Physicists Use AI to Enable Synthetic Microswimmers to Learn
Bio-inspired microscopic robots learn to navigate with AI machine learning.
Posted March 31, 2021
- Scientists have used artificial intelligence machine learning to "teach" navigation to synthetic microswimmers much smaller than a strand of human hair.
- To conduct the experiment gold nanoparticles were attached to the surface of thermophoretic microswimmers to absorb light from a laser.
- This technology may be a step toward future developments in precision medicine.
Artificial intelligence (AI) machine learning is enabling the convergence of multiple scientific disciplines. Synthetic microswimmers, inspired by biological microscopic organisms, are human-made microscopic particles. Recently, scientists in the Molecular Nanophotonics Group at the Universität Leipzig and their collaborators applied AI machine learning to enable synthetic microswimmers to learn to navigate; they published their findings in the March 24, 2021 issue of Science Robotics .
“Although the implementation of signaling and feedback by physical or chemical processes into a single artificial microswimmer is still a distant goal, the current hybrid solution opens a whole branch of new possibilities for understanding adaptive behavior of single microswimmers in noisy environments and the emergence of collective behavior of large ensembles of active systems,” the researchers reported.
The study was led by professor Frank Cichos at the Universität Leipzig, in collaboration with Santiago Muiños Landin, Alexander Fischer, and Viktor Holubec. Secondary affiliations include the AIMEN Technology Centre, Smart Systems and Smart Manufacturing–Artificial Intelligence and Data Analytics Laboratory in Spain (Landin), and the Charles University in the Czech Republic (Holubec).
Together, they applied artificial intelligence machine learning to real-world synthetic microswimmers smaller than one-thirtieth of the diameter of a single strand of human hair.
“Artificial microswimmers that can replicate the complex behavior of active matter are often designed to mimic the self-propulsion of microscopic living organisms,” wrote the researchers. “However, compared with their living counterparts, artificial microswimmers have a limited ability to adapt to environmental signals or to retain a physical memory to yield optimized emergent behavior.”
In physics, Brownian motion refers to the erratic random movement of microscopic particles in a fluid. The concept was named after the Scottish paleobotanist and botanist Robert Brown (1773–1858) who in 1827 used a microscope to observe the irregular and unceasing motion of minute particles from the pollen grains of the herbaceous perennial plant Clarkia pulchella .
“Different from macroscopic living systems and robots, both microscopic living organisms and artificial microswimmers are subject to Brownian motion, which randomizes their position and propulsion direction,” wrote the researchers. “Here, we combine real-world artificial active particles with machine learning algorithms to explore their adaptive behavior in a noisy environment with reinforcement learning.”
To conduct the study, the researchers used AI reinforcement learning (RL), where an agent learns from interacting with its environment through rewards or penalties. According to the scientists, recent computation studies show that AI reinforcement learning can provide “optimal strategies for the navigation of active particles through flows,” as well as the swarming of robots and development of collective motion.
The scientists used a light-controlled self-thermophoretic microswimmer with 30 percent of its surface covered with gold nanoparticles to conduct the experiment. For the microswimmer suspended in water to self-propel, light directed from a laser is absorbed by the gold nanoparticles which generates a temperature gradient and induces thermo-osmotic surface flows. The microswimmers “learn” with the aid of AI reinforcement learning how to navigate through liquid despite Brownian motion.
“Combining elements of machine learning and real-world artificial microswimmers would considerably extend the current computational studies into real-world applications for the future development of smart artificial microswimmers,” wrote the researchers.
At the intersection of physics, chemistry, math, computational analysis, and biology is the interdisciplinary branch of knowledge called biophysics—the science of explaining biological functions in terms of the physical properties of molecules. The understanding gained from biophysics can enable the development of novel drug treatments such as nanomedicines and biomolecules for precision medicine. The study is proof-of-concept that applying AI machine learning algorithms to microscopic agents can help reveal the physical phenomena that are vital to microscopic motion of biological organisms. It is a step towards intelligent autonomous microscopic robots that may be used for the development of novel smart drugs in the not-so-distant future.
Copyright © 2021 Cami Rosso All rights reserved.