본문

서브메뉴

Robust Representation Learning and Real-Time Serving of Deep Models for Health Time Series
Robust Representation Learning and Real-Time Serving of Deep Models for Health Time Series
Robust Representation Learning and Real-Time Serving of Deep Models for Health Time Series

Detailed Information

자료유형  
 학위논문 서양
최종처리일시  
20260209102934
ISBN  
9798263396688
DDC  
000
저자명  
Xu, Yanbo.
서명/저자  
Robust Representation Learning and Real-Time Serving of Deep Models for Health Time Series
발행사항  
[Sl] : Georgia Institute of Technology, 2023
발행사항  
Ann Arbor : ProQuest Dissertations & Theses, 2023
형태사항  
122 p
주기사항  
Source: Dissertations Abstracts International, Volume: 87-05, Section: B.
주기사항  
Advisor: Zhang, Chao;Tumanov, Alexey.
학위논문주기  
Thesis (Ph.D.)--Georgia Institute of Technology, 2023.
초록/해제  
요약Modern Electronic Health Record (EHR) systems provide large amount of data that enables machine learning (ML) researchers to develop ML methods to improve healthcare. However, development in a clinical setting presents unique challenges in ML model training and serving. For example, EHR data are usually captured from multiple sources over time in noisy environments such as in Intensive Care Units (ICUs). As a result, data are generated in the form of time series with multiple issues including heterogeneity, missingness, irregulrity, etc. Although ML methods such as deep neural networks have been successfully developed for many predictive health tasks, improvements are still in need for learning robust and efficient predictive models to harness such multi-modal, noisy, and massive time series data.In this thesis, we aim to tackle the following fundamental problems in developing ML models for health time series:• Multiple modalities in time series. Clinical time series are often generated on different devices at different frequencies. A typical ICU monitoring dataset can contain continuous signals like electrocardiogram (ECG), evenly charted tabular data like vital signs, and sparse discrete events like lab tests and medications. Simple binning methods on values can reduce rich information in dense data and mask important information in sparse data. To address this, we design an efficient ensembling algorithm for reweighting the models that are individualized for each data modality. Then for better capturing the underlying heterogeneity behind the multimodal data, we further design individualized embeddings per modality and fit self-attention Transformer on top of them for more robustly fusing the EHR time series.• Missing observations at random time steps. Data collection is often noisy in EHR systems. Missing data or mis-timestamped data happens due to random device disconnections, patient's body movement, human errors, etc. Models without consid erations on input missingness and noises can lead to overfit and biased predictions. We incorporate stochastic differential equations into spatial temporal modeling, enabling imputations on randomly missing fields in structural time series with support of uncertainty quantification. We further propose score-based diffusion models for generating missing data and denoising the observed discrete event sequences.• Large unlabelled data available across different sites. True labels are expensive to obtain in clinical applications. Although input signals can be easily collected in EHR systems, many labels of interest still require manual annotations and data reviews from clinical experts retrospectively. Thus large amount of unlabelled data, which can be collected across several different hospitals, become available to researchers whereas only a few are labelled. To address this challenge, we investigate self-supervised learning in deep models and learn robust representations from the large unlabelled data that can be later adapted and fine tuned for downstream tasks.• Timely serving in resource-limited systems. In clinical environments such as ICUs, care practitioners need to make appropriate decision in a timely manner. Thus far deep learning models have been mainly developed for increasing prediction accuracy in heathcare, but few of them consider whether or how they can be served in real time given a resource constrained deployment environment. To bridge the gap, we design cost-aware prediction pipelines that can cascade to differently sized models for balancing between prediction accuracy and serving cost.This thesis contributes to developing robust and efficient ML models for modeling health time series. We investigate two main machine learning techniques, robust representation learning and real-time serving, in deep learning models. My dissertation shows that our proposed approaches can make significant improvements in the four aforementioned problems.
일반주제명  
Mortality
일반주제명  
Neural networks
일반주제명  
Signal processing
일반주제명  
Electrocardiography
일반주제명  
Electrical engineering
일반주제명  
Medicine
기타저자  
Georgia Institute of Technology.
기본자료저록  
Dissertations Abstracts International. 87-05B.
전자적 위치 및 접속  
로그인 후 원문을 볼 수 있습니다.

MARC

 008260203s2023        us                              c    eng  d
■001000017366043
■00520260209102934
■006m          o    d                
■007cr#unu||||||||
■020    ▼a9798263396688
■035    ▼a(MiAaPQ)AAI32315775
■035    ▼a(MiAaPQ)GeorgiaTech75107
■040    ▼aMiAaPQ▼cMiAaPQ
■0820  ▼a000
■1001  ▼aXu,  Yanbo.
■24510▼aRobust  Representation  Learning  and  Real-Time  Serving  of  Deep  Models  for  Health  Time  Series
■260    ▼a[Sl]▼bGeorgia  Institute  of  Technology▼c2023
■260  1▼aAnn  Arbor▼bProQuest  Dissertations  &  Theses▼c2023
■300    ▼a122  p
■500    ▼aSource:  Dissertations  Abstracts  International,  Volume:  87-05,  Section:  B.
■500    ▼aAdvisor:  Zhang,  Chao;Tumanov,  Alexey.
■5021  ▼aThesis  (Ph.D.)--Georgia  Institute  of  Technology,  2023.
■520    ▼aModern  Electronic  Health  Record  (EHR)  systems  provide  large  amount  of  data  that  enables  machine  learning  (ML)  researchers  to  develop  ML  methods  to  improve  healthcare.  However,  development  in  a  clinical  setting  presents  unique  challenges  in  ML  model  training  and  serving.  For  example,  EHR  data  are  usually  captured  from  multiple  sources  over  time  in  noisy  environments  such  as  in  Intensive  Care  Units  (ICUs).  As  a  result,  data  are  generated  in  the  form  of  time  series  with  multiple  issues  including  heterogeneity,  missingness,  irregulrity,  etc.  Although  ML  methods  such  as  deep  neural  networks  have  been  successfully  developed  for  many  predictive  health  tasks,  improvements  are  still  in  need  for  learning  robust  and  efficient  predictive  models  to  harness  such  multi-modal,  noisy,  and  massive  time  series  data.In  this  thesis,  we  aim  to  tackle  the  following  fundamental  problems  in  developing  ML  models  for  health  time  series:•  Multiple  modalities  in  time  series.  Clinical  time  series  are  often  generated  on  different  devices  at  different  frequencies.  A  typical  ICU  monitoring  dataset  can  contain  continuous  signals  like  electrocardiogram  (ECG),  evenly  charted  tabular  data  like  vital  signs,  and  sparse  discrete  events  like  lab  tests  and  medications.  Simple  binning  methods  on  values  can  reduce  rich  information  in  dense  data  and  mask  important  information  in  sparse  data.  To  address  this,  we  design  an  efficient  ensembling  algorithm  for  reweighting  the  models  that  are  individualized  for  each  data  modality.  Then  for  better  capturing  the  underlying  heterogeneity  behind  the  multimodal  data,  we  further  design  individualized  embeddings  per  modality  and  fit  self-attention  Transformer  on  top  of  them  for  more  robustly  fusing  the  EHR  time  series.•  Missing  observations  at  random  time  steps.  Data  collection  is  often  noisy  in  EHR  systems.  Missing  data  or  mis-timestamped  data  happens  due  to  random  device  disconnections,  patient's  body  movement,  human  errors,  etc.  Models  without  consid  erations  on  input  missingness  and  noises  can  lead  to  overfit  and  biased  predictions.  We  incorporate  stochastic  differential  equations  into  spatial  temporal  modeling,  enabling  imputations  on  randomly  missing  fields  in  structural  time  series  with  support  of  uncertainty  quantification.  We  further  propose  score-based  diffusion  models  for  generating  missing  data  and  denoising  the  observed  discrete  event  sequences.•  Large  unlabelled  data  available  across  different  sites.  True  labels  are  expensive  to  obtain  in  clinical  applications.  Although  input  signals  can  be  easily  collected  in  EHR  systems,  many  labels  of  interest  still  require  manual  annotations  and  data  reviews  from  clinical  experts  retrospectively.  Thus  large  amount  of  unlabelled  data,  which  can  be  collected  across  several  different  hospitals,  become  available  to  researchers  whereas  only  a  few  are  labelled.  To  address  this  challenge,  we  investigate  self-supervised  learning  in  deep  models  and  learn  robust  representations  from  the  large  unlabelled  data  that  can  be  later  adapted  and  fine  tuned  for  downstream  tasks.•  Timely  serving  in  resource-limited  systems.  In  clinical  environments  such  as  ICUs,  care  practitioners  need  to  make  appropriate  decision  in  a  timely  manner.  Thus  far  deep  learning  models  have  been  mainly  developed  for  increasing  prediction  accuracy  in  heathcare,  but  few  of  them  consider  whether  or  how  they  can  be  served  in  real  time  given  a  resource  constrained  deployment  environment.  To  bridge  the  gap,  we  design  cost-aware  prediction  pipelines  that  can  cascade  to  differently  sized  models  for  balancing  between  prediction  accuracy  and  serving  cost.This  thesis  contributes  to  developing  robust  and  efficient  ML  models  for  modeling  health  time  series.  We  investigate  two  main  machine  learning  techniques,  robust  representation  learning  and  real-time  serving,  in  deep  learning  models.  My  dissertation  shows  that  our  proposed  approaches  can  make  significant  improvements  in  the  four  aforementioned  problems.
■590    ▼aSchool  code:  0078.
■650  4▼aMortality
■650  4▼aNeural  networks
■650  4▼aSignal  processing
■650  4▼aElectrocardiography
■650  4▼aElectrical  engineering
■650  4▼aMedicine
■690    ▼a0800
■690    ▼a0544
■690    ▼a0564
■71020▼aGeorgia  Institute  of  Technology.
■7730  ▼tDissertations  Abstracts  International▼g87-05B.
■790    ▼a0078
■791    ▼aPh.D.
■792    ▼a2023
■793    ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17366043▼nKERIS▼z이  자료의  원문은  한국교육학술정보원에서  제공합니다.

Preview

Export

ChatGPT Discussion

AI Recommended Related Books


    New Books MORE
    Statistics for the past 3 years. Go to brief

    Подробнее информация.

    • Бронирование
    • не существует
    • моя папка
    • Первый запрос зрения
    • Non-Book Loan Application
    • Nighttime Book Loan Application
    материал
    Reg No. Количество платежных Местоположение статус Ленд информации
    TF14975 전자도서 대출가능 My Folder 부재도서신고 비도서대출신청 야간 도서대출신청

    * Бронирование доступны в заимствований книги. Чтобы сделать предварительный заказ, пожалуйста, нажмите кнопку бронирование

    Books borrowed together with this book

    Related Popular Books

    Available after logging in.