Abstract
In our brain, information is exchanged among neurons in the form of spikes where both the space (which neuron fires) and time (when the neuron fires) contain relevant information. Every neuron is connected to other neurons by synapses, which are continuously created, updated, and stimulated to enable information processing and learning. Realizing the brain-like neuron/synapse network in silicon would enable artificial autonomous agents capable of learning, adaptation, and interaction with the environment. Toward this aim, the conventional microelectronic technology, which is based on complementary metal–oxide–semiconductor transistors and the von Neumann computing architecture, does not provide the desired energy efficiency and scaling potential. A generation of emerging memory devices, including resistive switching random access memory (RRAM) also known as the memristor, can offer a wealth of physics-enabled processing capabilities, including multiplication, integration, potentiation, depression, and time-decaying stimulation, which are suitable to recreate some of the fundamental phenomena of the human brain in silico. This work provides an overview about the status and the most recent updates on brain-inspired neuromorphic computing devices. After introducing the RRAM device technologies, we discuss the main computing functionalities of the human brain, including neuron integration and fire, dendritic filtering, and short- and long-term synaptic plasticity. For each of these processing functions, we discuss their proposed implementation in terms of materials, device structure, and brain-like characteristics. The rich device physics, the nano-scale integration, the tolerance to stochastic variations, and the ability to process information in situ make the emerging memory devices a promising technology for future brain-like hardware intelligence.
Highlights
The human brain is one of the most complex objects in the universe
Once the local graded potential (LGP) reaches the threshold after a certain number of pulses, the volatile resistive switching random access memory (RRAM) device switches to the high-conductance state, resulting in a fire output signal with high current
The results indicate long-term depression (LTD) for Δt < 0 and Long-term potentiation (LTP) for Δt > 0, where the change of conductance tends to vanish at increasing ∣Δt∣, which is in line with the observed biological spike-timing dependent plasticity (STDP).[7]
Summary
The human brain is one of the most complex objects in the universe. It is capable of executing high-level cognitive tasks, such as abstraction, generalization, prediction, decision making, recognition, and navigation in a continuously changing environment. Many research efforts aim at mimicking the type of computation of the human brain to achieve its outstanding energy efficiency This is the objective of neuromorphic engineering, where spiking neural networks (SNNs) are developed with artificial neurons and synapses. Time-dependent functions such as spike integration in an artificial neuron generally require large capacitors in CMOS technology, limiting the cost effectiveness of neuromorphic circuits.[5] Synaptic weights are generally stored in static random access memory (SRAM), which are volatile, i.e., all synaptic values are lost when the circuit is switched off.[6] In addition, SRAM devices are large and binary, i.e., they can only store 0 and 1; they are not suitable for gradual potentiation and depression that are typical of synaptic plasticity phenomena.[7,8,9]. V will focus on artificial synapses including learning functions via plasticity and sensing/computation via short-term memory (STM)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
