A groundbreaking achievement by a group of scientists has given birth to an artificial intelligence (AI) system that empowers robots to independently conduct up to 10,000 scientific experiments in a single day.
Named BacterAI, this AI system has the potential to revolutionize the pace of discovery across various fields including medicine, agriculture, and environmental science. In a recent publication in Nature Microbiology, the research team showcased the successful implementation of BacterAI in mapping the metabolic processes of two oral health-related microbes.
Unveiling the Amino Acid Requirements for Microbial Growth
Led by Professor Paul Jensen at the University of Michigan, the research team aimed to uncover the specific amino acid requirements for the growth of beneficial mouth microbes.
However, due to the immense number of potential combinations arising from the 20 available amino acids, determining the precise combination for each bacterial species presented a formidable challenge.
To overcome this obstacle, BacterAI was developed to test hundreds of amino acid combinations each day and adapt its approach based on the previous day’s outcomes. In just nine days, BacterAI successfully predicted the necessary amino acid combinations for growth with 90% accuracy.
Accelerating Discoveries through Automated Experimentation
The implications of automated experimentation and the utilization of BacterAI extend far beyond the realm of microbiology, as researchers from diverse fields can now formulate questions as puzzles for AI to solve through trial and error.
It is important to note that approximately 90% of bacteria remain largely unexplored, and conventional methods of acquiring fundamental scientific knowledge about them demand significant time and resources. The rapid pace of automated experimentation facilitated by BacterAI has the potential to accelerate these discoveries, paving the way for future breakthroughs in various domains.
BacterAI stands out as an exceptional innovation in the field of AI due to its ability to generate its own dataset through a series of experiments, diverging from the conventional approach of feeding labeled datasets into machine learning models.
The platform leverages the outcomes of prior trials to make predictions about experiments that yield the most significant amount of information. With fewer than 4,000 experiments, BacterAI has already deciphered most of the rules governing the precise feeding of bacteria.
Under the guidance of Professor Paul Jensen, the development of BacterAI represents a remarkable leap forward in harnessing AI to expedite scientific discovery. As demonstrated by the team’s research, BacterAI provides crucial insights into the microbial world, including those impacting our health.
With robots now capable of conducting up to 10,000 experiments per day, the potential for BacterAI to accelerate discoveries is immense. The study received funding from the National Institutes of Health, with support from NVIDIA, further highlighting the significance of this groundbreaking advancement.
