Thread by @predict_addict

Understanding the Historical Divide Between Machine Learning and Statistics

On social media, it's common to encounter strong reactions to statements like "Machine learning is not statistics."

Much of this stems from a lack of historical context about the development of ML as a field.

It's important to note that the modern foundations of machine learning were largely shaped in places like the USSR and the USA—not the UK.

While Alan Turing’s legacy is significant, the UK's direct contributions to core ML theory during the 20th century were almost non-existent.

For example, the first dedicated machine learning department in the UK, founded at Royal Holloway (RHUL), was built by prominent figures from elsewhere—Vladimir Vapnik and Alexey Chervonenkis from the USSR, Ray Solomonoff from the US, and others.

To clarify the distinction:

On one side, you had the Institute of Control Sciences in the USSR, a powerhouse that developed many of ML’s theoretical foundations—Statistical Learning Theory, VC dimension, Support Vector Machines, and kernel methods.

On the other side, the Central Economic Mathematical Institute (CEMI) focused on statistics and economics, producing luminaries like Nobel laureate Leonid Kantorovich. However, their work was not directly tied to machine learning.

The historical separation between ML and statistics wasn’t just geographical—it was conceptual. ML papers were often rejected by traditional statistics journals because the two fields had different goals, methodologies, and assumptions.

So the next time someone confidently asserts that "statistics is ML," it may be worth encouraging them to look at the actual history of both disciplines before drawing conclusions.

#machinelearning #statistics