Page 25

CIM29

N°29 I quarterly I APR IL 2013 Focus | 25 w From Inspiration to Application w Counterfeit-proof banknotes, cosmetics with no toxic pigments, television screens… The morpho (A), a butterfly native to the Amazon Basin, has already inspired researchers at the INSP1 and their industrial partners. What makes this insect so fascinating is the way light interacts with the nanometric structures (B) on the surface of its wings. “With our colleagues at the University of Namur (Belgium), we are also studying how certain animals, like fireflies, emit light,” reports Serge Berthier of the INSP. Here again, the first applications are already being developed, including an electroluminescent diode that incorporates a bioinspired device. 01. Institut des nanosciences de Paris (CNRS / UPMC ). Contact i nformation: Serge Berthier > serge.berthier@insp.jussieu.fr Nature is undeniably a treasure-trove of ingenuity. Imitating it opens a plethora of potential applications. Here is a selection. “In the field of probability, some algorithms are called ‘genetic’ because they are inspired by natural selection,” explains Pierre Del Moral of the IMB.1 Imagine having to select the best route on a map. The various possible solutions are represented by different individuals. “Each of them has a favorite itinerary and they all compare their proposals to evaluate which is the best,” the researcher explains. The individuals considered most efficient are duplicated and can create variants by mutation, while the others, deemed “unsuitable,” disappear… “This approach is based on solution diversity,” explains Del Moral, “unlike conventional reinforcement approaches, at work in artificial neural networks.” In neural network systems, which are inspired by biological neurons, the most effective solutions are reinforced, but no new ones are created. Genetic algorithms, used since the 1950s, are more efficient, although researchers have never quite understood why. The mathematical model developed by Del Moral now provides the answer.2 01. Institut de mathématiques de Bordeaux (CNRS / Universités Bordeaux-I and Bordeaux- Segalen / IP B / Inria). 02. P. Del Moral et al., “On the concentration properties of interacting particle processes,” Foundations and Trends in Machine Learning, 2012. 3: 225-389. Contact i nformation: Pierre Del Moral, pierre.del-moral@inria.fr Darwini an alg orithms Diatoms are minuscule algae with an unusual but very interesting characteristic: they are protected by a clear exoskeleton. Similar in composition to glass, this outer shell—or frustule—has a remarkable porous structure. Seen through a high-power microscope, it appears as a network of small cavities. Magnifying the image further reveals another network of smaller holes. This hierarchical structure, with pores varying in size from one micron down to a few nanometers, was taken as a model for the development of sensors and  catalysts  by researchers from the LCMCP,1 the CRPP, 2 and IRCELYON.3 The main advantage of this type of bioinspired structure is that it improves the devices’ reactivity. The total surface of the pores in which catalysis or chemical identification takes place is enormous: a single gram of these materials can cover up to 1000 m2. 01. Laboratoire de chimie de la matière condensée de Paris (CNRS / UP MC / ENSCP / Collège de France). 02. Centre de recherche Paul Pascal (CNRS ). 03. Institut de recherche sur la catalyse et l’environnement de Lyon (CNRS / UC BL). Contact i nformation: Clément Sanchez, clement.sanchez@upmc.fr Holes in holes... catalyst. Substance used to accelerate a chemical reaction. © national geographic creative/getty images © eye of science/phanie © s. gschmeissner/spl/cosmos A B q Seen under a microscope and colorized, this diatom reveals its complex structure: a network of cavities that are in turn perforated by other smaller holes.


CIM29
To see the actual publication please follow the link above