본문

서브메뉴

Integrating Machine Learning Techniques for Streamlined Predictive Modeling in Cosmological Applications
Integrating Machine Learning Techniques for Streamlined Predictive Modeling in Cosmologica...
Integrating Machine Learning Techniques for Streamlined Predictive Modeling in Cosmological Applications

상세정보

자료유형  
 학위논문 서양
최종처리일시  
20260202105534
ISBN  
9798263348496
DDC  
523
저자명  
Sethuram, Snigdaa S.
서명/저자  
Integrating Machine Learning Techniques for Streamlined Predictive Modeling in Cosmological Applications
발행사항  
[Sl] : Georgia Institute of Technology, 2025
발행사항  
Ann Arbor : ProQuest Dissertations & Theses, 2025
형태사항  
142 p
주기사항  
Source: Dissertations Abstracts International, Volume: 87-05, Section: B.
주기사항  
Advisor: Wise, John H.
학위논문주기  
Thesis (Ph.D.)--Georgia Institute of Technology, 2025.
초록/해제  
요약The field of astrophysics has developed alongside humanity for many centuries, and is unique in our inability to directly conduct "experiments" on astrophysical objects. Historically, astrophysics has largely been observational and theoretical in nature, but with the advent of computers and their increasing capabilities, the field of computational astrophysics has accelerated our understanding of the universe profoundly. Simulations provide a type of sandbox in which we can better understand the objects we observe and explore our universe in greater detail.As Moore's Law has played out in the development of computer architectures, computational astrophysics has, in parallel, evolved from simplistic analytical models to highly sophisticated simulations capturing a wide dynamic range of physical processes. The rapid development of deep learning models in recent years has the potential to transform this landscape, making simulations significantly more efficient and accessible. Traditional high-fidelity simulations are extremely useful and essential for capturing complex cosmic phenomena, but are often prohibitively expensive in terms of time and resources. By integrating deep learning techniques, we can now emulate intricate physical processes with remarkable speed and accuracy, exploring broader parameter spaces and delving into more detailed astrophysical models. This breakthrough not only enhances our ability to study the universe at unprecedented scales but also opens new avenues for investigating complex physics that were previously out of reach.In this dissertation, I leverage machine learning (ML) and data analysis techniques to optimize simulation runtimes and processing, with an emphasis on the early universe. The projects presented herein demonstrate how targeted computational strategies can streamline astrophysical simulations and provide valuable insights to inform observational efforts.1.1 Background and Motivation1.1.1 The Early Universe and the Epoch of ReionizationThe formation of the first stars and galaxies represents a crucial turning point in the evolution of cosmic structure, signaling the end of the cosmic "Dark Ages." Prior to this, around 400,000 years after the Big Bang, the universe had expanded and cooled sufficiently for its temperature to drop below 3,000 K. This cooling allowed free electrons and protons to combine into neutral hydrogen and helium, a process known as recombination. As a consequence, photons, no longer scattered by free electrons, began to travel freely through space, leading to the decoupling of radiation from matter. This moment left behind a relic imprint known as the Cosmic Microwave Background (CMB) and marked the onset of the Dark Ages, which lasted approximately 200 million years before the first luminous sources emerged. A general schematic of the evolution of the universe is shown in Figure 1.1.These primordial galaxies hosted a diverse population of astrophysical sources that played a transformative role in reshaping the early Universe. At their core were the first generations of stars: the metal-free massive, short-lived Population III (Pop III) stars, which were composed almost entirely of hydrogen and helium, as they formed before any significant metal enrichment had occurred. Their violent deaths - through supernovae and direct-collapse mechanisms - seeded the cosmos with the first heavy elements, paving the way for the emergence of second-generation metal-poor Population II (Pop II) stars, which contained trace amounts of metals and continued to shape early galactic evolution.
일반주제명  
Astrophysics
일반주제명  
Star & galaxy formation
일반주제명  
Deep learning
일반주제명  
Gravity
일반주제명  
Hydrogen
일반주제명  
Radiation
일반주제명  
Cosmology
일반주제명  
Middle Ages
일반주제명  
Cooling
일반주제명  
Space telescopes
일반주제명  
Black holes
일반주제명  
Dark matter
일반주제명  
Neural networks
일반주제명  
Stars & galaxies
일반주제명  
Supernovae
일반주제명  
Fluid mechanics
일반주제명  
Universe
일반주제명  
Astronomy
일반주제명  
Optics
일반주제명  
Theoretical physics
기타저자  
Georgia Institute of Technology.
기본자료저록  
Dissertations Abstracts International. 87-05B.
전자적 위치 및 접속  
로그인 후 원문을 볼 수 있습니다.

MARC

 008260126s2025        us                              c    eng  d
■001000017360484
■00520260202105534
■006m          o    d                
■007cr#unu||||||||
■020    ▼a9798263348496
■035    ▼a(MiAaPQ)AAI32310079
■035    ▼a(MiAaPQ)GeorgiaTech77936
■040    ▼aMiAaPQ▼cMiAaPQ
■0820  ▼a523
■1001  ▼aSethuram,  Snigdaa  S.
■24510▼aIntegrating  Machine  Learning  Techniques  for  Streamlined  Predictive  Modeling  in  Cosmological  Applications
■260    ▼a[Sl]▼bGeorgia  Institute  of  Technology▼c2025
■260  1▼aAnn  Arbor▼bProQuest  Dissertations  &  Theses▼c2025
■300    ▼a142  p
■500    ▼aSource:  Dissertations  Abstracts  International,  Volume:  87-05,  Section:  B.
■500    ▼aAdvisor:  Wise,  John  H.
■5021  ▼aThesis  (Ph.D.)--Georgia  Institute  of  Technology,  2025.
■520    ▼aThe  field  of  astrophysics  has  developed  alongside  humanity  for  many  centuries,  and  is  unique  in  our  inability  to  directly  conduct  "experiments"  on  astrophysical  objects.  Historically,  astrophysics  has  largely  been  observational  and  theoretical  in  nature,  but  with  the  advent  of  computers  and  their  increasing  capabilities,  the  field  of  computational  astrophysics  has  accelerated  our  understanding  of  the  universe  profoundly.  Simulations  provide  a  type  of  sandbox  in  which  we  can  better  understand  the  objects  we  observe  and  explore  our  universe  in  greater  detail.As  Moore's  Law  has  played  out  in  the  development  of  computer  architectures,  computational  astrophysics  has,  in  parallel,  evolved  from  simplistic  analytical  models  to  highly  sophisticated  simulations  capturing  a  wide  dynamic  range  of  physical  processes.  The  rapid  development  of  deep  learning  models  in  recent  years  has  the  potential  to  transform  this  landscape,  making  simulations  significantly  more  efficient  and  accessible.  Traditional  high-fidelity  simulations  are  extremely  useful  and  essential  for  capturing  complex  cosmic  phenomena,  but  are  often  prohibitively  expensive  in  terms  of  time  and  resources.  By  integrating  deep  learning  techniques,  we  can  now  emulate  intricate  physical  processes  with  remarkable  speed  and  accuracy,  exploring  broader  parameter  spaces  and  delving  into  more  detailed  astrophysical  models.  This  breakthrough  not  only  enhances  our  ability  to  study  the  universe  at  unprecedented  scales  but  also  opens  new  avenues  for  investigating  complex  physics  that  were  previously  out  of  reach.In  this  dissertation,  I  leverage  machine  learning  (ML)  and  data  analysis  techniques  to  optimize  simulation  runtimes  and  processing,  with  an  emphasis  on  the  early  universe.  The  projects  presented  herein  demonstrate  how  targeted  computational  strategies  can  streamline  astrophysical  simulations  and  provide  valuable  insights  to  inform  observational  efforts.1.1  Background  and  Motivation1.1.1  The  Early  Universe  and  the  Epoch  of  ReionizationThe  formation  of  the  first  stars  and  galaxies  represents  a  crucial  turning  point  in  the  evolution  of  cosmic  structure,  signaling  the  end  of  the  cosmic  "Dark  Ages."  Prior  to  this,  around  400,000  years  after  the  Big  Bang,  the  universe  had  expanded  and  cooled  sufficiently  for  its  temperature  to  drop  below  3,000  K.  This  cooling  allowed  free  electrons  and  protons  to  combine  into  neutral  hydrogen  and  helium,  a  process  known  as  recombination.  As  a  consequence,  photons,  no  longer  scattered  by  free  electrons,  began  to  travel  freely  through  space,  leading  to  the  decoupling  of  radiation  from  matter.  This  moment  left  behind  a  relic  imprint  known  as  the  Cosmic  Microwave  Background  (CMB)  and  marked  the  onset  of  the  Dark  Ages,  which  lasted  approximately  200  million  years  before  the  first  luminous  sources  emerged.  A  general  schematic  of  the  evolution  of  the  universe  is  shown  in  Figure  1.1.These  primordial  galaxies  hosted  a  diverse  population  of  astrophysical  sources  that  played  a  transformative  role  in  reshaping  the  early  Universe.  At  their  core  were  the  first  generations  of  stars:  the  metal-free  massive,  short-lived  Population  III  (Pop  III)  stars,  which  were  composed  almost  entirely  of  hydrogen  and  helium,  as  they  formed  before  any  significant  metal  enrichment  had  occurred.  Their  violent  deaths  -  through  supernovae  and  direct-collapse  mechanisms  -  seeded  the  cosmos  with  the  first  heavy  elements,  paving  the  way  for  the  emergence  of  second-generation  metal-poor  Population  II  (Pop  II)  stars,  which  contained  trace  amounts  of  metals  and  continued  to  shape  early  galactic  evolution.
■590    ▼aSchool  code:  0078.
■650  4▼aAstrophysics
■650  4▼aStar  &  galaxy  formation
■650  4▼aDeep  learning
■650  4▼aGravity
■650  4▼aHydrogen
■650  4▼aRadiation
■650  4▼aCosmology
■650  4▼aMiddle  Ages
■650  4▼aCooling
■650  4▼aSpace  telescopes
■650  4▼aBlack  holes
■650  4▼aDark  matter
■650  4▼aNeural  networks
■650  4▼aStars  &  galaxies
■650  4▼aSupernovae
■650  4▼aFluid  mechanics
■650  4▼aUniverse
■650  4▼aAstronomy
■650  4▼aOptics
■650  4▼aTheoretical  physics
■690    ▼a0800
■690    ▼a0204
■690    ▼a0596
■690    ▼a0606
■690    ▼a0752
■690    ▼a0753
■71020▼aGeorgia  Institute  of  Technology.
■7730  ▼tDissertations  Abstracts  International▼g87-05B.
■790    ▼a0078
■791    ▼aPh.D.
■792    ▼a2025
■793    ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17360484▼nKERIS▼z이  자료의  원문은  한국교육학술정보원에서  제공합니다.

미리보기

내보내기

chatGPT토론

Ai 추천 관련 도서


    신착도서 더보기
    최근 3년간 통계입니다.

    소장정보

    • 예약
    • 소재불명신고
    • 나의폴더
    • 우선정리요청
    • 비도서대출신청
    • 야간 도서대출신청
    소장자료
    등록번호 청구기호 소장처 대출가능여부 대출정보
    TF17450 전자도서 대출가능 마이폴더 부재도서신고 비도서대출신청 야간 도서대출신청

    * 대출중인 자료에 한하여 예약이 가능합니다. 예약을 원하시면 예약버튼을 클릭하십시오.

    해당 도서를 다른 이용자가 함께 대출한 도서

    관련 인기도서

    로그인 후 이용 가능합니다.