You'll have to remember this name: Dr. Fumiya Iida. This researcher from the University of Cambridge has created a robot capable of reproducing itself, making babies and improving its offspring by learning.
Ahilosophers readily distinguish between organisms and artifacts. From Bergson or Cangilhem to Miguel Benasayagthe organism is distinct from the machine because it adapts to evolve.
Researchers at the University of Cambridge and the ETH Zurich, led by Dr. Fumiya Iida, have taken this fundamental characteristic of living things as a starting point for building their robot. Their credo?" We want to see robots capable of innovation and creativity ".
By observing the process of evolution by natural selection dear to Darwin, researchers have thus built a robot-mother, capable of giving birth to robot-babies, educating them, observing them and improving them. All this without human intervention or specific computer programming.
The scientists recently published their work in the journal PLOS One under the heading " Morphological Evolution of Physical robots through Model-Free Phenotype Development ». There they explain the protocol of their experiment.
The parent robot, an UR5 articulated arm developed by Universal Robots, has been programmed to build small mobile robots. Small cubes with an electric motor to move around. These cubes had the task of moving from one point to another. Their "genome" was made up of five "genes" containing all the information about shape, construction, and motor commands. The mother robot gave birth to five generations of ten little robots. In each generation, the mother robot analyzed the behavior of her offspring to learn lessons that could be used to improve the next generation. This improvement was done by combining the five original "genes". On arrival, after 500 baby robots, the study showed that the latest generation baby robots were twice as efficient as the best baby robots from the first litter. It should be remembered that the whole process of observation and combination took place without human intervention or programming other than the initial command to build a robot capable of moving.
In this study, the evolution of robots takes place as in nature, through "mutations", where the components of a gene are modified, added, deleted or "crossed". The improvement in performance from generation to generation has been made possible because the mother has succeeded in inventing new forms and patterns of movement for her children over time.
" One of the big questions in biology is where does intelligence come from? We want to use robotics to explore that mystery.says Dr. Iida to the BBC. He goes on to explain that in nature, living organisms are able to adapt their physical characteristics to their environment over time. These adaptations allow biological organisms to survive in a wide variety of different environments. But machines are not adaptable in the same way. They are essentially stuck in a shape for their entire "life", and certainly changing their shape would make them more adaptable to changing environments.
Evolutionary robotics is a rapidly growing field of research; its goal is to create autonomous robots that learn on their own, without human intervention. This would allow machines to adapt to various environmental conditions, just as living organisms do. The applications are innumerable and this field of research is fascinating; nevertheless, by gradually blurring the boundaries between artefacts and organisms, it is singularly worrying.