Tag Archives: A.V. Emelyanov

Memristor-based neural network and the biosimilar principle of learning

Leave a reply

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.

Plastic memristors for neural networks

Leave a reply

There is a very nice explanation of memristors and computing systems from the Moscow Institute of Physics and Technology (MIPT). First their announcement, from a Jan. 27, 2016 news item on ScienceDaily,

A group of scientists has created a neural network based on polymeric memristors — devices that can potentially be used to build fundamentally new computers. These developments will primarily help in creating technologies for machine vision, hearing, and other machine sensory systems, and also for intelligent control systems in various fields of applications, including autonomous robots.

The authors of the new study focused on a promising area in the field of memristive neural networks – polymer-based memristors – and discovered that creating even the simplest perceptron is not that easy. In fact, it is so difficult that up until the publication of their paper in the journal Organic Electronics, there were no reports of any successful experiments (using organic materials). The experiments conducted at the Nano-, Bio-, Information and Cognitive Sciences and Technologies (NBIC) centre at the Kurchatov Institute by a joint team of Russian and Italian scientists demonstrated that it is possible to create very simple polyaniline-based neural networks. Furthermore, these networks are able to learn and perform specified logical operations.

A Jan. 27, 2016 MIPT press release on EurekAlert, which originated the news item, offers an explanation of memristors and a description of the research,

A memristor is an electric element similar to a conventional resistor. The difference between a memristor and a traditional element is that the electric resistance in a memristor is dependent on the charge passing through it, therefore it constantly changes its properties under the influence of an external signal: a memristor has a memory and at the same time is also able to change data encoded by its resistance state! In this sense, a memristor is similar to a synapse – a connection between two neurons in the brain that is able, with a high level of plasticity, to modify the efficiency of signal transmission between neurons under the influence of the transmission itself. A memristor enables scientists to build a “true” neural network, and the physical properties of memristors mean that at the very minimum they can be made as small as conventional chips.

Some estimates indicate that the size of a memristor can be reduced up to ten nanometers, and the technologies used in the manufacture of the experimental prototypes could, in theory, be scaled up to the level of mass production. However, as this is “in theory”, it does not mean that chips of a fundamentally new structure with neural networks will be available on the market any time soon, even in the next five years.

The plastic polyaniline was not chosen by chance. Previous studies demonstrated that it can be used to create individual memristors, so the scientists did not have to go through many different materials. Using a polyaniline solution, a glass substrate, and chromium electrodes, they created a prototype with dimensions that, at present, are much larger than those typically used in conventional microelectronics: the strip of the structure was approximately one millimeter wide (they decided to avoid miniaturization for the moment). All of the memristors were tested for their electrical characteristics: it was found that the current-voltage characteristic of the devices is in fact non-linear, which is in line with expectations. The memristors were then connected to a single neuromorphic network.

A current-voltage characteristic (or IV curve) is a graph where the horizontal axis represents voltage and the vertical axis the current. In conventional resistance, the IV curve is a straight line; in strict accordance with Ohm’s Law, current is proportional to voltage. For a memristor, however, it is not just the voltage that is important, but the change in voltage: if you begin to gradually increase the voltage supplied to the memristor, it will increase the current passing through it not in a linear fashion, but with a sharp bend in the graph and at a certain point its resistance will fall sharply.

Then if you begin to reduce the voltage, the memristor will remain in its conducting state for some time, after which it will change its properties rather sharply again to decrease its conductivity. Experimental samples with a voltage increase of 0.5V hardly allowed any current to pass through (around a few tenths of a microamp), but when the voltage was reduced by the same amount, the ammeter registered a figure of 5 microamps. Microamps are of course very small units, but in this case it is the contrast that is most significant: 0.1 μA to 5 μA is a difference of fifty times! This is more than enough to make a clear distinction between the two signals.

After checking the basic properties of individual memristors, the physicists conducted experiments to train the neural network. The training (it is a generally accepted term and is therefore written without inverted commas) involves applying electric pulses at random to the inputs of a perceptron. If a certain combination of electric pulses is applied to the inputs of a perceptron (e.g. a logic one and a logic zero at two inputs) and the perceptron gives the wrong answer, a special correcting pulse is applied to it, and after a certain number of repetitions all the internal parameters of the device (namely memristive resistance) reconfigure themselves, i.e. they are “trained” to give the correct answer.

The scientists demonstrated that after about a dozen attempts their new memristive network is capable of performing NAND logical operations, and then it is also able to learn to perform NOR operations. Since it is an operator or a conventional computer that is used to check for the correct answer, this method is called the supervised learning method.

Needless to say, an elementary perceptron of macroscopic dimensions with a characteristic reaction time of tenths or hundredths of a second is not an element that is ready for commercial production. However, as the researchers themselves note, their creation was made using inexpensive materials, and the reaction time will decrease as the size decreases: the first prototype was intentionally enlarged to make the work easier; it is physically possible to manufacture more compact chips. In addition, polyaniline can be used in attempts to make a three-dimensional structure by placing the memristors on top of one another in a multi-tiered structure (e.g. in the form of random intersections of thin polymer fibers), whereas modern silicon microelectronic systems, due to a number of technological limitations, are two-dimensional. The transition to the third dimension would potentially offer many new opportunities.

The press release goes to explain what the researchers mean when they mention a fundamentally different computer,

The common classification of computers is based either on their casing (desktop/laptop/tablet), or on the type of operating system used (Windows/MacOS/Linux). However, this is only a very simple classification from a user perspective, whereas specialists normally use an entirely different approach – an approach that is based on the principle of organizing computer operations. The computers that we are used to, whether they be tablets, desktop computers, or even on-board computers on spacecraft, are all devices with von Neumann architecture; without going into too much detail, they are devices based on independent processors, random access memory (RAM), and read only memory (ROM).

The memory stores the code of a program that is to be executed. A program is a set of instructions that command certain operations to be performed with data. Data are also stored in the memory* and are retrieved from it (and also written to it) in accordance with the program; the program’s instructions are performed by the processor. There may be several processors, they can work in parallel, data can be stored in a variety of ways – but there is always a fundamental division between the processor and the memory. Even if the computer is integrated into one single chip, it will still have separate elements for processing information and separate units for storing data. At present, all modern microelectronic systems are based on this particular principle and this is partly the reason why most people are not even aware that there may be other types of computer systems – without processors and memory.

*) if physically different elements are used to store data and store a program, the computer is said to be built using Harvard architecture. This method is used in certain microcontrollers, and in small specialized computing devices. The chip that controls the function of a refrigerator, lift, or car engine (in all these cases a “conventional” computer would be redundant) is a microcontroller. However, neither Harvard, nor von Neumann architectures allow the processing and storage of information to be combined into a single element of a computer system.

However, such systems do exist. Furthermore, if you look at the brain itself as a computer system (this is purely hypothetical at the moment: it is not yet known whether the function of the brain is reducible to computations), then you will see that it is not at all built like a computer with von Neumann architecture. Neural networks do not have a specialized computer or separate memory cells. Information is stored and processed in each and every neuron, one element of the computer system, and the human brain has approximately 100 billion of these elements. In addition, almost all of them are able to work in parallel (simultaneously), which is why the brain is able to process information with great efficiency and at such high speed. Artificial neural networks that are currently implemented on von Neumann computers only emulate these processes: emulation, i.e. step by step imitation of functions inevitably leads to a decrease in speed and an increase in energy consumption. In many cases this is not so critical, but in certain cases it can be.

Devices that do not simply imitate the function of neural networks, but are fundamentally the same could be used for a variety of tasks. Most importantly, neural networks are capable of pattern recognition; they are used as a basis for recognising handwritten text for example, or signature verification. When a certain pattern needs to be recognised and classified, such as a sound, an image, or characteristic changes on a graph, neural networks are actively used and it is in these fields where gaining an advantage in terms of speed and energy consumption is critical. In a control system for an autonomous flying robot every milliwatt-hour and every millisecond counts, just in the same way that a real-time system to process data from a collider detector cannot take too long to “think” about highlighting particle tracks that may be of interest to scientists from among a large number of other recorded events.

Bravo to the writer!

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

Hardware elementary perceptron based on polyaniline memristive devices by V.A. Demin. V. V. Erokhin, A.V. Emelyanov, S. Battistoni, G. Baldi, S. Iannotta, P.K. Kashkarov, M.V. Kovalchuk. Organic Electronics Volume 25, October 2015, Pages 16–20 doi:10.1016/j.orgel.2015.06.015

This paper is behind a paywall.