I aim to share the notes in this series about my deep learning. I hope that will be useful.
The main difference between deep learning and standard machine learning applications is that they need very large data for quality results. So, in fact, we can call it a multi-layered neural network for deep learning. It needs more data set. If deep learning is a rocket, this rocket’s fuel is data.
There are a few elements to consider when choosing a house. The quality of the schools around you, your purchasing power, the number of rooms, the walking distance to your work etc. elements. These elements are our inputs.
Supervised Learning and Neural Networks
Only one neural network model is not used in every problem. Data may show unexpected variations. The following are some data and the network models used for the classification of these data.
Structured and Unstructured Data
Users’ information and their ad click status or house sizes and the prices of these houses, these are structured data. However, an audio file or an image file is unstructured data.
Example: The frequencies in an audio file are constantly changing, and you need a neural network that can keep up with this variability, which can repeat itself.
The hardware is evolving every day and processing these data is easier than before. So thanks to gamers and bitcoin miners. 🙂