BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//linuxsoftware.nz//NONSGML Joyous v1.4//EN
BEGIN:VEVENT
SUMMARY:Learning Reduced Nonlinear Plasma Models from Data
DTSTART:20241204T120000Z
DTEND:20241204T140000Z
DTSTAMP:20260606T234610Z
UID:a1072b12-bf85-47bf-85c7-c8beba343836
SEQUENCE:1
CREATED:20241203T094736Z
DESCRIPTION: This thesis explores the use of Machine Learning (ML) techniq
 ues\, with a particular emphasis on Sparse Regression (SR)\, to obtain int
 erpretable reduced nonlinear plasma models from data of first-principle Pa
 rticle-In-Cell (PIC) simulations.In order to become familiar with the SR t
 echnique\, its advantages and limitations\, SR is first applied to the dat
 a of the electron two-stream instability. Particular emphasis is given to 
 exploring the importance of using an integral formulation of SR to deal wi
 th intrinsically noisy data from finite particle statistics.SR is then app
 lied to the study of nonlinear ion-acoustic dynamics on plasmas. PIC simul
 ations are used to identify the rich dynamics of nonlinear perturbations i
 n plasmas. SR is used to recover the ion momentum equation from data of th
 ese simulations under different conditions. Motivated by the results on io
 n-acoustic waves\, SR technique is then applied to recover the Kortweg de-
 Vries (KdV) equation\, both on data generated directly from the analytical
  solutions and from fully-kinetic PIC simulations. It is shown that in wea
 k non-linear regimes\, as appropriate for the KdV equation\, even very low
  noise levels put strong limitations for recovering high order derivatives
 . These results have important implications for future applications of SR 
 to discover reduced nonlinear models from data. These implications and fut
 ure research directions are discussed.
LAST-MODIFIED:20241203T094736Z
LOCATION:Sala de Formação Avançada do DF (Sala 2-8.11 - 2º Piso do Pav
 ilhão de Física)
URL:http://df.vps.tecnico.ulisboa.pt/en/events/learning-reduced-nonlinear-
 plasma-models-from-data/
X-ALT-DESC;FMTTYPE=text/html:<p data-block-key="zr3cx"> This thesis explor
 es the use of Machine Learning (ML) techniques\, with a particular emphasi
 s on Sparse Regression (SR)\, to obtain interpretable reduced nonlinear pl
 asma models from data of first-principle Particle-In-Cell (PIC) simulation
 s.</p><p data-block-key="817ia">In order to become familiar with the SR te
 chnique\, its advantages and limitations\, SR is first applied to the data
  of the electron two-stream instability. Particular emphasis is given to e
 xploring the importance of using an integral formulation of SR to deal wit
 h intrinsically noisy data from finite particle statistics.<br/><br/></p><
 p data-block-key="32hkm">SR is then applied to the study of nonlinear ion-
 acoustic dynamics on plasmas. PIC simulations are used to identify the ric
 h dynamics of nonlinear perturbations in plasmas. SR is used to recover th
 e ion momentum equation from data of these simulations under different con
 ditions.<br/><br/> </p><p data-block-key="dusf4">Motivated by the results 
 on ion-acoustic waves\, SR technique is then applied to recover the Kortwe
 g de-Vries (KdV) equation\, both on data generated directly from the analy
 tical solutions and from fully-kinetic PIC simulations. It is shown that i
 n weak non-linear regimes\, as appropriate for the KdV equation\, even ver
 y low noise levels put strong limitations for recovering high order deriva
 tives. These results have important implications for future applications o
 f SR to discover reduced nonlinear models from data. These implications an
 d future research directions are discussed.</p>
END:VEVENT
END:VCALENDAR