Variational Pseudo Marginal Methods for Jet Reconstruction in Particle Physics

Hanming Yang,Antonio Khalil Moretti,Sebastian Macaluso,Philippe Chlenski,Christian A. Naesseth,Itsik Pe'er

2024-06-05

Abstract:Reconstructing jets, which provide vital insights into the properties and histories of subatomic particles produced in high-energy collisions, is a main problem in data analyses in collider physics. This intricate task deals with estimating the latent structure of a jet (binary tree) and involves parameters such as particle energy, momentum, and types. While Bayesian methods offer a natural approach for handling uncertainty and leveraging prior knowledge, they face significant challenges due to the super-exponential growth of potential jet topologies as the number of observed particles increases. To address this, we introduce a Combinatorial Sequential Monte Carlo approach for inferring jet latent structures. As a second contribution, we leverage the resulting estimator to develop a variational inference algorithm for parameter learning. Building on this, we introduce a variational family using a pseudo-marginal framework for a fully Bayesian treatment of all variables, unifying the generative model with the inference process. We illustrate our method's effectiveness through experiments using data generated with a collider physics generative model, highlighting superior speed and accuracy across a range of tasks.

Machine Learning,Computation

What problem does this paper attempt to address?

This paper focuses on the problem of jet reconstruction in high energy particle physics, which is a key challenge in data analysis. It involves estimating the potential structure (binary tree) of jets and parameters such as energy, momentum, and type of particles. As the number of observed particles increases, the potential jet topologies grow exponentially, making it difficult for traditional Bayesian methods to handle uncertainty. To address this issue, the paper proposes a combined sequential Monte Carlo method to infer the potential structure of jets, and develops a variational inference algorithm for parameter learning using this method. By constructing a variational family of pseudo marginal frameworks, complete Bayesian processing of all variables is achieved, unifying the generative model with the inference process. The paper demonstrates the effectiveness and accuracy of the proposed method on various tasks by conducting experiments using data from particle physics generative models. Moreover, the method is also applicable to precise or high-quality approximations for small to medium-sized datasets, overcoming the limitations of existing methods in jet reconstruction.