Comparing Machine Learning and Interpolation Methods for Loop-Level Calculations

Abstract

The need to approximate functions is ubiquitous in science, either due to empirical constraints or high computational cost of accessing the function. In high-energy physics, the precise computation of the scattering cross-section of a process requires the evaluation of computationally intensive integrals. A wide variety of methods in machine learning have been used to tackle this problem, but often the motivation of using one method over another is lacking. Comparing these methods is typically highly dependent on the problem at hand, so we specify to the case where we can evaluate the function a large number of times, after which quick and accurate evaluation can take place. We consider four interpolation and three machine learning techniques and compare their performance on three toy functions, the four-point scalar Passarino-Veltman $D_0$ function, and the two-loop self-energy master integral $M$. We find that in low dimensions ($d=3$), traditional interpolation techniques like the Radial Basis Function perform very well, but in higher dimensions ($d=5,6,9$) we find that multi-layer perceptrons (a.k.a neural networks) do not suffer as much from the curse of dimensionality and provide the fastest and most accurate predictions.

arXiv comment30 pages, 17 figures, v2:added a few references, v3: new title, added a few references

References in corpus (32)

• Adam: A Method for Stochastic Optimization
• Scikit-learn: Machine Learning in Python
• XGBoost: A Scalable Tree Boosting System
• SciPy 1.0--Fundamental Algorithms for Scientific Computing in Python
• TensorFlow: A system for large-scale machine learning
• Gaussian Error Linear Units (GELUs)
• Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning
• Parton distributions for the LHC Run II
• Variational Inference with Normalizing Flows
• Precision Electroweak Measurements on the Z Resonance
• Collier: a fortran-based Complex One-Loop LIbrary in Extended Regularizations
• Package-X: A Mathematica package for the analytic calculation of one-loop integrals
• Reduction schemes for one-loop tensor integrals
• Gaussian Processes for Big Data
• Reduction of one-loop tensor 5-point integrals
• GPflow: A Gaussian process library using TensorFlow
• Scalar one-loop 4-point integrals
• Generalized Recurrence Relations for Two-loop Propagator Integrals with Arbitrary Masses
• TSIL: a program for the calculation of two-loop self-energy integrals
• DART: Dropouts meet Multiple Additive Regression Trees
• i-flow: High-dimensional Integration and Sampling with Normalizing Flows
• NNLO QCD corrections to three-photon production at the LHC
• Single-jet inclusive rates with exact color at $\mathcal{O}({\alpha }_s^4)$
• RBF Interpolation with CSRBF of Large Data Sets
• GANplifying Event Samples
• A factorisation-aware Matrix element emulator
• Optimising simulations for diphoton production at hadron colliders using amplitude neural networks
• Targeting Multi-Loop Integrals with Neural Networks
• Using neural networks for efficient evaluation of high multiplicity scattering amplitudes
• Efficient Numerical Evaluation of Feynman Integral
• Machine-learning interpolation of population-synthesis simulations to interpret gravitational-wave observations: a case study
• $\text{Xsec}$: the cross-section evaluation code

Cited by in corpus (6)

• Calomplification -- The Power of Generative Calorimeter Models
• Deep Generative Models for Detector Signature Simulation: A Taxonomic Review
• Hypergraphs in LHC Phenomenology -- The Next Frontier of IRC-Safe Feature Extraction
• Predicting Feynman periods in ${\varphi }^4$-theory
• Modeling the Asymptotic Behavior of Higher-Order Aftershocks with Deep Learning
• Using Neural Networks to Accelerate TALYS-2.0 Nuclear Reaction Simulations