Hi,
Can anyone help with the difference between
AI, ML and DL in simple terms.
This is just one of many mostly similar representations out on the interweb. The DL vs ML boundary is probably the easiest and least contentious; DL is a subset of ML that focuses on ‘deep’ neural networks where features are algorithmically extracted from large, often very large data sets. ML is a collection of algorithms that also learn or train using data, but maybe not as large data sets and maybe more direct human input and direction in selecting features. Notice that this picture shows Linear Neural Network in the ML region. Deep implies multiple layers in the neural network, especially what are referred to as hidden because they sit in between the input and output layers.
There is less agreement about what exactly AI implies and the AI/ML boundary. Some people like to append … like a human does to definitions of AI. For me, one starts to move to AI based on complexity and autonomy of a system that uses one or more ML components to make choices, regardless of whether the problem solving is performed like a human does or not. If a machine exhibits complex data-driven novel behavior but makes decisions not the way a human does, is it therefore not intelligent? 
FYI, this and similar questions are asked and addressed frequently, and you can find and read other threads using the search function. Here’s just one of them…
1 Like