Efficient sampling of constrained high-dimensional theoretical spaces with machine learning

Hollingsworth, Jacob; Ratz, Michael; Tañedo, Philip; Whiteson, D.

doi:10.1140/epjc/s10052-021-09941-9

Cited by 14 publications

(12 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…• Another approach for revealing structures is by looking at generative techniques of effective field theories satisfying certain UV [55,56] or IR constraints [57]. Again it would be very useful to understand which methods are best suited to identify systematics in phenomenologically interesting string vacua.…”

Section: Discussionmentioning

confidence: 99%

Revealing systematics in phenomenologically viable flux vacua with reinforcement learning

Krippendorf,

Kroepsch,

Syvaeri

2021

Preprint

View full text Add to dashboard Cite

The organising principles underlying the structure of phenomenologically viable string vacua can be accessed by sampling such vacua. In many cases this is prohibited by the computational cost of standard sampling methods in the high dimensional model space.Here we show how this problem can be alleviated using reinforcement learning techniques to explore string flux vacua. We demonstrate in the case of the type IIB flux landscape that vacua with requirements on the expectation value of the superpotential and the string coupling can be sampled significantly faster by using reinforcement learning than by using metropolis or random sampling. Our analysis is on conifold and symmetric torus background geometries. We show that reinforcement learning is able to exploit successful strategies for identifying such phenomenologically interesting vacua. The strategies are interpretable and reveal previously unknown correlations in the flux landscape.

show abstract

Section: Discussionmentioning

confidence: 99%

Revealing systematics in phenomenologically viable flux vacua with reinforcement learning

Krippendorf,

Kroepsch,

Syvaeri

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Another attempt [7], also using Machine Learning models, is to use generative deep learning to produce likely valid points. The authors trained Normalising Flow Networks on a collection of valid points in order to learn their distribution to sample more, novel points, from the same distribution.…”

Section: (Re)framing the Problemmentioning

confidence: 99%

“…( 3), and allows us to use single-objective optimisation algorithms. 4 In this work we will use the same constraints as in [7], namely the mass of the Higgs boson, m h 0 , and dark matter relic density, Ω DM h 2 . The values of the upper and lower bounds can be seen in table I.…”

Section: (Re)framing the Problemmentioning

confidence: 99%

“…In this way one ensures automatically that one complies with the precise measurement of the Z-boson mass and that one is in a minimum of the potential where SU(2) L × U(1) Y is correctly broken. 7 In the following we will focus on the Higgs mass and the dark matter relic density as observables as already mentioned above. We summarise here a few key aspects of these observables as this will be helpful to understand some aspects of our findings.…”

Section: The Physics Models Consideredmentioning

confidence: 99%

“…Such task has seen recent efforts and developments of the employment of AI/ML algorithms to mitigate the burden of such scans in an attempt to increase sampling efficiency. Recent attempts at tackling this problem have deployed techniques such as deep neural networks [2,3] to try to guess if a new point is valid; Bayesian neural networks [4] to try to predict the observable value for a given parameter space point; active-learning methods [5,6] to find boundaries of valid subspaces; and generative models [7] used to re-sample from a collection of valid points. However, these efforts often require a large amount of data to be gathered previously for machine learning training -which presumably are hard to come by -before they can be used to suggest new points with high-efficiency, effectively not solving the sampling bottleneck.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Exploring Parameter Spaces with Artificial Intelligence and Machine Learning Black-Box Optimisation Algorithms

Souza¹,

Romão²,

Castro³

et al. 2022

Preprint

View full text Add to dashboard Cite

Constraining Beyond the Standard Model theories usually involves scanning highly multidimensional parameter spaces and check observable predictions against experimental bounds and theoretical constraints. Such task is often timely and computationally expensive, especially when the model is severely constrained and thus leading to very low random sampling efficiency. In this work we tackled this challenge using Artificial Intelligence and Machine Learning search algorithms used for Black-Box optimisation problems. Using the cMSSM and the pMSSM parameter spaces, we consider both the Higgs mass and the Dark Matter Relic Density constraints to study their sampling efficiency and parameter space coverage. We find our methodology to produce orders of magnitude improvement of sampling efficiency whilst reasonably covering the parameter space.

show abstract

Visualization and efficient generation of constrained high-dimensional theoretical parameter spaces

Baretz,

Carrara,

Hollingsworth

et al. 2023

J. High Energ. Phys.

View full text Add to dashboard Cite

We describe a set of novel methods for efficiently sampling high-dimensional parameter spaces of physical theories defined at high energies, but constrained by experimental measurements made at lower energies. Often, theoretical models such as supersymmetry are defined by many parameters, $$ \mathcal{O} $$ O (10 − 100), expressed at high energies, while relevant experimental constraints are often defined at much lower energies, preventing them from directly ruling out portions of the space. Instead, the low-energy constraints define a complex, potentially non-contiguous subspace of the theory parameters. Naive scanning of the theory space for points which satisfy the low-energy constraints is hopelessly inefficient due to the high dimensionality, and the inverse problem is considered intractable. As a result, many theoretical spaces remain under-explored. We introduce a class of modified generative autoencoders, which attack this problem by mapping the high-dimensional parameter space to a structured low-dimensional latent space, allowing for easy visualization and efficient generation of theory points which satisfy experimental constraints. An extension without dimensional compression, which focuses on limiting potential information loss, is also introduced.

show abstract

Efficient sampling of constrained high-dimensional theoretical spaces with machine learning

Cited by 14 publications

References 25 publications

Revealing systematics in phenomenologically viable flux vacua with reinforcement learning

Revealing systematics in phenomenologically viable flux vacua with reinforcement learning

Exploring Parameter Spaces with Artificial Intelligence and Machine Learning Black-Box Optimisation Algorithms

Visualization and efficient generation of constrained high-dimensional theoretical parameter spaces

Contact Info

Product

Resources

About