Searching for fundamental theories of Nature with unsupervised machine learning
Lead Research Organisation:
University of Sussex
Department Name: Sch of Mathematical & Physical Sciences
Abstract
In High Energy Particle Physics we contrast data with new theories of Nature. Those
theories are proposed to solve mysteries such as 1.) what is the Dark Universe made of, 2.)
why there is so much more matter than antimatter in the Universe, and 3.) how can a light
Higgs particle exist.
To answer these questions, we propose mathematical models and compare with
observations. Sources of data are quite varied and include complex measurements from
the Large Hadron Collider, underground Dark Matter detection experiments and satellite
information on the Cosmic Microwave Background. We need to incorporate all this data in
a framework which allows us to test hypotheses, and this is usually done via a statistical
analysis, e.g. Bayesian, which provides a measure of how well a hypothesis can explain
current observations. Alas, this approach has so far been unfruitful and is driving the field
of Particle Physics to an impasse.
In this project, we will take a different and novel approach to search for new physics. We
will assume that our inability to discover new physics stems from strong theoretical biases
which have so far guided analyses. We will instead develop unsupervised searching
techniques, mining on data for new phenomena, avoiding as much theoretical prejudices
as possible. The project has a strong theoretical component, as the candidate will learn the
mathematical/physical basis of new physics theories including Dark Matter, the Higgs
particle and Inflation. The candidate will also learn about current unsupervised-learning
techniques and the interpretation of data in High-Energy Physics.
The strategy adopted for this project holds the potential to open a new avenue of research
in High Energy Physics. We are convinced that this departure from conventional statistical
analyses mentioned above is the most effective way to discover new physics from the huge
amount of data produced in the Large Hadron Collider and other experiments of similar
scale.
Reaching the scientific goals outlined here would require modelling huge amounts of data
at different levels of purity (raw measurements, pseudo-observables, re-interpreted data),
and finding patterns which had not been detected due to a focus on smaller sets of
information. Hence, we believe that research into unsupervised learning in this context will
have far reaching applications beyond academic pursuits. As the world becomes
increasingly data-orientated, so does our reliance on novel algorithms to make sense of the
information we have in our possession. To give some examples, we can easily expect the
development of unsupervised learning integrated into facial recognition software and assist
in the discovery of new drugs, which provides a boost in the security and medical sector
respectively.
theories are proposed to solve mysteries such as 1.) what is the Dark Universe made of, 2.)
why there is so much more matter than antimatter in the Universe, and 3.) how can a light
Higgs particle exist.
To answer these questions, we propose mathematical models and compare with
observations. Sources of data are quite varied and include complex measurements from
the Large Hadron Collider, underground Dark Matter detection experiments and satellite
information on the Cosmic Microwave Background. We need to incorporate all this data in
a framework which allows us to test hypotheses, and this is usually done via a statistical
analysis, e.g. Bayesian, which provides a measure of how well a hypothesis can explain
current observations. Alas, this approach has so far been unfruitful and is driving the field
of Particle Physics to an impasse.
In this project, we will take a different and novel approach to search for new physics. We
will assume that our inability to discover new physics stems from strong theoretical biases
which have so far guided analyses. We will instead develop unsupervised searching
techniques, mining on data for new phenomena, avoiding as much theoretical prejudices
as possible. The project has a strong theoretical component, as the candidate will learn the
mathematical/physical basis of new physics theories including Dark Matter, the Higgs
particle and Inflation. The candidate will also learn about current unsupervised-learning
techniques and the interpretation of data in High-Energy Physics.
The strategy adopted for this project holds the potential to open a new avenue of research
in High Energy Physics. We are convinced that this departure from conventional statistical
analyses mentioned above is the most effective way to discover new physics from the huge
amount of data produced in the Large Hadron Collider and other experiments of similar
scale.
Reaching the scientific goals outlined here would require modelling huge amounts of data
at different levels of purity (raw measurements, pseudo-observables, re-interpreted data),
and finding patterns which had not been detected due to a focus on smaller sets of
information. Hence, we believe that research into unsupervised learning in this context will
have far reaching applications beyond academic pursuits. As the world becomes
increasingly data-orientated, so does our reliance on novel algorithms to make sense of the
information we have in our possession. To give some examples, we can easily expect the
development of unsupervised learning integrated into facial recognition software and assist
in the discovery of new drugs, which provides a boost in the security and medical sector
respectively.
