Chemical Computing Peter Dittrich Bio Systems Analysis Department

Chemical Computing Peter Dittrich Bio Systems Analysis Department. of Mathematics and Computer Science Friedrich-Schiller-University Jena BMBF Grant No. 0312704 A Friedrich-Schiller-Universität Jena UPP 2004, Mt. S. Michel, 15. 09. 04 Jena Centre for Bioinformatics P. Dittrich - FSU & JCB Jena

Jena Downtown. . . UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 2

Here we are. . . UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 3

CS, Jena University Bio Systems Analysis Group Jena Centre for Bioinformatics UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 4

Chemical Computing 1. Computing helps Chemistry 2. Chemistry helps Computing UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 5

What is Chemistry? • Deals with – Substances composed of molecules – Reactions that transform substances, such that the composition of molecules changes – Dynamics UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 6

Where does Chemical Computing Occur in Nature? UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 7

Principles of Chemical Computing • • • Pattern recognition Formation of (spatial) structures Change of conformation Optical activity Chemical kinetics Energy minimization UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 8

Chemistry Helps Computing 1. Real chemical computing 2. Artificial chemical computing UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 9

Examples Where the Chemical Metaphor is Used in Computing • Real Chemical Computing (Liberman 1972, 1979), (Conrad 1972) (Seelig & Rössler 1972) and others – – – Enzymes DNA/RNA-Comp Optical Reaction-Diffusion Programmed Self-Assembly UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 10

Examples Where the Chemical Metaphor is Used in Computing Artificial Chemical Computing • Abstract Molecular Machine (Liang) • Rewriting systems (e. g. , GAMMA, CHAM, P-Systgems, ARMS, …) • Hormone systems in distributed robot control systems (e. g. COG) • Chemical-like systems to control the behavior and emotions in artificial agents (e. g. Creatures or PSI (D. Dörner) • Control of morpho-genetic systems (control of artificial gene expression and morphogenesis) • Control of growth of artificial neural networks (e. g. , Astor/Adami) • Control of amorphous computers • Communication among neurons in an ANN where neurons have spatial coordinates (e. g. , neural gas by P. Husbands) UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 11

Example for Microscopic Chemical Computing DNA Computing (Adleman) UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 12

Example for Macroscopic Chemical Computing Chemical Neuron [see Hjelmfelt, Weinberger, Ross 1991] UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 13

Example for Macroscopic Chemical Computing: Simple Hyper-cyclic Associative Memory Answer Query (Output) (Input) Hypercycle of replicating catalysts [Dittrich 1995] UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 14

Some interesting aspects … UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena

Fine Grained Parallelism Usually: • Distributed • Robust • Asynchronous • Emergent • Self-organizing → soft computing, organic computing, computational intelligence UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 16

“Invisible Networks” UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 17

“Invisible Networks” UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 18

“Invisible Networks” • A network larger than the neural network of the human brain: • M = {2, 3, …, 10 E 30} • A + B + X -> A + B + C with C = A/B if A mod B = 0, C=C otherwise. UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 19

Self-modification • Self-modification (s. higher-order & generative programming) • Strange loop • Dualism of – structure and function – data and program – Tape and machine UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 20

Challenges UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena

Challenges • • Efficiency Scalability Programmability Adaptability UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 22

The talks in the chemical computing session … UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena

Wolfgang Banzhaf University of Newfoundland Evolving Artificial Chemistries by Genetic Programming UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 24

Andrew Adamatzky University of the West of England Programming Reaction. Diffusion Computers UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 25

Tetsuya Asai Graduate School of Informaton Science and Technology, Sapporo Reaction Diffusion Processors UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 26

Klaus-Peter Zauner University of Southampton From Prescriptive Programming of Solid. State Devices to Orchestrated Self. Organization of Informed Matter UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 27

Winfried Kurth University of Cottbus Relational Growth Grammars UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 28

Yann Radenac IRISA, Rennes High-order Chemical Programming Style UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 29

Questions for discussion • How to program a chemical computer (whatever it is)? • How do chemical computing paradigms scale up? • Can the chemical metaphor lead to new computational systems with abilities superior to conventional approaches? • Or even to systems that can not be realized by conventional approaches? UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 30

Thank You UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 31

COG (MIT, Brooks) UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 32

Growing Artificial NNs [Astor/Adami] [J. S. Astor, Christophs Adami: A Developmental Model for the Evolution of Artificial Neural Networks. , Artificial Life 6(3), 189 -218, 2000 http: //norgev. alife. org/] UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 33

PSI (D. Dörner) UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 34

PSI (D. Dörner) UPP 2004, Mt. S. Michel, 15. 09. 04 P. Dittrich - FSU & JCB Jena 35