MathLiveSoftmax function
Turns a vector of K real numbers into a vector of K positive values that sum to one, so they can be read as probabilities. It is widely used as the output layer of classification models such as neural networks.
\sigma(\mathbf{z})_j = \frac{e^{z_j}}{\sum_{k=1}^{K} e^{z_k}}What the symbols mean
| input vector of K real-valued scores (logits) | |
| probability assigned to the j-th class, between 0 and 1 | |
| score of the j-th component of the input vector | |
| score of the k-th component, summed over all K components | |
| number of classes or components in the vector | |
| Euler's number, the base of the natural exponential |
About this formula
- FieldComputer science
- TopicMachine learning
- LevelUniversity
- Also callednormalized exponential function, softargmax, softmax activation function
- SourceWikidata Q7554146
Write it in Obsidian
MathLive is a free visual math editor for Obsidian. Run Find a formula, type “softmax function”, and it goes into your note as editable LaTeX. You can also scan formulas from screenshots, PDFs and handwriting.