What is called deep-learning
UNDERSTANDING DEEP-LEARNING IS EASY!
Machine learning corresponds to algorithms that adjust the parameters of their calculations according to the examples given to them. This makes it possible to adapt their operation to the data provided. Deep learning is an architecture that cascades several layers of such algorithms to prioritize the problem and achieve much higher performance. It’s not simple ! But to explain, if …
- Machine Learning: machine learning corresponds to algorithms that adjust the parameters of their calculations according to the examples given to them. This makes it possible to adapt their operation to the data provided.
- Deep Learning: deep learning is an architecture that cascades several layers of algorithms of this type in order to prioritize the problem and obtain much higher performances.
- Artificial Neuron: A function that combines inputs to calculate whether the output value is rather high or low, its calculation being adjusted by parameters.
- Neural network (artificial): the assembly of a large number of neurons to allow a very sophisticated input / output calculation.
- GPU: graphical processing unit, an additional processor in our computers in which all calculations related to graphical display operations are pre-wired to speed up processing
- Computer vision: a set of algorithms that start from the pixel values of the images to extract characteristics, locate the objects that are seen there, and label them.
ABOUT THE AUTHOR
David Louapre is a normalist and doctor in theoretical physics, now a researcher in the materials industry. Passionate about scientific culture, his career has allowed him to take a deep approach to many areas of science (fundamental physics, applied mathematics, physico-chemistry of materials, thermal, mechanical, etc.). He declares himself « Responsible R & D (day), scientific popularizer (the night) ».
One can follow his sharing of grains of « amazing science » on his blog and his video channel.