Challenges in materials and devices for Resistive-Switching-based Neuromorphic Computing (1812.01120v1)

Abstract: This tutorial describes challenges and possible avenues for the implementation of the components of a solid-state system, which emulates a biological brain. The tutorial is devoted mostly to a charge-based (i.e. electric controlled) implementation using transition metal oxides materials, which exhibit unique properties that emulate key functionalities needed for this application. In the Introduction, we compare the main differences between a conventional computational machine, based on the Turing-von Neumann paradigm, to a Neuromorphic machine, which tries to emulate important functionalities of a biological brain. We also describe the main electrical properties of biological systems, which would be useful to implement in a charge-based system. In Chapter II, we describe the main components of a possible solid-state implementation. In Chapter III, we describe a variety of Resistive Switching phenomena, which may serve as the functional basis for the implementation of key devices for Neuromorphic computing. In Chapter IV we describe why transition metal oxides, are promising materials for future Neuromorphic machines. Theoretical models describing different resistive switching mechanisms are discussed in Chapter V while existing implementations are described in Chapter VI. Chapter VII presents applications to practical problems. We list in Chapter VIII important basic research challenges and open issues. We discuss issues related to specific implementations, novel materials, devices and phenomena. The development of reliable, fault tolerant, energy efficient devices, their scaling and integration into a Neuromorphic computer may bring us closer to the development of a machine that rivals the brain.

Overview of Resistive-Switching-based Neuromorphic Computing

The paper "Challenges in materials and devices for Resistive-Switching-based Neuromorphic Computing" authored by del Valle et al. provides a detailed examination of the potential and challenges in implementing neuromorphic computing systems using resistive-switching (RS) phenomena. This research addresses the fundamental differences in architecture between conventional Turing-von Neumann (TvN) computing systems and neuromorphic systems inspired by biological neural networks. The paper thoroughly explores transition metal oxides (TMOs) as promising materials for these neuromorphic systems due to their ability to exhibit both volatile and non-volatile resistive switching, crucial for simulating neuronal and synaptic functionalities.

Key Findings

The paper highlights three primary types of RS mechanisms: unipolar, bipolar, and volatile switching. Each mechanism is rigorously examined with a focus on its material basis, usually involving TMOs, and their unique electrical characteristics useful for emulating neurological processes such as synaptic plasticity and neuronal firing.

1. Unipolar and Bipolar Switching: In TMOs, unipolar and bipolar switching are attributed to the movement and redistribution of oxygen vacancies under an applied electric field. The paper highlights how these mechanisms can create RS behaviors suitable for synaptic devices in neuromorphic systems.
2. Volatile Switching: Volatile RS effects are linked to metal-insulator transitions typically observed within Mott insulators. These transitions allow materials to emulate the "leaky-integrate-and-fire" (LIF) characteristics of biological neurons.
3. Material Considerations: The research underscores TMOs for their versatility, noting their extensive range of non-volatile RS across numerous elemental oxides due to their multiple valence states and electronic configurations. This versatility is key in supporting a broad array of RS phenomena.

Practical and Theoretical Implications

From a practical perspective, implementing RS in neuromorphic systems promises significant reductions in energy consumption relative to conventional TvN systems, thus aligning with global efforts towards more sustainable computational technology. Theoretically, advancing our understanding of RS mechanisms across different material systems could pave the way for developing more efficient computational architectures that may someday rival biological brains.

The authors provide a comprehensive discussion of current implementations, such as using crossbar architectures for matrix operations central to pattern recognition and data clustering tasks. However, challenges remain in mitigating RS variability and achieving scalable integration with existing fabrication technologies.

Future Directions

The research sets a robust foundation for future exploration into RS-based neuromorphic computing. It encourages further inquiry into:

• The potential of integrating diverse functionalities through hybrid material systems combining electrical, thermal, and ionic switching properties.
• The scaling of charge-based neuromorphic systems and their comparison with other post-CMOS technologies like spintronics or superconducting devices.
• Addressing architecture-related challenges such as minimizing thermal dissipation and achieving efficient three-dimensional interconnectivity.

In conclusion, while the paper does not claim a near-term completion of a fully functional neuromorphic computing system, it robustly demonstrates the feasibility of RS phenomena as a serious candidate for future technological breakthroughs in artificial intelligence and cognitive computing applications. The exploration into TMOs and other novel materials continues to be a promising domain in the quest to emulate the incredible efficiency and functionality of the human brain.