When a typical computer searches for a law of nature from data, it tries to guess a complete formula in one go. It's like a cook who, without knowing the recipe, dumps all the ingredients into a pot at once and hopes for soup. Most often, the result is an inedible mess: the formula is long but meaningless, and it fails on new data.
Scientists proposed a different path: let the machine follow in the footsteps of great physicists. First, it finds a simple pattern — for example, that a planet moves faster the closer it is to the Sun. Then it adds the next link, then another. In this way, several clear pieces come together into a chain that ultimately yields an accurate law. They named this method "symbolic regression chains."
Tests on three problems (turbulent boiling of a liquid, viscous flow through a pipe, and metal evaporation under a laser) showed that the new method outperforms older ones. And on data about air resistance for various aircraft, it found a pattern that had been overlooked before. This means the machine can not only rediscover known laws but also suggest new useful formulas to engineers.
This approach is like assembling a construction set: each piece is simple and understandable, and together they form a complex mechanism. And that is far more reliable than trying to guess the entire mechanism at once.
Symbolic regression is when a computer searches not just for a number, but for a formula that can be written on paper and understood in terms of how it works.