Models are a synthetic representation of the way that something in the real life works. There exists a function that describes every non-random occurrence in the universe. Knowing any particular real-world function is impossible without knowing its output for every input past and future. But estimating/approximating a function can be done if you have enough example outputs of that function. The estimation of this function is called a model.
In machine learning, models are the result of training. With parametric ML algorithms like neural networks, the model is the trained network parameters (the weights and biases, more on that soon). Essentially, models are:
Sometimes the word model and architecture are used interchangeably, however, model architecture usually describes the structure of the model (how many parameters, multiple combined models, depth of the model, etc.). A machine learning model usually refers to the product of training; the learned algorithm.
Next: The Macine Learning Pipeline
Previous: Data is Key
Return to the main page.
All source code in this document is licensed under the GPL v3 or any later version. All non-source code text is licensed under a CC-BY-SA 4.0 international license. You are free to copy, remix, build upon, and distribute this work in any format for any purpose under those terms. A copy of this website is available on GitHub.