Artificial intelligence is a computer program is capable of making a decision such that it performance in particular task increase constantly without any human intervention. There are two main type of branches in artificial intelligence that are learning and non-learning.
Non learning logic like fuzzy logic, system engineering, genetic programing, optimisation etc. do not have a learning phase. Fuzzy logic is based on comparing a spectrum of real number values between 0 and 1 not just 0 and 1 as in the case of Boolean programming. The values of 0 and 1 is only used for extreme cases. System engineering involves a domain expert coding the rules how to solve the particular problem. While genetic programing is based on evolution, where different program compete on the problem and where there performance is compared using a fitness function and the using a selection process these new program compete against their offspring and their mutated program.
In recent year optimisation and game theory are emerging area of research in non-learning artificial intelligence model even betting expert human and deep learning models.
Learning model is mainly dominated machine learning. Machine learning is mainly defined as a program which improve its performance as measured by a metric M on task T with respect to experience E. There are many algorithm used in machine learning like linear regression, polynomial regression, support vector machine, decision trees, t-SNE,PCA,KNN etc. Machine learning is further divided into supervised learning where in learning phase the output is available and non-supervised learning where in learning phase the output is not available. Non supervised machine learning is mainly used for clustering and visualization while supervised learning id used mainly for decision making. Machine learning model require data to adjust the weight of the variables in the algorithm.
Deep learning is a special kind of machine learning which uses stacked neural networks and lots of data. It has many algorithm under it like RNN, CNN, RBM, GAN etc. Deep learning is further divided into generative networks and discriminative networks. While data science is the science of collecting, pre-processing and describing data and using to inform decision making by either gaining new insight or testing a new hypothesis.
So how are deep learning, machine learning and data science related? In the case of deep learning and machine learning, the answer is clear deep learning is part of machine learning. Also many of the algorithm for algorithmic modelling for data science are common with machine learning, so in a way we can say the data science and machine learning intersect with each other.