Deep learning is part of machine learning that is more sophisticated because it utilizes neural networks with more hidden layers. With deep learning, a computer program or tool can learn from experience and generate modeling independently. However, it should be understood that the implementation of deep learning cannot be done haphazardly. To be realized, it is necessary: (1) A large amount of data (big data); and (2) Support of sophisticated and powerful computing machines.
There are differences between deep learning for vision systems and machine learning that you need to know. Machine learning is a form of Artificial Intelligence that in its work process must depend on the programmer (human). To determine which data contains an image of a cat, for example, machine learning needs to be told the characteristics that indicate data as a cat. This is distinguishable from deep learning where the computer timetable or device that utilizes it does not require to be told in advance about the symbols that reveal data is an idea of a cat or not.
The advantages and disadvantages of deep learning are as follows:
– Features in data can be learned on their own, unlike machine learning
– High variation in data can be learned better and is automated
– Flexible and can be used to study various new phenomena or in different fields
– Requires large amounts of data
– Expensive and need high power computer support
– More difficult to learn than machine learning
How does deep learning work? Have you ever watched children learning the names of objects? Computer programs or tools that use deep learning work like a human child. Because the neural network, which is the core of this technology, is inspired by the work of biological neurons that can work autonomously. In front of the neural network in the form of spheres, a series of data called training data will be given.