Deep learning is a rebranding for a concept called Neural Networking, and is based off of predictions. We supply the training data for a complex model and the machine takes this input, and supply’s us with a function that would predict the model we are aiming for. Then, once this functions applied to new data, it helps predict outcomes we as scientist can use.
Most of the workflow is spent wrangling data, and ensuring each part has appropriate weight. Once the model works, it can be used in everything from MRI classification to market predictions. If this interests you, learn more at groups like Auckland’s AI Meetup or UoA’s Data Science Club!