18. Deep Learning Models Can Do Both Feature Extraction and Classification

An advantage of Deep Learning Models (special type of Machine Learning models, built on the idea of neurons signaling in the brain to give us intelligence and govern our movements, thoughts, emotions, actions, personality, etc.) is that they do not need for you to do feature extraction beforehand. You can just input a bunch of data into the model...even labeled image data...and the deep learning models will learn the features and do the prediction! The only downside tends to be that sometimes we do want to know what the neurons ended up learning...what information and what features ultimately were extracted by them. And this may not be so easy to do :(
But deep learning has made great strides, such as with Alpha Fold and the protein prediction problem.

Please do subscribe to Saniya's YouTube channel and reach out with any and all questions :)