Home / Maths for AI / Symbols / Machine learning notation

Theta (parameters)

Say it: "theta"

What it means

The usual name for all of a model's learnable parameters (weights and biases) together. fθ(x)f_\theta(x) is "the model with parameters θ\theta".

Example

θ∗=arg⁡min⁡θ1n∑i=1nL(fθ(xi),yi)\theta^* = \arg\min_\theta \frac{1}{n}\sum_{i=1}^{n} L(f_\theta(\mathbf{x}_i), y_i)

Training finds the best parameters.

In code

import numpy as np
from sklearn.linear_model import LinearRegression
X = np.array([[1.0], [2.0], [3.0]]); y = np.array([3.0, 5.0, 7.0])
m = LinearRegression().fit(X, y)
print(m.coef_, m.intercept_)   # the learned theta

Output

[2.] 1.0000000000000009

Where it appears in machine learning

Every training objective is written in terms of θ\theta.

Related symbols

Learn it properly in the free lesson Optimisation, with practice questions that check your answers.