A team of scientists has created what they are calling the first-ever, self-replicating living robots, which are extracted from frog cells and have been programmed using an evolutionary artificial intelligence algorithm. Called Xenobots, they were first artificially synthesized in laboratory conditions in January last year. Thanks to the efforts of a team that includes experts from The University of Vermont, the Allen Discovery Center at Tufts University and the Wyss Institute for Biologically Inspired Engineering at Harvard University. The team started by creating an evolutionary algorithm to create thousands of simulated cell combinations in various shapes and forms.
The algorithm made cellular combinations based on the biophysics of what a particular mixture of frog skin and cardiac cells can do. First, simulated organism designs that are selected by the algorithm as viable were saved as a cellular blueprint. Scientists then gathered stem cells from the embryos of frogs, separated them into cells and left them for incubation. This was followed by cutting and joining these cells under a microscope to form the same design as the algorithm suggested. So far, the team had observed that these tiny living robots could move and swim independently, thanks to coordinated cellular activity, just like a normal living organism or a collection of cells work in nature. Earlier this year, researchers in Singapore created a soft polymer foam that acts like self-healing skin for robotic limbs.
Now, the team behind the original Xenobot project has achieved a significant breakthrough by initiating self-reproduction in these robots that are being called Xenobots 3.0 to signify their upgraded status. The findings have been published as research titled "Kinematic self-replication in reconfigurable organisms" in the Proceedings of the National Academy of Sciences of the United States of America. The name, Xenobot, is a term based on Xenopus laevis, a species of African frog whose embryo was used to harvest stem cells for the project. Scientists claim that the mode of reproduction shown by these self-replicating and hand-assembled biological robots has never been observed before in nature. Using the evolutionary algorithm running on the Deep Green supercomputer cluster at UVM's Vermont Advanced Computing Core, the team tested billions of body shapes in simulation to find the one that can allow these biological robots to replicate on their own.
The goal was to achieve a process called motion-based "kinematic" replication. Interestingly, one of the cell cluster shapes recommended by the algorithm — and one that allowed the Xenobots to replicate on their own — looked like the classic video game character Pac-Man with a rounded profile and a distinct open mouth. As for the method of self-replication itself, these Pac-Man-shaped Xenobots found isolated single cells on their own, then collected hundreds of them, and eventually turned the cluster into a copy of themselves — all of it happening in the mouth region of the parent Xenobot.
A cellular cluster soon turned into a baby Xenobot, which then moved out on its own to find new cells and produce more self-replicating Xenobots. The premise sounds scary, but the team behind the research says that it is their obligation to carry forward the study and devise scenarios where these self-replicating programmable robots can be used for productive tasks like delivering drugs to body systems. Moreover, using such an evolutionary artificial intelligence algorithm comes with benefits of its own such as speeding up the vaccine development process for if a pandemic such as COVID-19 hits humanity again.
Sources: Harvard University, Proceedings of the National Academy of Sciences of the United States of America
from ScreenRant - Feed https://ift.tt/3DdwYU1
0 Comments