Introduction to AI/ML

Terminology

Intelligence, AI, ML, GenAI and data science: what the words mean and how they relate.

What is AI?

Mimicry of human intelligence

So first: what is intelligence?

What is intelligence?

Definitions of intelligence by Binet, Wechsler, Wagnon and Sternberg

Intelligence as an emergent system's ability to respond to its environment, from Future Minds by Richard Yonck

Intelligence = cognitive abilities

Pattern Recognition Abstract Reasoning Problem-Solving Language Comprehension Memory Capacity to Learn and Un-learn Emotional Intelligence Creativity Imagination Planning Self-Awareness Adaptability Social Intelligence

Machine learning is mostly about one of them: pattern recognition.

How to test human intelligence?

A shape-arithmetic puzzle: triangles, circles and squares summing to numbers

A puzzle of shaded circles, squares and triangles adding to numbers

A matrix-reasoning puzzle with one missing tile

Every one of these is a pattern to find.

Are LLMs intelligent?

ChatGPT 4o says there are two Rs in strawberry

ChatGPT 5 thinks for 11 seconds and says the letter R has 3 strawberries

Level of mimicry

Artificial narrow intelligence, then general intelligence, then super intelligence

What is ML?

A computer program is said to learn from experience E with respect to some class of tasks T and performance measure P, if its performance at tasks in T, as measured by P, improves with experience E.— Tom Mitchell (1997)

The field of study that gives computers the ability to learn without being explicitly programmed.— Arthur Samuel (1959)

Finding patterns.— History

ML vs programming

What is GenAI?

Sketch: input and pattern go into a normal program, output comes out

Sketch: input and output go into an ML program, a pattern comes out

Sketch: a pattern goes into a GenAI program, input and output come out

How are these terms related?

The same picture, from two sources

The AI Universe: nested circles from artificial intelligence down to generative AI — Source

Artificial intelligence containing machine learning, deep learning and symbolic AI — Source

What about data science?

Venn diagram: data science overlapping AI, ML and deep learning

Nested ovals: data science containing AI, ML and deep learning

Data science where AI meets maths, statistics, visualisation and EDA

There is no single agreed picture: different authors draw the boundary differently.

ML foundations and pillars

Machine learning as a temple: regression, dimensionality reduction, density estimation and classification on foundations of linear algebra, probability, optimisation and more — Source: Deisenroth, M. P., Faisal, A. A., & Ong, C. S. (2020). *Mathematics for Machine Learning*. Cambridge University Press.