TL;DR:
- AI Scientist combines theories and data to uncover scientific equations.
- In 1918, Irving Langmuir published a seminal paper that analyzed the behavior of gas molecules adhering to a solid surface.
- A hundred years later, researchers at IBM Research, Samsung AI, and UMBC developed an AI Scientist that successfully replicated a crucial part of Langmuir’s Nobel Prize-winning work and rediscovered Kepler’s third law of planetary motion and produced a good approximation of Einstein’s time-dilation law.
- The groundbreaking results were published in Nature Communications on April 12, 2023.
- AI-Descartes is a machine-learning tool that has the ability to reason and uses symbolic regression to find equations that fit the data.
- AI-Descartes offers several advantages over other systems, including its logical reasoning capability, which sets it apart from “generative AI” programs.
- The name AI-Descartes pays homage to 17th-century mathematician and philosopher René Descartes, who believed in the natural world being described by fundamental physical laws and logical deduction playing a key role in scientific discovery.
Main AI News:
In 1918, Irving Langmuir, an American chemist, published a seminal paper that analyzed the behavior of gas molecules when they adhere to a solid surface. This groundbreaking work was the result of meticulous experiments and a theory that posited solids offer discrete sites for gas molecules to occupy. With the help of this theory and the outcomes of his experiments, Langmuir derived a set of equations that precisely determine how much gas will adhere, depending on the pressure.
Fast forward a century, and researchers at IBM Research, Samsung AI, and UMBC have developed an “AI Scientist” that has successfully replicated a crucial component of Langmuir’s Nobel Prize-winning work. This cutting-edge artificial intelligence system not only confirmed Langmuir’s findings but also rediscovered Kepler’s third law of planetary motion, which calculates the time it takes for one celestial body to orbit another based on the distance between them. Additionally, the AI Scientist produced an accurate approximation of Einstein’s time-dilation law, which explains how time slows down for objects moving at high speeds.
The groundbreaking results of this study were recently published in the esteemed journal Nature Communications on April 12, 2023.
The newly developed “AI-Descartes” is a machine-learning tool that has the ability to reason and join the ranks of other innovative computing systems aimed at accelerating scientific discovery. This AI Scientist utilizes a concept called symbolic regression, which is the process of finding equations that fit the data. The system generates numerous candidate equations using basic mathematical operations such as addition, multiplication, and division and then selects the equation that most accurately describes the relationships in the data.
According to Cristina Cornelio, a research scientist at Samsung AI in Cambridge, England, and the first author of the paper, AI-Descartes offers several advantages over other systems, with its most unique feature being its logical reasoning capability. In instances where there are multiple candidate equations that fit the data, the system determines which equation aligns best with existing scientific theory. This ability to reason sets AI-Descartes apart from “generative AI” programs such as ChatGPT, which may produce errors in basic mathematics due to limited logical skills.
“Our work merges a first-principles approach, commonly used by scientists for centuries to derive new formulas based on existing scientific theory, with a data-driven approach that is prevalent in the machine learning era,” Cornelio explains. “By combining both approaches, we are able to create more accurate and meaningful models for a wide range of applications.“
The name AI-Descartes pays homage to the 17th-century mathematician and philosopher René Descartes, who believed that the natural world could be explained by a few fundamental physical laws and that logical deduction played a critical role in scientific discovery.
Conlcusion:
The development of AI-Descartes, a machine-learning tool with the ability to reason, has the potential to revolutionize the scientific industry. Its unique logical reasoning capability sets it apart from other AI programs and allows it to find accurate and meaningful models for a wide range of applications. Its success in replicating crucial components of Langmuir’s Nobel Prize-winning work, rediscovering Kepler’s law, and producing a good approximation of Einstein’s time-dilation law, highlights its potential for advancing scientific discovery. The combination of traditional scientific theories with data-driven approaches makes AI-Descartes a valuable tool for the scientific market.
