Christof Teuscher

CryptoBooster: A Reconfigurable and Modular Cryptographic Coprocessor

The CryptoBooster is a modular and reconfigurable cryptographic coprocessor that takes full advantage of current high-performance reconfigurable circuits (FPGAs) and their partial reconfigurability. The CryptoBooster works as a coprocessor with a host system in order to accelerate cryptographic operations. A series of cryptographic modules for different encryption algorithms are planned. The first module we implemented is IDEACore, an encryption core for the International Data Encryption Algorithm (IDEA?).

Turing’s Connectionism

In a “little known” paper entitled “Intelligent Machinery,” Turing had already investigated connectionist models as early as 1948. Unfortunately, his work was dismissed by his employer and went unpublished until 1968, 14 years after his death.

Critical Values in Asynchronous Random Boolean Networks

Wherever we see life, we see different kinds of complex networks, reason why they are studied across various fields of science. Random Boolean Networks (RBNs) form a special class in which the links between the nodes and the boolean functions are specified at random. Whereas synchronous RBNs were investigated in detail, there has little been done around their asynchronous counterpart, although there is evidence that most living systems are governed by asynchronous updating. Derrida’s annealed approximation predicts a critical connectivity value of K=2 for synchronous RBNs. We present a similar and original approach for asynchronous RBNs and show that they do not possess such a critical connectivity value. The asynchronous and nondeterministic updating scheme introduces perturbations that reach about 25% of the nodes and thus prevents the networks to become stable. Further, our numerical simulations show that asynchronous RBN tend to amplify small and to reduce big perturbations.

BioWatch: a giant electronic bio-inspired watch

The Embryonics project is inspired by some of the basic processes of molecular biology, such as the embryonic development of living beings. Transposing these processes into digital electronics integrated circuits, we design artificial organism endowed with properties typical of to the living work, such as self-repair and self-healing. In order to illustrate the original features of the Embyronics project, we define the cellular and molecular architecture of a giant artificial organism, the BioWatch. The hardware implementation of our watch exploits a new reconfigurable tissue, the bio-inspired electronic wall or BioWall.

A networked FPGA-based hardware implementation of a neural network application

Describes a networked FPGA-based implementation of the FAST (Flexible Adaptable-Size Topology) architecture, an artificial neural network (ANN) that dynamically adapts its size. Most ANN models base their ability to adapt to problems on changing the strength of the interconnections between computational elements according to a given learning algorithm. However, constrained interconnection structures may limit such ability. Field programmable hardware devices are very well adapted for the implementation of ANNs with in-circuit structure adaptation. To realize this implementation, we used a network of Labomat-3 boards (a reconfigurable platform developed in our laboratory), which communicate with each other using TCP/IP or a faster direct hardware connection.

A reconfigurable hardware membrane system

P systems are massively parallel systems and software simulations do no usually allow to exploit this parallelism. We present a parallel hardware implementation of a special class of membrane systems. The implementation is based on a universal membrane hardware component that allows to efficiently run membrane systems on specialized hardware such as FPGAs. The implementation is presented in detail as well as performance results and an example.

Hypercomputation: hype or computation?

Can we physically implement a hypercomputer? So far no one has.

A Self-Repairing and Self-Healing Electronic Watch: The BioWatch

The Embryonics project is inspired by some of the basic processes of molecular biology, such as the embryonic development of living beings. Transposing these processes in to digital electronic integrated circuits, we design artificial organisms endowed with properties typical of to the living world, such as self-repair and self-healing. In order to illustrate the original features of the Embryonics project, we define the cellular and molecular architecture of a giant artificial organism, the BioWatch. The hardware implementation of a microprogrammed version of our watch exploits a new reconfigurable tissue, the bio-inspired electronic wall or BioWall.

On fireflies, cellular systems, and evolware

Many observers have marveled at the beauty of the synchronous flashing of fireflies that has an almost hypnotic effect. In this paper we consider the issue of evolving two-dimensional cellular automata as well as random boolean networks to solve the firefly synchronization task. The task was successfully solved by means of cellular programming based co-evolution performing computations in a completely local manner, each cell having access only to its immediate neighbors states. An FPGA-based Evolware implementation on the BioWalls cellular tissue and different other simulations show that the approach is very efficient and easily implementable in hardware.

A Revival of Turing's Forgotten Connectionist Ideas: Exploring Unorganized Machines

Turing had already investigated connectionist networks at the end of the forties and was probably the first person to consider building machines out of very simple, neuron-like elements connected together in a mostly random manner. The present paper aims to revive and shed more light on Turing’s ideas. Turing’s unorganized machines are analyzed and new types of machines are proposed by the authors. An example of a pattern classification task is presented using a Turing net and genetic algorithms for building and training the networks.