Artificial Intelligence and Machine Learning in Physics
Contents
What is this talk about?
Many to thank
A simple perspective on the interface between ML and Physics
ML in Nuclear Physics
AI/ML and some statements you may have heard
Scientific Machine Learning
Physics driven Machine Learning
Machine Learning and Physics
Lots of room for creativity
Types of machine learning
The plethora of machine learning algorithms/methods
Examples of applications of ML in physics
And more
Selected references
What are the basic ingredients?
"Unsupervised learning in nuclear physics, Argon-46 by Solli, Bazin, Kuchera, MHJ, Strauss, NIMA 1010, 165461 (2020)":"https://www.sciencedirect.com/science/article/abs/pii/S0168900221004460?via%3Dihub"
Quantum Monte Carlo and deep learning
Monte Carlo methods and Neural Networks
Deep learning neural networks, "Variational Monte Carlo calculations of \( A\le 4 \) nuclei with an artificial neural-network correlator ansatz by Adams et al.":"https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.127.022502"
"Gnech et al, Variational Monte Carlo calculations of \( A\le 6 \) nuclei Few Body Systems 63, (2022)":"https://link.springer.com/article/10.1007/s00601-021-01706-0"
Neutron matter with \( N=14 \) neutron
The electron gas in three dimensions with \( N=14 \) electrons (Wigner-Seitz radius \( r_s=4 \) Ry)
Extrapolations and model interpretability
Physics based statistical learning and data analysis
Bayes' Theorem
Quantified limits of the nuclear landscape
Constraining the equation of state for dense nuclear matter
Observations (or conclusions if you prefer)
Possible start to raise awareness about ML in your own field
Education
QuSTEAM Model
Possible courses
Important Issues to think of
Observations
Future Needs/Problems
Educational initiatives since 2019 in nuclear physics
Additional material
Examples of Machine Learning methods and applications in nuclear physics
Machine learning and nuclear theory: Why?
Examples of Machine Learning methods and applications in nuclear physics, continues
Addendum: Quantum Monte Carlo Motivation
Quantum Monte Carlo Motivation
Quantum Monte Carlo Motivation
The trial wave function
The correlation part of the wave function
Resulting ansatz
Energy derivatives
Derivatives of the local energy
How do we define our cost function?
Meet the variance and its derivatives
The variance defines the cost function
Why Boltzmann machines?
A standard BM setup
The structure of the RBM network
The network
Joint distribution
Defining different types of RBMs
Representing the wave function
Choose the cost/loss function
Electrons in a harmonic oscillator trap in two dimensions
Quantum dots and Boltzmann machines, onebody densities \( N=6 \), \( \hbar\omega=0.1 \) a.u.
Onebody densities \( N=30 \), \( \hbar\omega=1.0 \) a.u.
Onebody densities \( N=30 \), \( \hbar\omega=0.1 \) a.u.
Neutron matter with \( N=14 \) neutron
Neutron matter with \( N=14 \) neutron
Neutron matter with \( N=14 \) neutron
The electron gas in three dimensions with \( N=14 \) electrons (Various Wigner-Seitz Radii)
The electron gas in three dimensions with \( N=14 \) electrons (Various Wigner-Seitz Radii)
The electron gas in three dimensions with \( N=14 \) electrons (Various Wigner-Seitz Radii)
The electron gas in three dimensions with \( N=14 \) electrons (Various Wigner-Seitz Radii)
The electron gas in three dimensions with \( N=14 \) electrons (Various Wigner-Seitz Radii)
The electron gas in three dimensions with \( N=14 \) electrons (Various Wigner-Seitz Radii)
The electron gas in three dimensions with \( N=14 \) electrons (Various Wigner-Seitz Radii)
The electron gas in three dimensions with \( N=14 \) electrons (Various Wigner-Seitz Radii)
Neutron matter with \( N=14 \) neutron
Folding and unfolding and response functions
Experimental design
Selected references
Mehta et al.
and
Physics Reports (2019)
.
Machine Learning and the Physical Sciences by Carleo et al, Rev Mod Phys 91, 045002
Ab initio solution of the many-electron Schrödinger equation with deep neural networks by Pfau et al., Physical Review Research 2, 033429
Report from the A.I. For Nuclear Physics Workshop by Bedaque et al., Eur J. Phys. A 57, (2021)
Particle Data Group summary on ML methods
«
1
...
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
...
78
»