Giovanni Egidio Pazienza

Advances in Neuromorphic Memristor Science and Applications

Physical implementation of the memristor at industrial scale sparked the interest from various disciplines, ranging from physics, nanotechnology, electrical engineering, neuroscience, to intelligent robotics. As any promising new technology, it has raised hopes and questions; it is an extremely challenging task to live up to the high expectations and to devise revolutionary and feasible future applications for memristive devices. The possibility of gathering prominent scientists in the heart of the Silicon Valley given by the 2011 International Joint Conference on Neural Networks held in San Jose, CA, has offered us the unique opportunity of organizing a series of special events on the present status and future perspectives in neuromorphic memristor science. This book presents a selection of the remarkable contributions given by the leaders of the field and it may serve as inspiration and future reference to all researchers that want to explore the extraordinary possibilities given by this revolutionary concept.

Teaching Memristors to EE Undergraduate Students [Class Notes]

Generations of electrical engineers have learned that there are three fundamental passive two-terminal circuit elements: resistor, capacitor, and inductor. Nevertheless, this apparently immovable situation changed in 2008 when Nature published an article on the memristor, which was proved to be the fourth fundamental circuit element. Since then, researchers have devoted time and effort to find how this device may possibly change the future of electronics. It is time then to introduce the memristor in EE undergraduate courses, but how? The great majority of works on memristor published so far have been aimed at experienced researchers and not at young students. The goal of this paper is providing an original point of view on this issue and describing a simple approach to memristor which is suitable to be used in EE undergraduate courses.

2012 Special Issue: Metamodeling and the Critic-based approach to multi-level optimization

Large-scale networks with hundreds of thousands of variables and constraints are becoming more and more common in logistics, communications, and distribution domains. Traditionally, the utility functions defined on such networks are optimized using some variation of Linear Programming, such as Mixed Integer Programming (MIP). Despite enormous progress both in hardware (multiprocessor systems and specialized processors) and software (Gurobi) we are reaching the limits of what these tools can handle in real time. Modern logistic problems, for example, call for expanding the problem both vertically (from one day up to several days) and horizontally (combining separate solution stages into an integrated model). The complexity of such integrated models calls for alternative methods of solution, such as Approximate Dynamic Programming (ADP), which provide a further increase in the performance necessary for the daily operation. In this paper, we present the theoretical basis and related experiments for solving the multistage decision problems based on the results obtained for shorter periods, as building blocks for the models and the solution, via Critic-Model-Action cycles, where various types of neural networks are combined with traditional MIP models in a unified optimization system. In this system architecture, fast and simple feed-forward networks are trained to reasonably initialize more complicated recurrent networks, which serve as approximators of the value function (Critic). The combination of interrelated neural networks and optimization modules allows for multiple queries for the same system, providing flexibility and optimizing performance for large-scale real-life problems. A MATLAB implementation of our solution procedure for a realistic set of data and constraints shows promising results, compared to the iterative MIP approach.

A brief analysis of the main SPICE models of the memristor

In the last few years, the memristor has sparked the interest of numerous researchers. The lack of a commercially available off-the-shelf device has encouraged the publication of numerous SPICE models, which are not always easy to use. We believe that unclear schematics and obscure, and sometimes flawed, codes have prevented the widespread application of yet valid models. In this paper, we devote our efforts to the clear and systematic presentation of a few remarkable SPICE models, and provide a complete library containing all working codes and essential schematics.

Memristor as an archetype of dynamic data-driven systems and applications to sensor networks

Since its introduction a decade ago, DDDAS has been applied to a wide range of science and technology fields, with specific focus of areas requiring fast and reliable processing of massive data streams from diverse resources. It is crucial to explore architectures and systems which are naturally suited to the DDDAS framework. In this paper, we show that the memristor – the fourth fundamental two-terminal passive circuit element alongside the well-known resistor, capacitor, and inductor – affords the efficient implementation of the working principles of DDDAS. Hence memristors can be considered as an archetype of DDDAS nanoscale hardware embodiment, being the smallest and most basic dynamic data driven application system. Memristors are electrical components with inherent memory processes; they have been predicted about four decade ago and have been physically implemented recently. We discuss the role that DDDAS may play in the development of computing platforms and sensor networks based on memristors in the next few years.

Are Memristors the Future of AI

We review the state-of-the-art of neuromorphic memristor science and technology. We cover principles of memristors and neuromorphic systems, computational models of memristors, and hardware implementations. Potential applications of memristors are also described, including supercomputing, image processing, computer vision, intelligent control, and robotics. This review is based on the chapters of the present volume, which extend the materials of the invited and plenary talks given at the series of events on memristors in 2011. We elaborate on challenges and future perspectives of this promising new research field.

Metamodeling for large-scale optimization tasks based on object networks

Optimization in large-scale networks - such as large logistical networks and electric power grids involving many thousands of variables - is a very challenging task. In this paper, we present the theoretical basis and the related experiments involving the development and use of visualization tools and improvements in existing best practices in managing optimization software, as preparation for the use of “metamodeling” - the insertion of complex neural networks or other universal nonlinear function approximators into key parts of these complicated and expensive computations; this novel approach has been developed by the new Center for Large-Scale Integrated Optimization and Networks (CLION) at University of Memphis, TN.

Applications of the virtual cellular machine to many-core processors

There is a large variety of multi- and many-core platforms but established design methodologies for such processor arrays are still missing. We propose the concept of virtual cellular machine as an architecture, which does not correspond necessarily to a physical device, built according to the specific requirements needed by the programmer. The features of a virtual cellular machine (processors, memories etc.) can then be mapped onto existing physical many-core processors.

How to teach memristors in EE undergraduate courses

Since a couple of years, the memristor has been in the media spotlight and it is expected to have a major impact on the future technology. Many scientists think that it should be taught in EE undergraduate courses, but there is no general agreement on how to do so. This paper presents several approaches to memristor and a thorough discussion about them, and it is the result of numerous discussions with experts in this area.

Percolation in memristive networks

Numerous scientists claim that the memristor may be a real breakthrough in the fields of electronic and circuit design. For this reason, it is important to study what dynamics arise in memristive networks and speculate about how they could be used for meaningful tasks. In this paper, we focus on the phenomenon of percolation in memristive networks, studying the theoretical aspects and performing SW simulations.