NEW AI SYSTEM IDENTIFIES PHYSICAL RULES DIRECTLY FROM DATA

Researchers have developed a new artificial intelligence system designed to identify underlying physical laws by analyzing raw data, offering a different approach to how scientific models are built.

- 

Instead of relying on predefined equations or assumptions supplied by human researchers, the system examines observations and infers mathematical relationships that describe how a system behaves.

The work builds on a longstanding challenge in physics and related sciences: translating complex measurements into concise, interpretable rules. Traditionally, this process depends on human intuition, theoretical frameworks, and iterative testing. The new Al-driven method shifts part of that burden to computation, allowing patterns to emerge directly from experimental or simulated data.

According to the researchers behind the project, the system is not intended to replace theory-driven science but to assist in situations where the governing equations are unknown, incomplete, or difficult to derive using conventional methods.

HOW THE SYSTEM INFERS

PHYSICAL LAWS

The Al operates by searching for mathematical structures that best explain observed behavior while favoring simplicity and consistency. Rather than producing opaque predictions, it outputs symbolic equations that can be inspected, tested, and compared against known physical principles.