Abstract
dc:description.abstractThe Standard Model (SM) stands as one of the most successful theories ever conceived, representing a triumph of human intellect and collaboration. The SM provides a very good description of visible matter, with the proton being a central part of this understanding. The structure of the proton, in terms of its elementary constituents, can be parametrised in terms of parton distribution functions (PDFs). PDFs, essential ingredients in theoretical predictions, cannot be calculated from first principles and must be extracted from fits to experimental data. Despite the immense success of the SM, we know that it cannot be the most complete theory of nature as it leaves important questions unanswered. In this context, the high energy physics community looks for physics beyond the SM (BSM). One way to conduct these searches is by studying subtle deviations of our theoretical predictions from the observed experimental values. In this context, as it reaches unprecedented levels of precision, the Large Hadron Collider (LHC) is carrying out one of the most rigorous tests of the SM. However, despite the large amounts of data collected so far, no direct evidence of BSM physics has been found. PDFs are crucial inputs in LHC physics. However, they are typically fitted under the assumption of the SM, which can introduce potential inconsistencies when they are used to generate predictions in BSM searches. In this way, a simultaneous determination of PDFs and BSM parameters is crucial to avoid biases in the interpretation of LHC data and to maximise the potential of the LHC to discover new physics. Additionally, BSM physics can be absorbed by the PDFs, rendering the interpretation of the data very challenging. The interplay between PDFs and BMS physics is an issue that is often overlooked in the literature and, in this thesis, we aim to explore it further. In Chapter 1, we begin by providing a review of the SM and, in particular, quantum chromodynamics and PDFs. Additionally, we introduce aspects of BSM physics and effective field theories (EFTs). We also formulate the problem of the PDF-BSM interplay and discuss the potential implications of this interplay in the context of LHC physics. After that, in Chapter 2, we discuss aspects of modern machine learning (ML) techniques that are useful for the analyses carried out in this thesis and we introduce SIMUnet: an open-source deep learning methodology to perform simultaneous fits of PDFs and BSM parameters, and to assess the potential absorption of new physics effects in PDF fits. In Chapter 3 we apply the SIMUnet methodology to perform fits of PDFs and EFT coefficients in the top sector and in global datasets. We compare the results of these fits with those obtained from standard SM PDF-only and EFT-only fits, and we discuss the implications of these results in terms of a simultaneous determination. Then, in Chapter 4, we apply the SIMUnet methodology to address the possible contamination of BSM physics in PDF fits. We show how this absorption can lead to apparent but spurious tensions between theoretical predictions and experimental measurements, and how to disentangle the effects of this contamination in the observables. Afterwards, in Chapter 5, we discuss whether BSM physics, formulated in terms of EFTs, can modify the renormalisation group evolution of the PDFs in the DGLAP equations, which describes how PDFs evolve with the energy scale. In Chapter 6, we shift our focus from PDFs to ways in which novel ML techniques can be used for precision calculations in LHC physics in general. To pave the way for the use of interpretable ML methodologies in collider physics, we explore how symbolic regression can be used for precision calculations in collider observables. In Chapter 7 we summarise the main results of this thesis and discuss potential future directions. Quotes are used to introduce each chapter. If the original quote is not in English, the original and the translated version are provided. If an official or well-known translation is available, it is used. Otherwise, a translation by the author, indicated by ‘trad. M. M. A.’, is provided.
Degree
thesis:*- Name dc:type.qualificationname
- Doctor of Philosophy (PhD)
- Level dc:type.qualificationlevel
- Doctoral
- Grantor dc:publisher.institution
- University of Cambridge
- Year dc:date.issued
- 2024
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Morales Alvarado, Manuel
- Advisor dc:contributor.advisor
-
- Ubiali, Maria
Subjects
dc:subject × 6Rights
dc:rightsIdentifiers
dc:identifier.*- DOI dc:identifier.doi
- https://doi.org/10.17863/CAM.116056
- OAI identifier oai:identifier
- oai:www.repository.cam.ac.uk:1810/380433