From the course: Training Neural Networks in C++ (2021)

Unlock the full course today

Join today to access over 22,400 courses taught by industry experts or purchase this course individually.

A simple model of a neuron

A simple model of a neuron - C++ Tutorial

From the course: Training Neural Networks in C++ (2021)

Start my 1-month free trial

A simple model of a neuron

- [Instructor] So based on the roles of the dendrites, nucleus and axons, let's just assign those roles to a set of inputs, a calculation unit and an output respectively to come up with the block diagram you can see at the right. We'll call the inputs x0 through xn minus one for a neuron with n inputs. Now, biological neurons seems to react in different sensitivities to different incoming signals. So we'll model that mathematically by assigning a weight to each input channel. Let's call them w0 through wn minus one. Next, the computation performed in order to fire or not is a simple weighted sum as shown inside the circle. And that's exactly what our output will be. Here's an example of how a neuron would help us classify. Let's suppose we have a two-input neuron and we feed data with two features into it. These features are shown in the plot at the right, which shows the risk of getting some medical condition, x, based on…

Contents