\approx

Approximately equal

Notation: structural shorthand

In plain English

"Close enough for the purpose at hand." $\pi \approx 3.14$ . In papers, $\approx$ often marks a deliberate simplification: dropping small terms, using a first-order Taylor expansion, or quoting a fitted value. Related symbols: $\sim$ ("of the same order of magnitude" or "distributed as"), $\simeq$ (asymptotically equal), $\lesssim$ (less than, up to a constant).

As code

const close = (a, b, tol = 1e-3) => Math.abs(a - b) <= tol * Math.max(1, Math.abs(b));
close(Math.PI, 3.1416); // true

Where the analogy bends: In code you choose the tolerance. In a paper ≈ often leaves the tolerance unstated, so ask how good the approximation is.

See it

x (radians)0.75 · error 10.0%

\sin x \approx x

for small

x

. An approximation always has a regime where it holds. Slide right and watch the error grow.

What it means, field by field

Physics

Small-angle approximation sin θ ≈ θ; non-relativistic limits; perturbation theory.

AI / ML

C ≈ 6ND training FLOPs; fitted exponents (α ≈ 0.076); stochastic estimates of a gradient.

Biology

Rounded rates, fitted exponents, simplified kinetics (Michaelis–Menten approximations).

Epistemics

Every ≈ hides an error bar. Find it or ask for it.

In a research paper

Rule of thumb for transformer training compute in FLOPs (Kaplan et al. 2020), with N parameters and D training tokens.

C \approx 6\,N D

Training costs about six floating-point operations per parameter per token: two for the forward pass, four for the backward pass.

Common misreadings

Treating an ≈ result as exact downstream and stacking approximations without tracking error.
Reading ~ always as "approximately". In probability, x ~ p means "x is drawn from distribution p".

Connected symbols

\propto

Proportional to

\Delta

Delta (change / difference)

\infty

Infinity

\mathbb{E}

Expectation (expected value)

Where it appears on this site

§ 1. Learning to Read Equations § 2. The Physics of Systems § 3. Biology as an Information System § 5. Complexity and Systems Thinking § 7. The Physics–Biology–AI Connection Lab § 8. The Scientific Skeptic’s Toolkit