Neuromorphic hardware implementation of Boltzmann Machine using a network of stochastic neurons can allow non-deterministic polynomial-time (NP) hard combinatorial optimization problems to be efficiently solved. Efficient implementation of such Boltzmann Machine with simulated annealing desires the statistical parameters of the stochastic neurons to be dynamically tunable, however, there has been limited research on stochastic semiconductor devices with controllable statistical distributions. Here, we demonstrate a reconfigurable tin oxide (SnOx)/molybdenum disulfide (MoS2) heterogeneous memristive device that can realize tunable stochastic dynamics in its output sampling characteristics. The device can sample exponential-class sigmoidal distributions analogous to the Fermi-Dirac distribution of physical systems with quantitatively defined tunable “temperature” effect. A BM composed of these tunable stochastic neuron devices, which can enable simulated annealing with designed “cooling” strategies, is conducted to solve the MAX-SAT, a representative in NP-hard combinatorial optimization problems. Quantitative insights into the effect of different “cooling” strategies on improving the BM optimization process efficiency are also provided.
Ferroelectric tunnel junctions (FTJs) have been intensively explored for future low power data storage and information processing applications. Among various ferroelectric (FE) materials studied, HfO2 and H0.5Zr0.5O2 (HZO) have the advantage of CMOS process compatibility. The validity of the simple effective mass approximation, for describing the tunneling process in these materials, is examined by computing the complex band structure from ab initio simulations. The results show that the simple effective mass approximation is insufficient to describe the tunneling current in HfO2 and HZO materials, and quantitative accurate descriptions of the complex band structures are indispensable for calculation of the tunneling current. A compact k · p Hamiltonian is parameterized to and validated by ab initio complex band structures, which provides a method for efficiently and accurately computing the tunneling current in HfO2 and HZO. The device characteristics of a metal/FE/metal structure and a metal/FE/semiconductor (M-F-S) structure are investigated by using the non-equilibrium Green's function formalism with the parameterized effective Hamiltonian. The result shows that the M-F-S structure offers a larger resistance window due to an extra barrier in the semiconductor region at off-state. A FTJ utilizing M-F-S structure is beneficial for memory design.
Artificial neuronal devices that
functionally resemble biological
neurons are important toward realizing advanced brain emulation and
for building bioinspired electronic systems. In this Communication,
the stochastic behaviors of a neuronal oscillator based on the charge-density-wave
(CDW) phase transition of a 1T-TaS2 thin film are reported,
and the capability of this neuronal oscillator to generate spike trains
with statistical features closely matching those of biological neurons
is demonstrated. The stochastic behaviors of the neuronal device result
from the melt-quench-induced reconfiguration of CDW domains during
each oscillation cycle. Owing to the stochasticity, numerous key features
of the Hodgkin-Huxley description of neurons can be realized in this
compact two-terminal neuronal oscillator. A statistical analysis of
the spike train generated by the artificial neuron indicates that
it resembles the neurons in the superior olivary complex of a mammalian
nervous system, in terms of its interspike interval distribution,
the time-correlation of spiking behavior, and its response to acoustic
stimuli.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.