Two-dimensional materials-based probabilistic synapses and reconfigurable neurons for measuring inference uncertainty using Bayesian neural networks

Sebastian, Amritanand; Pendurthi, Rahul; Kozhakhmetov, Azimkhan; Trainor, Nicholas; Robinson, Joshua A.; Redwing, Joan M.; Das, Saptarshi

doi:10.1038/s41467-022-33699-7

Cited by 27 publications

(37 citation statements)

References 52 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Two types of mobile ionic species are primarily used in memristors: metal cations, such as those of Ag, Cu, and Ti, and anions, such as oxygen (or its vacancy) [ 76 , 77 , 78 , 79 ]. In the case of a metal cation film memory, metal ions are typically introduced into the host material from the electrode through an electroforming process, or the active layer itself is mixed in a stoichiometric ratio [ 80 , 81 ]. Various reported ion-mediated memristors and their rudimentary mechanisms are discussed below.…”

Section: Single Memristor Devicementioning

confidence: 99%

Synapse-Mimetic Hardware-Implemented Resistive Random-Access Memory for Artificial Neural Network

Seok

Son

Jathar

et al. 2023

Sensors

View full text Add to dashboard Cite

Memristors mimic synaptic functions in advanced electronics and image sensors, thereby enabling brain-inspired neuromorphic computing to overcome the limitations of the von Neumann architecture. As computing operations based on von Neumann hardware rely on continuous memory transport between processing units and memory, fundamental limitations arise in terms of power consumption and integration density. In biological synapses, chemical stimulation induces information transfer from the pre- to the post-neuron. The memristor operates as resistive random-access memory (RRAM) and is incorporated into the hardware for neuromorphic computing. Hardware composed of synaptic memristor arrays is expected to lead to further breakthroughs owing to their biomimetic in-memory processing capabilities, low power consumption, and amenability to integration; these aspects satisfy the upcoming demands of artificial intelligence for higher computational loads. Among the tremendous efforts toward achieving human-brain-like electronics, layered 2D materials have demonstrated significant potential owing to their outstanding electronic and physical properties, facile integration with other materials, and low-power computing. This review discusses the memristive characteristics of various 2D materials (heterostructures, defect-engineered materials, and alloy materials) used in neuromorphic computing for image segregation or pattern recognition. Neuromorphic computing, the most powerful artificial networks for complicated image processing and recognition, represent a breakthrough in artificial intelligence owing to their enhanced performance and lower power consumption compared with von Neumann architectures. A hardware-implemented CNN with weight control based on synaptic memristor arrays is expected to be a promising candidate for future electronics in society, offering a solution based on non-von Neumann hardware. This emerging paradigm changes the computing algorithm using entirely hardware-connected edge computing and deep neural networks.

show abstract

Section: Single Memristor Devicementioning

confidence: 99%

Synapse-Mimetic Hardware-Implemented Resistive Random-Access Memory for Artificial Neural Network

Seok

Son

Jathar

et al. 2023

Sensors

View full text Add to dashboard Cite

show abstract

“…Two recent publications have proposed new devices where the inherent probability distribution of their resistance can be tuned. These solutions, which exploit two-dimensional materials 24 and magnetic devices 25 , were validated with simulations of Bayesian neural networks.…”

Section: /16mentioning

confidence: 99%

“…• Each synapse is implemented using two memristors that are programmed independently, allowing the partial decorrelation of mean values and standard deviations of synaptic weights. This idea was presented in simulation studies 24,25 .…”

Section: /16mentioning

confidence: 99%

Bringing uncertainty quantification to the extreme-edge with memristor-based Bayesian neural networks

Bonnet

Hirtzlin

Majumdar

et al. 2023

Preprint

View full text Add to dashboard Cite

Safety-critical sensory processing applications, like medical diagnosis, require making accurate decisions based on a small amount of noisy input data. For these applications, using Bayesian neural networks, able to quantify the uncertainty of the predictions, is a superior approach to using conventional artificial neural networks. However, because of the probabilistic nature of Bayesian neural networks, they can be computationally intensive to use for inference stage and thus not well suited for extreme-edge applications. An emerging idea to solve this problem is to use the intrinsic probabilistic nature of memristors to efficiently implement Bayesian neural network inference: the variability in the resistance of memristors would represent the probability distribution of weights in Bayesian neural networks. However, when using memristors, statistical effects follow the laws of device physics, whereas in Bayesian neural networks, those effects can take on arbitrary shapes. In this work, we overcome this difficulty by adopting a dedicated synapse architecture based on two memristors, and by training Bayesian neural networks with a dedicated variational inference technique that includes a “technological loss” to take into account specificities of memristor physics. This technique allowed us to program a two-layer Bayesian neural network on 75 physical crossbar arrays of 1,024 memristors, incorporating CMOS periphery circuitry to do in-memory computing, to classify arrhythmia in electrocardiograms. Our experimental neural network classified heartbeats with high accuracy, and estimated the certainty of its predictions. In the case of uncertain predictions, it differentiated between ambivalent heartbeats (aleatoric uncertainty), and heartbeats with never-seen patterns (epistemic uncertainty). We show that our technique can also be used with phase change memories, by employing a different “technological loss” term. The great advantage of this approach is its low energy consumption: we estimate an 800 times improvement in energy efficiency compared to a GPU performing the same task.

show abstract

“…While earlier reports utilized exfoliated flakes, advancements in large area wafer-scale growth techniques over the years have strengthened their appeal for development of post-silicon era technological nodes [6][7][8]. Apart from providing the platform for developing electronics [9][10][11], optoelectronics [12,13] and sensing devices [14,15] at the scaling limits, certain TMDs have also found niche applications as electrode materials for batteries [16], energy storage devices [17], and electrocatalysts [18].…”

Section: Introductionmentioning

confidence: 99%

An Atomically Thin and Photosensitive Vanadium Disulfide Memtransistor

Das

Wali

Patel

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

Intrinsically ferromagnetic and semiconducting two-dimensional (2D) H-phase vanadium disulfide (VS2) holds tremendous promise for future electronics, optoelectronics, spintronics and valleytronics applications. However, its thermodynamic instability and the formation of intermediate stoichiometric polymorphs during its growth have stymied any progress towards synthesis of high quality 2D VS2 films. In this article, we circumvent these challenges and accomplish large area growth of monolayer VS2 films using atmospheric pressure chemical vapor deposition (APCVD) technique. By incorporating excess sulfur during the growth process which suppresses the formation of intermediate compounds, good quality large-area VS2 film can be synthesized. Furthermore, the electronic and optoelectronic properties of VS2 were explored by fabricating photosensitive memtransistor devices, which reveal an n-type carrier transport along with a high responsivity to red, green, and blue wavelengths of light. In addition the device exhibited multiple non-volatile conductance states through electrical programming. To the best of our knowledge, this is the first comprehensive report on memtransistors built from large area grown H-phase VS2 that integrate compute, sense, and storage functionalities in a single device.

show abstract

Two-dimensional materials-based probabilistic synapses and reconfigurable neurons for measuring inference uncertainty using Bayesian neural networks

Cited by 27 publications

References 52 publications

Synapse-Mimetic Hardware-Implemented Resistive Random-Access Memory for Artificial Neural Network

Synapse-Mimetic Hardware-Implemented Resistive Random-Access Memory for Artificial Neural Network

Bringing uncertainty quantification to the extreme-edge with memristor-based Bayesian neural networks

An Atomically Thin and Photosensitive Vanadium Disulfide Memtransistor

Contact Info

Product

Resources

About