Cellular Wave Computer Research Articles

Cellular wave computing via nanoscale chip architectures

Nanoscale electronic technologies, both CMOS-based and beyond, provide the means to fabricate massively parallel systems with thousands or millions of processors/cores in a small package. Most of these architectures are in prototype stage. Owing to physical constraints the cellular property prevails: they have blocks of mainly identical processing cells in arrays, and the precedence of geometric locality is mandatory. The cells might vary both in function as well as in their modes of operation, including arithmetic/analog, logic and symbolic cells, possibly with integrated sensory elements. The interplay between local dynamics and global dynamics plays an increasingly important role. Cellular wave computers rely on this architecture endowed with a new world of algorithms based on spatial-temporal dynamics as protagonist instructions. Today, crossing the billion-transistor-chip boundary, desktop supercomputers are emerging. A typical product might consist of close to a million logic processors and tens of thousands of DSPs in FPGA arrays. Other architectures with GPUs and array processors are becoming ubiquitous. On the other hand, cellular visual microprocessors and other portable low-power devices move towards wireless and other portable technologies. The maturity of 3D integration provides another mechanism for complexity and efficiency increase. The biological counterparts are starting to cover not only the single modality effects, but multimodal organizations are considered as well. Moreover, they inspire learning algorithms. Not only are new devices emerging, but new algorithmic thinking and related software technologies are also developing hand-in-hand. The issue is not just to ‘parallelize’ existing algorithms, but to invent new spatial-temporal algorithms for tasks that are operating on either topographic or non-topographic data or objects. The recent success of the 12th IEEE CNNA 2010 Workshop in Berkeley (February 2010) has made a boost in the development of this field and many new devices entering the market by protagonist vendors accelerate the need for a new kind of computer science and for new computer engineering tools. Physical implementation, CMOS, and non-CMOS devices Virtual and physical cellular machines and algorithmic principles Algorithmic innovation and software design tools Biological modeling and inspiration Application case studies Theoretical foundation. Deadline for submission of manuscripts is 20 December 2010, via the electronic submission system of the Journal.

Cellular wave computer algorithms with spatial semantic embedding for handwritten text recognition

AbstractThe recognition of cursive handwritten texts is a complex, in some cases unsolvable, task. One problem is that in most cases it is difficult or impossible to identify each letter, even if the words are separated. In our new method, the identification of letters is not needed due to the extensive and iterative use of semantic and morphological information of a given language. We are using a spatial feature code, generated by a cellular nonlinear network (CNN) based cellular wave computer algorithm, and combine it with the linguistic properties of the given language. Most general‐purpose handwriting recognition systems lack the ability to integrate linguistic background knowledge because they use it only for post‐processing recognition results. The high‐level a priori background knowledge is, however, crucial in human reading and similarly it can boost recognition rates dramatically in case of recognition systems. In our new system we do not treat the visual source as the only input: geometric and linguistic information are given equal importance. On the geometric side we use word‐level holistic feature detection without letter segmentation by analogic CNN algorithms designed for cellular wave computers (IEEE Trans. Circuits Syst. 1993; 40:163–173; Cellular Neural Networks and Visual Computing, Foundations and Applications. Cambridge University Press: Cambridge, U.K., New York, 2002). The linguistic side is based on a morpho‐syntactic linguistic system (Proceedings of COLING‐2002, vol. II, Taipei, Taiwan, 2002; 1263–1267). A novel shape coding method is used to interface them, and their interaction is enhanced via an inverse filtering technique based on features that are global or of a low confidence value. A statistical context selection method is also applied to further reduce the output word lists. Copyright © 2008 John Wiley & Sons, Ltd.

Cellular wave computer architectures in a new era of computing—15 years later

Cellular wave computer architectures in a new era of computing—15 years later

Statistical physics on cellular neural network computers

The computational paradigm represented by Cellular Neural/nonlinear Networks (CNN) and the CNN Universal Machine (CNN-UM) as a Cellular Wave Computer, gives new perspectives also for computational statistical physics. Thousands of locally interconnected cells working in parallel, analog signals giving the possibility of generating truly random numbers, continuity in time and the optical sensors included on the chip are just a few important advantages of such computers. Although CNN computers are mainly used and designed for image processing, here we argue that they are also suitable for solving complex problems in computational statistical physics. This study presents two examples of stochastic simulations on CNN: the site-percolation problem and the two-dimensional Ising model. Promising results are obtained using an ACE16K chip with 128×128 cells. In the second part of the work we discuss the possibility of using the CNN architecture in studying problems related to spin-glasses. A CNN with locally variant parameters is used for developing an optimization algorithm on spin-glass type models. Speed of the algorithms and further trends in developing the CNN chips are discussed.

Circuits, computers, and beyond Boolean logic

AbstractHistorically, the invention of the stored programmable computer architecture, introduced by John Von Neumann, was also influenced by electrical circuit implementation aspects, as well as tied to fundamental insight of logic reasoning. It can also be considered as a mind‐inspired machine. Since then, the implementation of logic gates, control and memories has developed independently of the architecture. The Cellular Wave Computer architecture (IEEE Trans. Circuits Syst. II 1993; 40:163–173; Electron. Lett. 2007; 43:427–449; J. Circuits Syst. Comput. 2003; 5(2):539–562) as a spatial–temporal universal machine on flows has also been influenced by circuit aspects of very large‐scale integration (VLSI) technology, as well as some motivating living neural circuits, via the cellular nonlinear (neural) network (CNN). It might be considered as a brain‐inspired machine.In this paper, after summarizing the main properties of the Cellular Wave Computer, we highlight a few basic properties of this new kind of computer and computing. In particular, phenomena related to (i) the one‐pass solution of a set of implicit equations due to real‐time spatial array feedback, (ii) the true random signal array generation via the insertion of the continuous physical noise signals, (iii) the finite synchrony radius due to the functional delay of wires, as well as to (iv) biology relevance. We also show that the Cellular Wave Computer is performing spatial–temporal inference that goes beyond Boolean logic, a characteristic of living neural circuits. Copyright © 2007 John Wiley & Sons, Ltd.

Tactile sensing-processing: Interface-cover geometry and the inverse-elastic problem

The skin-like elastic cover of a tactile-sensor array plays a fundamental role in determining how the sensor is responding to an arbitrary surface stimulus. Indeed, this cover is the first spatial–temporal sensory instruction in a tactile cellular wave-computer, or in living, neural-tactile signal-processing organs. While the sensor under the elastic layer measures the local strain/stress, we are interested in the stimulating force distribution at the contacting surface. This paper deals with the described inverse problem in the stationary case over a three-axial tactile-sensor array, measuring three components of the local strain tensor of the elastic cover. Our goal here is to design an interface flexible cover to enhance specific capabilities, sensitivity and accuracy, when using the three-axial sensor array. First, on the simplest, flat rubber surface, we create a kind of tactile hyper-accuracy by giving a simple analytical solution to the case when the cover is indented by an arbitrary point load. Second, instead of generalizing our solution to more complex indentation profiles, we try to abolish the need for the complex inverse solution by specifying our cover's geometry and thus reducing the complexity of the sensor's coding mechanism. We create hemispherical bumps over the elastic surface and use a finite-element model to show that with the bumps we can localize the (otherwise continuous) input over the taxels and detect normal and shear force components independently. Finally, we confirm our theoretical results with real-time experimental data and use the measurements for sensor calibration and texture classification.

Cellular wave computers for nano-tera-scale technology—beyond Boolean, spatial-temporal logic in million processor devices

Cellular wave computer architecture as a spatial-temporal universal machine on flows reflects (i) the prophecies of Turing and von Neumann, (ii) physical constraints of nano-scale technology, (iii) typical constructs in neural organisation and mechanisms, and (iv) provides for a new way of algorithmic and hardware-software framework, already implemented in some new products, including visual microprocessors. A few aspects of this new kind of computer and computing are highlighted, including the `function in layout' principle, the biological relevance in retinal modelling, and the new type of spatial-temporal algorithmic thinking beyond Boolean logic

Stochastic simulations on the cellular wave computers

The computational paradigm represented by Cellular Neural/nonlinear Networks (CNN) and the CNN Universal Machine (CNN-UM) as a Cellular Wave Computer, gives new perspectives for computational physics. Many numerical problems and simulations can be elegantly addressed on this fully parallelized and analogic architecture. Here we study the possibility of performing stochastic simulations on this chip. First a realistic random number generator is implemented on the CNN-UM, and then as an example the two-dimensional Ising model is studied by Monte Carlo type simulations. The results obtained on an experimental version of the CNN-UM with 128 * 128 cells are in good agreement with the results obtained on digital computers. Computational time measurements suggests that the developing trend of the CNN-UM chips - increasing the lattice size and the number of local logic memories - will assure an important advantage for the CNN-UM in the near future.

Cellular wave computers for brain-like spatial-temporal sensory computing

Present day classical computers, developed during the last sixty years are logic machines, based on binary logic and arithmetic, acting on discrete valued (binary coded) data. Its unique property is algorithmic (stored) programmability, invented by John von Neumann. The mathematical concept is based on a universal machine on integers (Turing machine). Cellular automata, introduced also by J. von Neumann, are fully parallel array processors with all discrete space, time and state values. Their beautiful properties have been recently rediscovered showing the deeper qualitative properties, if we allow the states and time to be continuous values like in CNN, a broader class of dynamics will be generated. Even more, the fundamental conditions to generate complex features at the edge of chaos have been established: the need of local activity. Taking one step further, and using the CNN-UM architecture, a new world of algorithms is opening.