Patrick Sheridan

Complementary resistive switching in tantalum oxide-based resistive memory devices

Complementary resistive switches (CRS) are considered as a potential solution for the sneak path problem in large-scale integration of passive crossbar resistive memory arrays. A typical CRS is composed of two bipolar memory cells that are connected anti-serially. Here, we report a tantalum-oxide based resistive memory that achieves the complementary switching functionality within a single memory cell. The complementary switching effect is accompanied by switching polarity reversal in different voltage bias regimes. These effects were explained by the redistribution of oxygen vacancies inside the tantalum-oxide layers. The effects of symmetry breaking on bipolar switching and complementary switching were also discussed.

Experimental Demonstration of a Second-Order Memristor and Its Ability to Biorealistically Implement Synaptic Plasticity

Memristors have been extensively studied for data storage and low-power computation applications. In this study, we show that memristors offer more than simple resistance change. Specifically, the dynamic evolutions of internal state variables allow an oxide-based memristor to exhibit Ca(2+)-like dynamics that natively encode timing information and regulate synaptic weights. Such a device can be modeled as a second-order memristor and allow the implementation of critical synaptic functions realistically using simple spike forms based solely on spike activity.

Sparse coding with memristor networks

Sparse representation of information provides a powerful means to perform feature extraction on high-dimensional data and is of broad interest for applications in signal processing, computer vision, object recognition and neurobiology. Sparse coding is also believed to be a key mechanism by which biological neural systems can efficiently process a large amount of complex sensory data while consuming very little power. Here, we report the experimental implementation of sparse coding algorithms in a bio-inspired approach using a 32 × 32 crossbar array of analog memristors. This network enables efficient implementation of pattern matching and lateral neuron inhibition and allows input data to be sparsely encoded using neuron activities and stored dictionary elements. Different dictionary sets can be trained and stored in the same system, depending on the nature of the input signals. Using the sparse coding algorithm, we also perform natural image processing based on a learned dictionary.

Device and SPICE modeling of RRAM devices

We report the development of physics based models for resistive random-access memory (RRAM) devices. The models are based on a generalized memristive system framework and can explain the dynamic resistive switching phenomena observed in a broad range of devices. Furthermore, by constructing a simple subcircuit, we can incorporate the device models into standard circuit simulators such as SPICE. The SPICE models can accurately capture the dynamic effects of the RRAM devices such as the apparent threshold effect, the voltage dependence of the switching time, and multi-level effects under complex circuit conditions. The device and SPICE models can also be readily expanded to include additional effects related to internal state changes, and will be valuable to help in the design and simulation of memory and logic circuits based on resistive switching devices.

Data Clustering using Memristor Networks

Memristors have emerged as a promising candidate for critical applications such as non-volatile memory as well as non-Von Neumann computing architectures based on neuromorphic and machine learning systems. In this study, we demonstrate that memristors can be used to perform principal component analysis (PCA), an important technique for machine learning and data feature learning. The conductance changes of memristors in response to voltage pulses are studied and modeled with an internal state variable to trace the analog behavior of the device. Unsupervised, online learning is achieved in a memristor crossbar using Sanger’s learning rule, a derivative of Hebb’s rule, to obtain the principal components. The details of weights evolution during training is investigated over learning epochs as a function of training parameters. The effects of device non-uniformity on the PCA network performance are further analyzed. We show that the memristor-based PCA network is capable of linearly separating distinct classes from sensory data with high clarification success of 97.6% even in the presence of large device variations.

Experimental Demonstration of Feature Extraction and Dimensionality Reduction Using Memristor Networks

Memristors have been considered as a leading candidate for a number of critical applications ranging from nonvolatile memory to non-Von Neumann computing systems. Feature extraction, which aims to transform input data from a high-dimensional space to a space with fewer dimensions, is an important technique widely used in machine learning and pattern recognition applications. Here, we experimentally demonstrate that memristor arrays can be used to perform principal component analysis, one of the most commonly used feature extraction techniques, through online, unsupervised learning. Using Sangers rule, that is, the generalized Hebbian algorithm, the principal components were obtained as the memristor conductances in the network after training. The network was then used to analyze sensory data from a standard breast cancer screening database with high classification success rate (97.1%).

Feature Extraction Using Memristor Networks

Crossbar arrays of memristive elements are investigated for the implementation of dictionary learning and sparse coding of natural images. A winner-take-all training algorithm, in conjunction with Ojas rule, is used to learn an overcomplete dictionary of feature primitives that resemble Gabor filters. The dictionary is then used in the locally competitive algorithm to form a sparse representation of input images. The impacts of device nonlinearity and parameter variations are evaluated and a compensating procedure is proposed to ensure the robustness of the sparsification. It is shown that, with proper compensation, the memristor crossbar architecture can effectively perform sparse coding with distortion comparable with ideal software implementations at high sparsity, even in the presence of large device-to-device variations in the excess of 100%.

Pattern recognition with memristor networks

In this paper we develop the concept of implementing pattern recognition algorithms in analog memristor networks. First, a device model is presented with experimental results demonstrating the feasibility of using WOx-based memristors to represent the tunable weights in a neural network. Next, simulation results demonstrate that an array of these memristors can be used to implement an unsupervised learning algorithm for pattern recognition. Handwritten digits are classified as an example problem while the concept is developed for more general use.

Modeling and implementation of oxide memristors for neuromorphic applications

We report the fabrication, modeling and implementation of nanoscale tungsten-oxide (WOx) memristive (memristor) devices for neuromorphic applications. The device behaviors can be predicted accurately by considering both ion drift and diffusion. Short-term memory and memory enhancement phenomena, and the effects of spike rate, timing and associativity have been demonstrated. SPICE modeling has been achieved that allows circuit-level implementations.

3-D Vertical Dual-Layer Oxide Memristive Devices

Dual-layer resistive switching memory devices with WOx switching layer formed at the sidewall of the horizontal electrodes have been fabricated and characterized. The devices exhibit well-characterized analog switching characteristics and small mismatch in electrical characteristics for devices formed at the two layers. The 3-D vertical device structure allows higher storage density and larger connectivity for neuromorphic computing applications. We show that the vertical devices exhibit potentiation and depression characteristics similar to planar devices, and can be programmed independently with no crosstalk between the layers.