Artificial intelligence is
a way of programming that, through a process of learning and improving performance like a person,
lets a system carry out human-like communication and tasks.
Machine learning is
a programming method within AI that performs classification or prediction based primarily on statistical techniques,
where the required learning-answer data or learning algorithm is designed by humans.
Deep learning is
a programming method within machine learning that performs classification and prediction based on neural networks,
where the needed data and algorithms are designed by the neural network itself, which then produces the result.
Overall, deep learning's performance is considerably ahead of machine learning's.
That said, with machine learning you can trace the steps and the process, while with deep learning that is not easy,
and getting to a finished model requires enormous computing power and time.
That's why most large-scale tech firms lead the tech, and the small- and mid-sized tech firms
tend to customize or reuse what the big firms have already built — paid/free APIs and infrastructure released to the public.
The rich-get-richer gap is widening beyond the point of being caught up. But we shouldn't lose sight of either the macro view — that most tech firms are not nationally-bound or nationally-operated — or the micro? view — that AI, machine learning, and deep learning all rely on data, and that data ultimately rests on people's lives (lifestyles).
