Skip to main content

The technique behind Ufinq

Symbolic Regression.

Finding the formula in the data.

Symbolic Regression searches the space of mathematical expressions for one that explains your data. Not the parameters of a fixed-form model — the formula itself: shape, operations, constants, all discovered.

Ufinq is Diafunc's Symbolic Regression technology, and it goes far beyond evolutionary search. Every candidate program Ufinq considers is a local optimum: its constants are tuned and its structure is simplified before it competes. Decompositional learning splits hard problems into pieces that can be solved independently and recomposed, bringing Symbolic Regression to scale. These are only foundations; there is much more.

A typical search
x1
x1 · x2
sin(x1) + x2
0.42 · sin(2π · x1) + 1.8 · x22 − 0.31

Generations of competing programs. One wins.

An expression tree rebuilding itself across generations — candidate programs evolving toward a discovered formula.

Diafunc's Symbolic Regression technology — research, papers, benchmarks — is at ufinq.com.

Visit ufinq.com