What is Deep Learning?

Deep learning is a machine learning method based on artificial intelligence. With deep learning, machines are able to learn by themselves. That differs from programming, where they simply execute predetermined rules.

Deep learning relies on artificial neural networks inspired by the human brain. These networks are made up of tens or hundreds of layers of neurons, each receiving and interpreting information from the preceding layer. Deep learning is based on these algorithms, as well as image recognition and robotic vision. The higher the number of neurons, the deeper the network. The system learns to recognise letters and words in text, or faces in photos, which it can then identify.

Let’s imagine we want to use a neural network to recognise photos in which at least one cat is present. In order to identify cats in the images, the algorithm needs to be able to distinguish different types of cat, and to recognise the cat no matter what angle the picture is taken from.

To do this, the neural network has to be trained. It has to compile a selection of images to practice deep learning. This would be made up of thousands of examples of different cats, mixed with images of objects which are not cats. These images are then converted into data and transferred onto the network. The artificial neurons assign a weight to each different element. The final layer of neurons reassembles all the information to determine if it is indeed a cat or not.

The neural network then compares this response to correct answers given by humans. If the answers match, the network remembers this success and will reuse it later in order to recognise cats. If wrong, the network also notes the error and adjusts the weight on the various neurons to correct it. The process repeats thousands of times until the network is able to recognise a cat in an image without fail.

See acast.com/privacy for privacy and opt-out information.

Learn more about your ad choices. Visit megaphone.fm/adchoices