Memristive systems offer biomimetic functions that are being actively explored for energy‐efficient neuromorphic circuits. In addition to providing ultimate geometric scaling limits, 2D semiconductors enable unique gate‐tunable responses including the recent realization of hybrid memristor and transistor devices known as memtransistors. In particular, monolayer MoS2memtransistors exhibit nonvolatile memristive switching where the resistance of each state is modulated by a gate terminal. Here, further control over the memtransistor neuromorphic response through the introduction of a second gate terminal is gained. The resulting dual‐gated memtransistors allow tunability over the learning rate for non‐Hebbian training where the long‐term potentiation and depression synaptic behavior is dictated by gate biases during the reading and writing processes. Furthermore, the electrostatic control provided by dual gates provides a compact solution to the sneak current problem in traditional memristor crossbar arrays. In this manner, dual gating facilitates the full utilization and integration of memtransistor functionality in highly scaled crossbar circuits. Furthermore, the tunability of long‐term potentiation yields improved linearity and symmetry of weight update rules that are utilized in simulated artificial neural networks to achieve a 94% recognition rate of hand‐written digits.
Higher order neural processing with input-adaptive dynamic weights on MoS2 memtransistor crossbars
The increasing complexity of deep learning systems has pushed conventional computing technologies to their limits. While the memristor is one of the prevailing technologies for deep learning acceleration, it is only suited for classical learning layers where only two operands, namely weights and inputs, are processed simultaneously. Meanwhile, to improve the computational efficiency of deep learning for emerging applications, a variety of non-traditional layers requiring concurrent processing of many operands are becoming popular. For example, hypernetworks improve their predictive robustness by simultaneously processing weights and inputs against the application context. Two-electrode memristor grids cannot directly map emerging layers’ higher-order multiplicative neural interactions. Addressing this unmet need, we present crossbar processing using dual-gated memtransistors based on two-dimensional semiconductor MoS 2 . Unlike the memristor, the resistance states of memtransistors can be persistently programmed and can be actively controlled by multiple gate electrodes. Thus, the discussed memtransistor crossbar enables several advanced inference architectures beyond a conventional passive crossbar. For example, we show that sneak paths can be effectively suppressed in memtransistor crossbars, whereas they limit size scalability in a passive memristor crossbar. Similarly, exploiting gate terminals to suppress crossbar weights dynamically reduces biasing power by ∼20% in memtransistor crossbars for a fully connected layer of AlexNet. On emerging layers such as hypernetworks, collocating multiple operations within the same crossbar cells reduces operating power by ∼ 15 × on the considered network cases.
more »
« less
- NSF-PAR ID:
- 10350571
- Date Published:
- Journal Name:
- Frontiers in Electronic Materials
- Volume:
- 2
- ISSN:
- 2673-9895
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
more » « less
Abstract -
more » « less
Network features found in the brain may help implement more efficient and robust neural networks. Spiking neural networks (SNNs) process spikes in the spatiotemporal domain and can offer better energy efficiency than deep neural networks. However, most SNN implementations rely on simple point neurons that neglect the rich neuronal and dendritic dynamics. Herein, a bio‐inspired columnar learning network (CLN) structure that employs feedforward, lateral, and feedback connections to make robust classification with sparse data is proposed. CLN is inspired by the mammalian neocortex, comprising cortical columns each containing multiple minicolumns formed by interacting pyramidal neurons. A column continuously processes spatiotemporal signals from its sensor, while learning spatial and temporal correlations between features in different regions of an object along with the sensor's movement through sensorimotor interaction. CLN can be implemented using memristor crossbars with a local learning rule, spiking timing‐dependent plasticity (STDP), which can be natively obtained in second‐order memristors. CLN allows inputs from multiple sensors to be simultaneously processed by different columns, resulting in higher classification accuracy and better noise tolerance. Analysis of networks implemented on memristor crossbars shows that the system can operate at very low power and high throughput, with high accuracy and robustness to noise.
-
Abstract Artificial neural networks have demonstrated superiority over traditional computing architectures in tasks such as pattern classification and learning. However, they do not measure uncertainty in predictions, and hence they can make wrong predictions with high confidence, which can be detrimental for many mission-critical applications. In contrast, Bayesian neural networks (BNNs) naturally include such uncertainty in their model, as the weights are represented by probability distributions (e.g. Gaussian distribution). Here we introduce three-terminal memtransistors based on two-dimensional (2D) materials, which can emulate both probabilistic synapses as well as reconfigurable neurons. The cycle-to-cycle variation in the programming of the 2D memtransistor is exploited to achieve Gaussian random number generator-based synapses, whereas 2D memtransistor based integrated circuits are used to obtain neurons with hyperbolic tangent and sigmoid activation functions. Finally, memtransistor-based synapses and neurons are combined in a crossbar array architecture to realize a BNN accelerator for a data classification task.more » « less
-
BACKGROUND Optical sensing devices measure the rich physical properties of an incident light beam, such as its power, polarization state, spectrum, and intensity distribution. Most conventional sensors, such as power meters, polarimeters, spectrometers, and cameras, are monofunctional and bulky. For example, classical Fourier-transform infrared spectrometers and polarimeters, which characterize the optical spectrum in the infrared and the polarization state of light, respectively, can occupy a considerable portion of an optical table. Over the past decade, the development of integrated sensing solutions by using miniaturized devices together with advanced machine-learning algorithms has accelerated rapidly, and optical sensing research has evolved into a highly interdisciplinary field that encompasses devices and materials engineering, condensed matter physics, and machine learning. To this end, future optical sensing technologies will benefit from innovations in device architecture, discoveries of new quantum materials, demonstrations of previously uncharacterized optical and optoelectronic phenomena, and rapid advances in the development of tailored machine-learning algorithms. ADVANCES Recently, a number of sensing and imaging demonstrations have emerged that differ substantially from conventional sensing schemes in the way that optical information is detected. A typical example is computational spectroscopy. In this new paradigm, a compact spectrometer first collectively captures the comprehensive spectral information of an incident light beam using multiple elements or a single element under different operational states and generates a high-dimensional photoresponse vector. An advanced algorithm then interprets the vector to achieve reconstruction of the spectrum. This scheme shifts the physical complexity of conventional grating- or interference-based spectrometers to computation. Moreover, many of the recent developments go well beyond optical spectroscopy, and we discuss them within a common framework, dubbed “geometric deep optical sensing.” The term “geometric” is intended to emphasize that in this sensing scheme, the physical properties of an unknown light beam and the corresponding photoresponses can be regarded as points in two respective high-dimensional vector spaces and that the sensing process can be considered to be a mapping from one vector space to the other. The mapping can be linear, nonlinear, or highly entangled; for the latter two cases, deep artificial neural networks represent a natural choice for the encoding and/or decoding processes, from which the term “deep” is derived. In addition to this classical geometric view, the quantum geometry of Bloch electrons in Hilbert space, such as Berry curvature and quantum metrics, is essential for the determination of the polarization-dependent photoresponses in some optical sensors. In this Review, we first present a general perspective of this sensing scheme from the viewpoint of information theory, in which the photoresponse measurement and the extraction of light properties are deemed as information-encoding and -decoding processes, respectively. We then discuss demonstrations in which a reconfigurable sensor (or an array thereof), enabled by device reconfigurability and the implementation of neural networks, can detect the power, polarization state, wavelength, and spatial features of an incident light beam. OUTLOOK As increasingly more computing resources become available, optical sensing is becoming more computational, with device reconfigurability playing a key role. On the one hand, advanced algorithms, including deep neural networks, will enable effective decoding of high-dimensional photoresponse vectors, which reduces the physical complexity of sensors. Therefore, it will be important to integrate memory cells near or within sensors to enable efficient processing and interpretation of a large amount of photoresponse data. On the other hand, analog computation based on neural networks can be performed with an array of reconfigurable devices, which enables direct multiplexing of sensing and computing functions. We anticipate that these two directions will become the engineering frontier of future deep sensing research. On the scientific frontier, exploring quantum geometric and topological properties of new quantum materials in both linear and nonlinear light-matter interactions will enrich the information-encoding pathways for deep optical sensing. In addition, deep sensing schemes will continue to benefit from the latest developments in machine learning. Future highly compact, multifunctional, reconfigurable, and intelligent sensors and imagers will find applications in medical imaging, environmental monitoring, infrared astronomy, and many other areas of our daily lives, especially in the mobile domain and the internet of things. Schematic of deep optical sensing. The n -dimensional unknown information ( w ) is encoded into an m -dimensional photoresponse vector ( x ) by a reconfigurable sensor (or an array thereof), from which w′ is reconstructed by a trained neural network ( n ′ = n and w′ ≈ w ). Alternatively, x may be directly deciphered to capture certain properties of w . Here, w , x , and w′ can be regarded as points in their respective high-dimensional vector spaces ℛ n , ℛ m , and ℛ n ′ .more » « less
-
Emerging resistive random-access memory (ReRAM) has recently been intensively investigated to accelerate the processing of deep neural networks (DNNs). Due to the in-situ computation capability, analog ReRAM crossbars yield significant throughput improvement and energy reduction compared to traditional digital methods. However, the power hungry analog-to-digital converters (ADCs) prevent the practical deployment of ReRAM-based DNN accelerators on end devices with limited chip area and power budget. We observe that due to the limited bitdensity of ReRAM cells, DNN weights are bit sliced and correspondingly stored on multiple ReRAM bitlines. The accumulated current on bitlines resulted by weights directly dictates the overhead of ADCs. As such, bitwise weight sparsity rather than the sparsity of the full weight, is desirable for efficient ReRAM deployment. In this work, we propose bit-slice `1, the first algorithm to induce bit-slice sparsity during the training of dynamic fixed-point DNNs. Experiment results show that our approach achieves 2 sparsity improvement compared to previous algorithms. The resulting sparsity allows the ADC resolution to be reduced to 1-bit of the most significant bit-slice and down to 3-bit for the others bits, which significantly speeds up processing and reduces power and area overhead.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.3389/femat.2022.950487
