I am a bit confused when can I say exactly that I use Machine Learning (ML) to solve my problem (Objective function) and what is the difference between using ML or using (Linear Algebra, Statistics, Optimization, heuristics, ..etc) - in other words what is new in ML ??
Machine learning is a category of many different methods where you want a computer to find structure in data, particularly, such structures that are hard for humans to model.
Deep learning is particularly popular these days. It is essentially a way to approximate an unknown function. You provide many examples of input data to the "function" and the corresponding desired outputs, and then train a neural network to give the right outputs (to approximate the function).
Linear algebra, optimization, and statistics are three topics that are utilized in deep learning to train the neural networks.
Machine learning will not replace classical methods, but it can be used as a complement to design heuristic solutions to problems for which we don't have any good solutions already. In most cases, you need a deep understanding of the problems that you need to solve, otherwise you cannot expect to get any good results from deep learning.
Machine learning is a category of many different methods where you want a computer to find structure in data, particularly, such structures that are hard for humans to model.
Deep learning is particularly popular these days. It is essentially a way to approximate an unknown function. You provide many examples of input data to the "function" and the corresponding desired outputs, and then train a neural network to give the right outputs (to approximate the function).
Linear algebra, optimization, and statistics are three topics that are utilized in deep learning to train the neural networks.
Machine learning will not replace classical methods, but it can be used as a complement to design heuristic solutions to problems for which we don't have any good solutions already. In most cases, you need a deep understanding of the problems that you need to solve, otherwise you cannot expect to get any good results from deep learning.
- Badges
- Science topic