A new spin on nanoscale computing

A nanoscale magnetic device that mimics the behaviour of neurons has been used to recognize audio signals. Such a device could be adapted to tackle tasks with greater efficiency than conventional computers. See Letter p.428

Neuromorphic (brain-like) computers offer many advantages over conventional systems, including energy efficiency, a high data-transfer speed and the ability to be trained. On page 428, Torrejon et al.¹ report one of the first nanoscale neuromorphic computers to perform a classification task — in this case, speech recognition. The core of the computer is a magnetic device called a spintronic oscillator that operates at gigahertz frequencies. Torrejon and colleagues' work is interesting not so much because of the application for speech recognition, the results of which are similar to those of other state-of-the-art technologies², but because of how the recognition is achieved.

How does a spintronic oscillator work? The device has a magnetization that can be thought of as an arrow that points in a particular direction. This direction can be regulated by applying an electrical current to the device — a state known as the equilibrium configuration. When the device is stimulated by a second electrical current (the input), the arrow begins to oscillate in a stable way, producing an oscillating voltage. Crucially, the device's response depends on the timing of the input. The arrow continues to oscillate until the input is removed, at which time the device returns to equilibrium.

Spintronic oscillators have a few key properties: they are tens to hundreds of nanometres in size; they are nonlinear (they can exhibit stable isolated oscillations); they can be analysed using signal-processing methods; and they produce analog, rather than digital, signals. Spintronic oscillators also have useful capabilities. For example, they can perform many distinct tasks simultaneously by combining (multiplexing) signals and they are capable of phase locking — a property that stabilizes the oscillations. The transistors used in conventional computing can be as small as spintronic oscillators, approaching the size of a single atom. However, a network of transistors that emulates the properties of a spintronic oscillator would be larger and more complex than the corresponding oscillator.

The approach of using oscillations for computations is based on biology. Recordings of electrical activity in the brain show that neurons transmit signals whose oscillations have a wide range of frequencies. Furthermore, biological rhythms operate on time scales ranging from milliseconds to months³. These oscillations are forms of analog information processing. One notable feature, which spintronic oscillators share, is that the oscillations are remarkably stable in the presence of noise and other perturbations.

In the 1940s and 1950s, the mathematician John von Neumann proposed using microwave-frequency oscillators for general-purpose computations⁴. By using one oscillation in voltage to represent '0' and the antiphase oscillation to represent '1', von Neumann showed that all arithmetic operations can be performed using simple electronic circuits called NAND gates. However, his proposal came immediately before the advent of transistors. Transistors took over the computing world because they are simple in design and, with ingenious engineering, can be interconnected to form complex switching circuits that perform the required arithmetic operations.

In the past decade, there has been an explosion in applications of artificial intelligence, machine learning and, in particular, 'deep' learning that require powerful computers to simulate massive artificial neural networks. At the same time, there have been concerns that transistors are reaching their limit in terms of size, functionality and cost effectiveness. New types of transistor and alternative technologies are being investigated throughout the computing industry, with the aim of producing ever-smaller computer circuits. Some researchers are revisiting von Neumann's ideas to use oscillators for arithmetic computation⁵, whereas others are developing computers based on quantum mechanics⁶. Torrejon and colleagues' work is the first step in a different direction — it suggests that spintronic oscillators could pave the way to building specialized chips for large-scale neural networks. The present era feels similar to that of 60 years ago, when transistors were first used to replace vacuum tubes in computing machines.

Torrejon et al. used an approach called reservoir computing, which is derived from studies of neural networks in the prefrontal cortex of primate brains⁷. In this approach, an input signal is fed into a computing system called a reservoir. Another computer is trained to read the state of the reservoir and map this state to the desired output.

The authors' reservoir is a spintronic oscillator comprising a non-magnetic material sandwiched between two magnetic layers (Fig. 1). As the input signal, the authors used an audio file of an isolated digit (0 to 9) pronounced by one of five different speakers. They then transformed the audio signal into an electrical current using signal-processing methods (the pre-processing stage). The current drives the oscillator, producing a voltage that measures the deflection of the magnetization from equilibrium. Finally, the authors identified the spoken digit (the output) from this voltage using machine-learning methods (the post-processing stage).

Figure 1: Spoken-digit recognition using a spintronic oscillator.

Torrejon et al.¹ show that a nanoscale magnetic device called a spintronic oscillator can be used for speech recognition. Their oscillator comprises a non-magnetic material (yellow), sandwiched between two magnetic materials (blue and grey). Shown here is a simplified version of their approach. The authors transform an audio signal for the word 'one' into an electrical current using signal-processing methods. The current causes the oscillator's magnetization (black arrow) to rotate (red arrow), producing an oscillating voltage. Torrejon and colleagues identify the spoken digit from this voltage using machine-learning methods, in which data are classified on the basis of the results of previous training. Unlike conventional electronics that would require a combination of several components and a larger microchip area, the authors' spintronic oscillator provides functionality in a single unit. Audio signal adapted from Fig. 2a in ref. 1.

Torrejon et al. achieve digit-recognition rates of up to 99.6%, independent of the speaker — a result that is competitive with other state-of-the-art technologies². Currently, the pre-processing of inputs and the post-processing of outputs rely on digital computation, so the authors' system is a hybrid digital–analog machine. The reservoir cannot be tuned during the recognition process, but the pre- and post-processing systems can be (for example, during training).

Neuromorphic computers might not become general-purpose computational machines. It is more likely that they will make up arrays of specialized computers that communicate and synchronize their activities — much like the brain does — but at speeds of gigahertz rather than hertz, and on length scales of nanometres rather than micrometres. Such computers could also be hybrids of digital and analog devices, thereby taking advantage of the strengths of both technologies.

A natural next step for the authors is to investigate networking of spintronic oscillators to design and build more-complex arrays that have greater functionality. Connections between such oscillators could be achieved using electrical or optical pathways, or through excitations called spin waves that propagate in a common magnetic medium. In addition, input and output processing might eventually reach the scale and functionality of spintronic oscillators. Torrejon and colleagues' system is a breakthrough in terms of using oscillators for computing. The system works, and it holds promise for major gains in classification, computation, control and switching.

¹