Skip to content

What is a Perceptron?

The perceptron is an algorithm from the supervised learning field and represents a Neural Network‘s basic building block. In its simplest form, it consists of a single neuron that calculates an output value using weighted input values.

How does a Perceptron work?

The perceptron is originally a mathematical model and was only later used in computer science and Machine Learning due to its ability to learn complex relationships. In its simplest form, it consists of exactly one so-called neuron, which imitates the structure of the human brain.

The perceptron has several inputs at which it receives numerical information, i.e. numerical values. Depending on the application, the number of inputs can differ. The inputs have different weights, which indicate how influential the inputs are for the final output. During the learning process, the weights are changed to produce the best possible results.

Aufbau eines Perceptrons
Exemplary Structure of a Perceptron

The neuron itself then forms the sum of the input values multiplied by the weights of the inputs. This weighted sum is passed on to the so-called activation function. In the simplest form of a neuron, there are exactly two outputs, so only binary outputs can be predicted, for example, “Yes” or “No” or “Active” or “Inactive”, etc.

Dieses Diagramm zeigt die Sigmoid Funktion mit der charakteristischen S-Form zwischen den Werten 0 und 1 in Anlehnung an Artificial Neural Networks.
Course of the Sigmoid Function

If the neuron has binary output values, a function is used whose values also lie between 0 and 1. An example of a frequently used activation function is the sigmoid function. The values of the function vary between 0 and 1 and actually take these values almost exclusively. Except for x = 0, there is a steep increase and the jump from 0 to 1. Thus, if the weighted sum of the perceptron exceeds x = 0 and the perceptron uses sigmoid as an activation function, the output also changes accordingly from 0 to 1.

Which Example illustrates how a Perceptron works?

As an example of how a perceptron works, let’s take a closer look at the work of a politician. She is a member of parliament and a new law has to be voted on. Thus, the politician has to decide whether she agrees or disagrees with the bill (abstention is not possible in our example). The perceptron thus has a binary output, namely approval or rejection.

There are various sources of information available to politicians as inputs for their decision. On the one hand, there is an information paper with background information issued by the parliament. Furthermore, the politician can inform herself about various issues on the Internet or discuss them with colleagues. The politician weights her input, i.e. her sources of information, according to how trustworthy she considers them to be. She assigns a relatively low weight to the parliamentary information paper, for example, because she fears that the research is not detailed enough and should already tend in a certain direction. She then takes the sum of the information available to her, along with the weights, and passes it on to the activation function.

In this example, we can imagine this as the head of our politician. She now decides, on the basis of the inputs, whether she should agree to the proposed law or not. Even small details in the inputs can lead to a massive change in the politician’s opinion.

What is a Multi-Layer Perceptron?

This simple model can already be used for many applications, but quickly reaches its limits when it comes to understanding and learning more complex relationships. Therefore, in practice, the focus is primarily on several layers of perceptrons, which are then assembled to form a so-called Artificial Neural Network.

These so-called multi-layer perceptrons are connected to each other in such a way that only the first layer of neurons receives inputs from outside the network. In the subsequent layers, on the other hand, the outputs of the previous layers serve as inputs for the current layer. Only the neurons in the last layer of the Neural Network are outputs, which are then used as the actual result.

Das Bild zeigt den Grundaufbau eines Künstlichen Neuronalen Netzwerkes mit den verschiedenen Schichten.
Multi-Layer Perceptron

Related to our policy example, we could look more closely at the politician’s sources of information. For example, the parliament’s information document is composed of the work of various research assistants whose weighted work has been incorporated into the final product. In turn, these collaborators could also be represented as individual perceptrons in the network.

What are the Bias Neurons?

In multi-layer perceptrons, so-called bias neurons are used. These are special input neurons, which, however, have no external meaning, but assume a constant value that can also change during the learning process. This allows the curve of the activation function to be shifted downward or upward.

The bias is similar to the constant used in linear regression, for example, and helps the model to fit better to the output values and thus improve the result.

Which Applications use Perceptrons?

Neural Networks are based on perceptrons and are mainly used in the field of Machine Learning. The goal here is mainly to learn structures in previous data and then predict new values. Some examples are:

  • Object Recognition in Images: Artificial Neural Networks can recognize objects in images or assign images to a class. Companies use this property in autonomous driving, for example, to recognize objects to which the car should react. Another area of application is in medicine when X-ray images are to be examined to detect an early stage of cancer, for example.
  • Prediction: If companies are able to predict future scenarios or states very accurately, they can weigh different decision options well against each other and choose the best option. For example, high-quality regression analysis for expected sales in the next year can be used to decide how much budget to allocate to marketing.
  • Customer Sentiment Analysis: Through the Internet, customers have many channels to make their reviews of the brand or a product public. Therefore, companies need to keep track of whether customers are mostly satisfied or not. With a few reviews, which are classified as good or bad, efficient models can be trained, which can then automatically classify a large number of comments.
  • Spam Detection: In many mail programs there is the possibility to mark concrete emails as spam. This data is used to train Machine Learning models that directly mark future emails as spam so that the end user does not even see them.
  • Analysis of Stock Prices: Neural Networks can also be used to predict the development of a stock based on previous stock prices. Various influencing variables play a role here, such as the overall economic situation or new information about the company.

What are the Advantages and Disadvantages of Perceptrons?

The use of perceptrons is characterized by the following advantages and disadvantages.

Advantages

  • With the help of single-layer perceptrons and especially multi-layer perceptrons, the so-called Neural Networks, complex predictions can be learned in the field of supervised learning.
  • Depending on the data, the number of inputs can be easily increased to include different, relevant values in the prediction.
  • A trained perceptron is relatively easy to interpret and the learned weights can be used to make a statement about how important the inputs are.

Disadvantages

  • Single perceptrons rarely provide really good results. For powerful models, different layers with many neurons usually have to be built up.
  • Especially multi-layer perceptrons only work really well with large data sets.
  • Training multi-layer perceptrons is usually time-consuming and resource-intensive.
  • In many layers, the interpretability of the weights is lost and a “black box” develops whose good predictions cannot really be explained.

This is what you should take with you

  • The perceptron is an algorithm from the field of supervised learning and represents the basic building block of a neural network.
  • When individual perceptrons are built and connected in multiple layers, it is called a multi-layer perceptron or a neural network.
  • The perceptron consists of the inputs, the weights, the activation function, and the outputs.
  • It can be used to learn complex relationships in data and apply them to new, previously unseen data.

Other Articles on the Topic of Perceptrons

  • A detailed explanation of perceptrons and their implementation in JavaScript can be found at w3schools.
close
Das Logo zeigt einen weißen Hintergrund den Namen "Data Basecamp" mit blauer Schrift. Im rechten unteren Eck wird eine Bergsilhouette in Blau gezeigt.

Don't miss new articles!

We do not send spam! Read everything in our Privacy Policy.

Cookie Consent with Real Cookie Banner