Python math.exp(): Natural Exponential Function

Syntax & Parameters 🔝

The math.exp() function is part of the standard math module:

import math

result = math.exp(x)

Return Value: Returns a floating-point number (float) representing $e^x$.

Basic Examples 🔝

Evaluating $e^x$ with zero, positive, and negative values:

import math

print("e^0: ", math.exp(0))   # Output: 1.0 (any non-zero number to power 0 is 1)
print("e^1: ", math.exp(1))   # Output: 2.718281828459045 (Euler's number e)
print("e^5: ", math.exp(5))   # Output: 148.4131591025766
print("e^-2:", math.exp(-2))  # Output: 0.1353352832366127 (1 / e^2)

e^0:  1.0
e^1:  2.718281828459045
e^5:  148.4131591025766
e^-2: 0.1353352832366127

math.exp(x) vs. math.e ** x 🔝

While you can mathematically compute $e^x$ using Python's exponentiation operator (math.e ** x) or the built-in pow(math.e, x), math.exp(x) is superior:

• Accuracy: math.e ** x first rounds math.e to a 53-bit floating-point constant and then raises it to the power, compounding rounding errors. math.exp(x) uses direct hardware-optimized polynomial approximations from the C standard library.
• Performance: math.exp(x) is significantly faster because it evaluates base $e$ directly in compiled machine instructions rather than generic multi-base exponentiation.

Handling OverflowError & Extreme Negative Underflow 🔝

Because exponential growth accelerates rapidly, values of $x > 709.78$ exceed the maximum number representable by 64-bit IEEE 754 floating-point numbers (≈ $1.79 imes 10^{308}$), raising an OverflowError:

import math

try:
    val = math.exp(710)
except OverflowError as err:
    print("Overflow error caught:", err)
# Output: Overflow error caught: math range error

Conversely, for large negative numbers ($x < -745$), $e^x$ approaches zero so closely that Python cleanly underflows to 0.0 without raising an exception:

print(math.exp(-800))  # Output: 0.0

High-Precision expm1() for Values Near Zero 🔝

When $x$ is extremely close to zero (e.g. $10^{-15}$), calculating $e^x - 1$ using math.exp(x) - 1 suffers from catastrophic cancellation, resulting in zero or losing trailing digits. Use math.expm1() for full numerical precision:

import math

tiny = 1e-15
print("Using exp() - 1:  ", math.exp(tiny) - 1)   # 1.1102230246251565e-15 (imprecise)
print("Using expm1():     ", math.expm1(tiny))       # 1.0000000000000005e-15 (accurate!)

Hands-on Exercise: Sigmoid Activation Function with Assertions 🔝

In machine learning and neural networks, the logistic sigmoid activation function squashes any real-valued number into a probability between $0$ and $1$: $$\sigma(x) = rac{1}{1 + e^{-x}}$$ Let's build a numerically stable sigmoid function with math.exp() and verify it with unit assertions:

import math

def sigmoid(x):
    # Numerical stability guard against overflow
    if x < -700:
        return 0.0
    elif x > 700:
        return 1.0
    return 1.0 / (1.0 + math.exp(-x))

# Unit Assertions
assert sigmoid(0) == 0.5
assert round(sigmoid(2), 4) == 0.8808
assert round(sigmoid(-2), 4) == 0.1192
assert sigmoid(-800) == 0.0
assert sigmoid(800) == 1.0

print("Sigmoid(0): ", sigmoid(0))
print("Sigmoid(2): ", round(sigmoid(2), 4))
print("Sigmoid(-2):", round(sigmoid(-2), 4))
print("All sigmoid activation assertions passed successfully!")

Practice in Google Colab & GitHub 🔝

Open in Google Colab View on GitHub

Run all math.exp() examples interactively, test continuous growth formulas, prevent OverflowError, and verify the sigmoid activation function with unit assertions directly in Google Colab.

More Details & Related Guides 🔝

• Python Math Module Overview — Master arithmetic, exponential, and trigonometric functions in Python.
• math.log() Function — Natural logarithm and inverse of math.exp().
• math.expm1() Function — High-precision $e^x - 1$ calculation for numbers close to zero.
• pow() Function — Raising base numbers to fractional and integer powers.
• float() Constructor — Converting numbers and strings into IEEE 754 floats.
• Exception Handling in Python — Catching OverflowError, ValueError, and math domain errors.

Subhendu Mohapatra

Author & Instructor at plus2net

I write and maintain practical tutorials on Python, PHP, SQL, JavaScript, HTML, jQuery, and web development at plus2net. The tutorials focus on clear explanations, working examples, and code that readers can test and adapt while learning.

About the author   Watch on YouTube

Connect: