Generalisation (and its zibling)

Generalisation… I am comparing State Spaces and Solution Spaces and realised that I may be talking about generalisation….

From Oxford Language

a general statement or concept obtained by inference from specific cases

From Wikipedia

A generalization is a form of abstraction whereby common properties of specific instances are formulated as general concepts or claims. Generalizations posit the existence of a domain or set of elements, as well as one or more common characteristics shared by those elements (thus creating a conceptual model). As such, they are the essential basis of all valid deductive inferences (particularly in logic, mathematics and science), where the process of verification is necessary to determine whether a generalization holds true for any given situation.

My view (that may change…) in relation to AI

It is my view that no model can generalize, this includes humans, we are not capable of pure generalisation and need to learn and adapt when our environment is different. The adaptability comes from a feedback loop that enables us to build and update models based on new states in the “test time” state space, this is the key to “generalisation” of models in AI.

And here we have a more detailed explanation of generalisation in learning.

I still prefer to view it as an adaptation, using prior knowledge as a malleable tool.

This paper poses a direct, high level challenge, to my view

Generalization is also considered to be an important factor in procedural memory, such as the near-automatic memory processes necessary for driving a car.

Banich, M. T., Dukes, P., & Caccamise, D. (2010). Generalization of knowledge: Multidisciplinary perspectives. Psychology Press.