Using bootstrap two methods to create general knowledge representation frameworks

Dear all

The following article caught my attention:
https://arxiv.org/abs/2401.00350
Bootstrap method in theoretical physics.

According to a keyword search, " bootstrapping" refers to several methods
used in scientific research.

The first is statistical and is used in data science as well. Then there
are several bootstrapping methods, one which is used to determine geometric
space issues in conformal field theory and the article above mentioned
refers to another.

Since AI uses a lot of statistical modelling and algorithms for
(mathematical) pattern detection, my question is can bootstrapping be used
to come up with generalized models for knowledge representation

This is of particular importance if we want to detect generalized
theoretical models for unifying or incorporating multiple theories (each
with datasets available).

We can use hypergraph and category theory modeling for hypothetical
generalized frameworks and use both bootstraps methods, statistical and
physical, to come up with the most plausible (best fitting) generalized
categories by using geometric space modeling for expanding mathematical
frameworks.

For those familiar with the mathematical Langlands program it is about
building the bridges between theoretical frameworks and consequently
finding unified knowledge representation frameworks.

Bootstrapping is also used in biological sciences and I assume is or could
be used in neuroscience as well."

Any thoughts on this are welcome and also insights to whether any form of
bootstrapping is currently used in AI and for finding Biologically Inspired
Cognitive Architectures.


Milton Ponson
Rainbow Warriors Core Foundation
CIAMSD Institute-ICT4D Program
+2977459312
PO Box 1154, Oranjestad
Aruba, Dutch Caribbean