An iterative deep learning procedure for determining electron scattering cross-sections from transport coefficients
We propose improvements to the artificial neural network (ANN) method of determining electron scattering cross-sections from swarm data proposed by coauthors. A limitation inherent to this problem, known as the inverse swarm problem, is the non-unique nature of its solutions, particularly when there...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IOP Publishing
2024-01-01
|
| Series: | Machine Learning: Science and Technology |
| Subjects: | |
| Online Access: | https://doi.org/10.1088/2632-2153/ad2fed |
| _version_ | 1797262547045318656 |
|---|---|
| author | Dale L Muccignat Gregory G Boyle Nathan A Garland Peter W Stokes Ronald D White |
| author_facet | Dale L Muccignat Gregory G Boyle Nathan A Garland Peter W Stokes Ronald D White |
| author_sort | Dale L Muccignat |
| collection | DOAJ |
| description | We propose improvements to the artificial neural network (ANN) method of determining electron scattering cross-sections from swarm data proposed by coauthors. A limitation inherent to this problem, known as the inverse swarm problem, is the non-unique nature of its solutions, particularly when there exists multiple cross-sections that each describe similar scattering processes. Considering this, prior methods leveraged existing knowledge of a particular cross-section set to reduce the solution space of the problem. To reduce the need for prior knowledge, we propose the following modifications to the ANN method. First, we propose a multi-branch ANN (MBANN) that assigns an independent branch of hidden layers to each cross-section output. We show that in comparison with an equivalent conventional ANN, the MBANN architecture enables an efficient and physics informed feature map of each cross-section. Additionally, we show that the MBANN solution can be improved upon by successive networks that are each trained using perturbations of the previous regression. Crucially, the method requires much less input data and fewer restrictive assumptions, and only assumes knowledge of energy loss thresholds and the number of cross-sections present. |
| first_indexed | 2024-04-24T23:58:50Z |
| format | Article |
| id | doaj.art-efa870389a434aacb0ca0e4bc17e2336 |
| institution | Directory Open Access Journal |
| issn | 2632-2153 |
| language | English |
| last_indexed | 2024-04-24T23:58:50Z |
| publishDate | 2024-01-01 |
| publisher | IOP Publishing |
| record_format | Article |
| series | Machine Learning: Science and Technology |
| spelling | doaj.art-efa870389a434aacb0ca0e4bc17e23362024-03-14T09:20:32ZengIOP PublishingMachine Learning: Science and Technology2632-21532024-01-015101504710.1088/2632-2153/ad2fedAn iterative deep learning procedure for determining electron scattering cross-sections from transport coefficientsDale L Muccignat0https://orcid.org/0000-0002-6693-6851Gregory G Boyle1https://orcid.org/0000-0002-8581-4307Nathan A Garland2https://orcid.org/0000-0003-0343-0199Peter W Stokes3https://orcid.org/0000-0002-0956-5927Ronald D White4https://orcid.org/0000-0001-5353-7440College of Science & Engineering, James Cook University , Townsville, QLD 4814, AustraliaCollege of Science & Engineering, James Cook University , Townsville, QLD 4814, AustraliaCentre for Quantum Dynamics, Griffith University , Nathan, QLD 4111, Australia; School of Environment and Science, Griffith University , Nathan, QLD 4111, AustraliaCollege of Science & Engineering, James Cook University , Townsville, QLD 4814, Australia; Department of Medical Physics, Townsville University Hospital , Townsville, QLD 4814, AustraliaCollege of Science & Engineering, James Cook University , Townsville, QLD 4814, AustraliaWe propose improvements to the artificial neural network (ANN) method of determining electron scattering cross-sections from swarm data proposed by coauthors. A limitation inherent to this problem, known as the inverse swarm problem, is the non-unique nature of its solutions, particularly when there exists multiple cross-sections that each describe similar scattering processes. Considering this, prior methods leveraged existing knowledge of a particular cross-section set to reduce the solution space of the problem. To reduce the need for prior knowledge, we propose the following modifications to the ANN method. First, we propose a multi-branch ANN (MBANN) that assigns an independent branch of hidden layers to each cross-section output. We show that in comparison with an equivalent conventional ANN, the MBANN architecture enables an efficient and physics informed feature map of each cross-section. Additionally, we show that the MBANN solution can be improved upon by successive networks that are each trained using perturbations of the previous regression. Crucially, the method requires much less input data and fewer restrictive assumptions, and only assumes knowledge of energy loss thresholds and the number of cross-sections present.https://doi.org/10.1088/2632-2153/ad2fedswarm analysisinverse problemBoltzmann equationmachine learning |
| spellingShingle | Dale L Muccignat Gregory G Boyle Nathan A Garland Peter W Stokes Ronald D White An iterative deep learning procedure for determining electron scattering cross-sections from transport coefficients Machine Learning: Science and Technology swarm analysis inverse problem Boltzmann equation machine learning |
| title | An iterative deep learning procedure for determining electron scattering cross-sections from transport coefficients |
| title_full | An iterative deep learning procedure for determining electron scattering cross-sections from transport coefficients |
| title_fullStr | An iterative deep learning procedure for determining electron scattering cross-sections from transport coefficients |
| title_full_unstemmed | An iterative deep learning procedure for determining electron scattering cross-sections from transport coefficients |
| title_short | An iterative deep learning procedure for determining electron scattering cross-sections from transport coefficients |
| title_sort | iterative deep learning procedure for determining electron scattering cross sections from transport coefficients |
| topic | swarm analysis inverse problem Boltzmann equation machine learning |
| url | https://doi.org/10.1088/2632-2153/ad2fed |
| work_keys_str_mv | AT dalelmuccignat aniterativedeeplearningprocedurefordeterminingelectronscatteringcrosssectionsfromtransportcoefficients AT gregorygboyle aniterativedeeplearningprocedurefordeterminingelectronscatteringcrosssectionsfromtransportcoefficients AT nathanagarland aniterativedeeplearningprocedurefordeterminingelectronscatteringcrosssectionsfromtransportcoefficients AT peterwstokes aniterativedeeplearningprocedurefordeterminingelectronscatteringcrosssectionsfromtransportcoefficients AT ronalddwhite aniterativedeeplearningprocedurefordeterminingelectronscatteringcrosssectionsfromtransportcoefficients AT dalelmuccignat iterativedeeplearningprocedurefordeterminingelectronscatteringcrosssectionsfromtransportcoefficients AT gregorygboyle iterativedeeplearningprocedurefordeterminingelectronscatteringcrosssectionsfromtransportcoefficients AT nathanagarland iterativedeeplearningprocedurefordeterminingelectronscatteringcrosssectionsfromtransportcoefficients AT peterwstokes iterativedeeplearningprocedurefordeterminingelectronscatteringcrosssectionsfromtransportcoefficients AT ronalddwhite iterativedeeplearningprocedurefordeterminingelectronscatteringcrosssectionsfromtransportcoefficients |