In a digital computer, data move back and forth between a microprocessor and memory, which consumes a lot of energy and creates a bottleneck when attempting to perform multiple tasks at the same time. On the other hand, in the brain, memory and information processing are co-located and fully integrated, resulting in orders of magnitude higher energy efficiency. Our synaptic transistor similarly achieves concurrent memory and information processing functionality to more faithfully mimic the brain.”

Hersam is the Walter P. Murphy Professor of Materials Science and Engineering at Northwestern’s McCormick School of Engineering. He also is chair of the department of materials science and engineering, director of the Materials Research Science and Engineering Center and member of the International Institute for Nanotechnology. Hersam co-led the research with Qiong Ma of Boston College and Pablo Jarillo-Herrero of MIT.

Recent advances in artificial intelligence (AI) have motivated researchers to develop computers that operate more like the human brain. Conventional, digital computing systems have separate processing and storage units, causing data-intensive tasks to devour large amounts of energy. With smart devices continuously collecting vast quantities of data, researchers are scrambling to uncover new ways to process it all without consuming an increasing amount of power. Currently, the memory resistor, or “memristor,” is the most well-developed technology that can perform combined processing and memory function. But memristors still suffer from energy costly switching.

“For several decades, the paradigm in electronics has been to build everything out of transistors and use the same silicon architecture,” Hersam said. “Significant progress has been made by simply packing more and more transistors into integrated circuits. You cannot deny the success of that strategy, but it comes at the cost of high power consumption, especially in the current era of big data where digital computing is on track to overwhelm the grid. We have to rethink computing hardware, especially for AI and machine-learning tasks.”

To rethink this paradigm, Hersam and his team explored new advances in the physics of moiré patterns, a type of geometrical design that arises when two patterns are layered on top of one another. When two-dimensional materials are stacked, new properties emerge that do not exist in one layer alone. And when those layers are twisted to form a moiré pattern, unprecedented tunability of electronic properties becomes possible.

For the new device, the researchers combined two different types of atomically thin materials: bilayer graphene and hexagonal boron nitride. When stacked and purposefully twisted, the materials formed a moiré pattern. By rotating one layer relative to the other, the researchers could achieve different electronic properties in each graphene layer even though they are separated by only atomic-scale dimensions. With the right choice of twist, researchers harnessed moiré physics for neuromorphic functionality at room temperature.

“With twist as a new design parameter, the number of permutations is vast,” Hersam said. “Graphene and hexagonal boron nitride are very similar structurally but just different enough that you get exceptionally strong moiré effects.”

To test the transistor, Hersam and his team trained it to recognize similar — but not identical — patterns. Just earlier this month, Hersam introduced a new nanoelectronic device capable of analyzing and categorizing data in an energy-efficient manner, but his new synaptic transistor takes machine learning and AI one leap further.

“If AI is meant to mimic human thought, one of the lowest-level tasks would be to classify data, which is simply sorting into bins,” Hersam said. “Our goal is to advance AI technology in the direction of higher-level thinking. Real-world conditions are often more complicated than current AI algorithms can handle, so we tested our new devices under more complicated conditions to verify their advanced capabilities.”

First the researchers showed the device one pattern: 000 (three zeros in a row). Then, they asked the AI to identify similar patterns, such as 111 or 101. “If we trained it to detect 000 and then gave it 111 and 101, it knows 111 is more similar to 000 than 101,” Hersam explained. “000 and 111 are not exactly the same, but both are three digits in a row. Recognizing that similarity is a higher-level form of cognition known as associative learning.”

In experiments, the new synaptic transistor successfully recognized similar patterns, displaying its associative memory. Even when the researchers threw curveballs — like giving it incomplete patterns — it still successfully demonstrated associative learning.

“Current AI can be easy to confuse, which can cause major problems in certain contexts,” Hersam said. “Imagine if you are using a self-driving vehicle, and the weather conditions deteriorate. The vehicle might not be able to interpret the more complicated sensor data as well as a human driver could. But even when we gave our transistor imperfect input, it could still identify the correct response.”

The study, “Moiré synaptic transistor with room-temperature neuromorphic functionality,” was primarily supported by the National Science Foundation.

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

Moiré synaptic transistor with room-temperature neuromorphic functionality by Xiaodong Yan, Zhiren Zheng, Vinod K. Sangwan, Justin H. Qian, Xueqiao Wang, Stephanie E. Liu, Kenji Watanabe, Takashi Taniguchi, Su-Yang Xu, Pablo Jarillo-Herrero, Qiong Ma & Mark C. Hersam. Nature volume 624, pages 551–556 (2023) DOI: https://doi.org/10.1038/s41586-023-06791-1 Published online: 20 December 2023 Issue Date: 21 December 2023

This paper is behind a paywall.

Striking similarity between memory processing of artificial intelligence (AI) models and hippocampus of the human brain

Leave a reply

A December 18, 2023 news item on ScienceDaily shifted my focus from hardware to software when considering memory in brainlike (neuromorphic) computing,

An interdisciplinary team consisting of researchers from the Center for Cognition and Sociality and the Data Science Group within the Institute for Basic Science (IBS) [Korea] revealed a striking similarity between the memory processing of artificial intelligence (AI) models and the hippocampus of the human brain. This new finding provides a novel perspective on memory consolidation, which is a process that transforms short-term memories into long-term ones, in AI systems.

…

A November 28 (?), 2023 IBS press release (also on EurekAlert but published December 18, 2023, which originated the news item, describes how the team went about its research,

In the race towards developing Artificial General Intelligence (AGI), with influential entities like OpenAI and Google DeepMind leading the way, understanding and replicating human-like intelligence has become an important research interest. Central to these technological advancements is the Transformer model [Figure 1], whose fundamental principles are now being explored in new depth.

The key to powerful AI systems is grasping how they learn and remember information. The team applied principles of human brain learning, specifically concentrating on memory consolidation through the NMDA receptor in the hippocampus, to AI models.

The NMDA receptor is like a smart door in your brain that facilitates learning and memory formation. When a brain chemical called glutamate is present, the nerve cell undergoes excitation. On the other hand, a magnesium ion acts as a small gatekeeper blocking the door. Only when this ionic gatekeeper steps aside, substances are allowed to flow into the cell. This is the process that allows the brain to create and keep memories, and the gatekeeper’s (the magnesium ion) role in the whole process is quite specific.

The team made a fascinating discovery: the Transformer model seems to use a gatekeeping process similar to the brain’s NMDA receptor [see Figure 1]. This revelation led the researchers to investigate if the Transformer’s memory consolidation can be controlled by a mechanism similar to the NMDA receptor’s gating process.

In the animal brain, a low magnesium level is known to weaken memory function. The researchers found that long-term memory in Transformer can be improved by mimicking the NMDA receptor. Just like in the brain, where changing magnesium levels affect memory strength, tweaking the Transformer’s parameters to reflect the gating action of the NMDA receptor led to enhanced memory in the AI model. This breakthrough finding suggests that how AI models learn can be explained with established knowledge in neuroscience.

C. Justin LEE, who is a neuroscientist director at the institute, said, “This research makes a crucial step in advancing AI and neuroscience. It allows us to delve deeper into the brain’s operating principles and develop more advanced AI systems based on these insights.”

CHA Meeyoung, who is a data scientist in the team and at KAIST [Korea Advanced Institute of Science and Technology], notes, “The human brain is remarkable in how it operates with minimal energy, unlike the large AI models that need immense resources. Our work opens up new possibilities for low-cost, high-performance AI systems that learn and remember information like humans.”

What sets this study apart is its initiative to incorporate brain-inspired nonlinearity into an AI construct, signifying a significant advancement in simulating human-like memory consolidation. The convergence of human cognitive mechanisms and AI design not only holds promise for creating low-cost, high-performance AI systems but also provides valuable insights into the workings of the brain through AI models.

Fig. 1: (a) Diagram illustrating the ion channel activity in post-synaptic neurons. AMPA receptors are involved in the activation of post-synaptic neurons, while NMDA receptors are blocked by magnesium ions (Mg²⁺) but induce synaptic plasticity through the influx of calcium ions (Ca²⁺) when the post-synaptic neuron is sufficiently activated. (b) Flow diagram representing the computational process within the Transformer AI model. Information is processed sequentially through stages such as feed-forward layers, layer normalization, and self-attention layers. The graph depicting the current-voltage relationship of the NMDA receptors is very similar to the nonlinearity of the feed-forward layer. The input-output graph, based on the concentration of magnesium (α), shows the changes in the nonlinearity of the NMDA receptors. Courtesy: IBS

This research was presented at the 37th Conference on Neural Information Processing Systems (NeurIPS 2023) before being published in the proceedings, I found a PDF of the presentation and an early online copy of the paper before locating the paper in the published proceedings.

PDF of presentation: Transformer as a hippocampal memory consolidation model based on NMDAR-inspired nonlinearity

PDF copy of paper:

Transformer as a hippocampal memory consolidation model based on NMDAR-inspired nonlinearity by Dong-Kyum Kim, Jea Kwon, Meeyoung Cha, C. Justin Lee.

This paper was made available on OpenReview.net:

OpenReview is a platform for open peer review, open publishing, open access, open discussion, open recommendations, open directory, open API and open source.

It’s not clear to me if this paper is finalized or not and I don’t know if its presence on OpenReview constitutes publication.

Finally, the paper published in the proceedings,

Transformer as a hippocampal memory consolidation model based on NMDAR-inspired nonlinearity by Dong Kyum Kim, Jea Kwon, Meeyoung Cha, C. Justin Lee. Part of Advances in Neural Information Processing Systems 36 (NeurIPS 2023) Main Conference Track

This link will take you to the abstract, access the paper by clicking on the Paper tab.