Full Text Available
Note: Clicking the button above will open the full text document at the original institutional repository in a new window.
Machine learning methods for large deviations
Thesis (MSc)--Stellenbosch University, 2025.
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | English |
| Published: |
Stellenbosch : Stellenbosch University
2025
|
| Subjects: | |
| Tags: |
No Tags, Be the first to tag this record!
|
| _version_ | 1867613920524500992 |
|---|---|
| access_status_str | Open Access |
| author | Cloete, Daniël Willem Hendrik |
| author2 | Touchette, Hugo |
| author_browse | Cloete, Daniël Willem Hendrik Touchette, Hugo |
| author_facet | Touchette, Hugo Cloete, Daniël Willem Hendrik |
| author_sort | Cloete, Daniël Willem Hendrik |
| collection | Thesis |
| dc_rights_str_mv | Stellenbosch University |
| description | Thesis (MSc)--Stellenbosch University, 2025. |
| format | Thesis |
| id | oai:scholar.sun.ac.za:10019.1/134573 |
| institution | Stellenbosch University (South Africa) |
| language | English |
| last_indexed | 2026-06-10T12:43:48.768Z |
| license_str | Other — see source repository |
| provenance_str_mv | Harvested via OAI-PMH from SUNScholar — Stellenbosch University Repository |
| publishDate | 2025 |
| publishDateRange | 2025 |
| publishDateSort | 2025 |
| publisher | Stellenbosch : Stellenbosch University |
| publisherStr | Stellenbosch : Stellenbosch University |
| record_format | dspace |
| source_str | SUNScholar — Stellenbosch University Repository |
| spelling | oai:scholar.sun.ac.za:10019.1/134573 Machine learning methods for large deviations Cloete, Daniël Willem Hendrik Touchette, Hugo Stellenbosch University. Faculty of Science. Dept. of Applied Mathematics. Large deviations Random variables Stochastic processes -- Mathematical models Markov processes Fluctuations (Physics) Machine learning Reinforcement learning Thesis (MSc)--Stellenbosch University, 2025. Cloete, D. W. H. 2025. Machine Learning Methods for Large Deviations. Unpublished masters thesis. Stellenbosch: Stellenbosch University [online]. Available: https://scholar.sun.ac.za/items/742eae32-afb3-4f4e-9dc1-f9f0c2540c74 ENGLISH ABSTRACT: Rare fluctuations of random variables dependent on the trajectory of a noisy dynamical system are inherently difficult to simulate due to their scarcity. Gaining an understanding of the probabilities of such events, as well as the associated system behaviour, is vital to uncovering the mechanisms that drive rare events. We study the application of machine learning methods to the simulation of Markov chains, jump processes, and diffusions in which time-additive observables exhibit large deviations from their typical values. A variational formulation of the optimal sampling process forms the theoretical basis of these methods, from which an optimization objective is defined. To solve this optimization problem, we investigate two methods, namely, a trajectory-based global stochastic optimization algorithm and an online reinforcement learning approach. We provide details of the implementation and use of these methods and present illustrative results. The stochastic optimization method proves effective across all cases considered. In contrast, our preliminary investigation of the reinforcement learning approach, while showing promising performance for Markov chains, remains incomplete and highlights several challenges for future research. AFRIKAANSE OPSOMMING: Geen opsomming beskikbaar. Masters 2025-12-15T06:10:50Z 2025-12-15T06:10:50Z 2025-12 Thesis https://scholar.sun.ac.za/handle/10019.1/134573 en Stellenbosch University v, 116 pages : illustrations application/pdf Stellenbosch : Stellenbosch University |
| spellingShingle | Large deviations Random variables Stochastic processes -- Mathematical models Markov processes Fluctuations (Physics) Machine learning Reinforcement learning Cloete, Daniël Willem Hendrik Machine learning methods for large deviations |
| title | Machine learning methods for large deviations |
| title_full | Machine learning methods for large deviations |
| title_fullStr | Machine learning methods for large deviations |
| title_full_unstemmed | Machine learning methods for large deviations |
| title_short | Machine learning methods for large deviations |
| title_sort | machine learning methods for large deviations |
| topic | Large deviations Random variables Stochastic processes -- Mathematical models Markov processes Fluctuations (Physics) Machine learning Reinforcement learning |
| url | https://scholar.sun.ac.za/handle/10019.1/134573 |
| work_keys_str_mv | AT cloetedanielwillemhendrik machinelearningmethodsforlargedeviations |