본문

서브메뉴

Towards Controllability, Efficiency, and Trustworthiness of Text Generation Systems
Towards Controllability, Efficiency, and Trustworthiness of Text Generation Systems
Towards Controllability, Efficiency, and Trustworthiness of Text Generation Systems

상세정보

자료유형  
 학위논문 서양
최종처리일시  
20260202105418
ISBN  
9798291566022
DDC  
004
저자명  
Cao, Shuyang.
서명/저자  
Towards Controllability, Efficiency, and Trustworthiness of Text Generation Systems
발행사항  
[Sl] : University of Michigan, 2025
발행사항  
Ann Arbor : ProQuest Dissertations & Theses, 2025
형태사항  
204 p
주기사항  
Source: Dissertations Abstracts International, Volume: 87-03, Section: B.
주기사항  
Advisor: Wang, Lu.
학위논문주기  
Thesis (Ph.D.)--University of Michigan, 2025.
초록/해제  
요약As reliance on automated text generation continues to grow, understanding and overcoming the limitations of current systems are critical. This thesis addresses the challenges of controllability, efficiency, and trustworthiness in text generation systems, which can enhance their practical applicability across diverse domains. The first part of this work focuses on controllability, where text outputs must meet specific user needs or constraints. We introduce a type-controlled question generation system that employs exemplar templates and optionally generated fine-grained templates to achieve improved control over target attributes. The template-based control signals significantly improve the system's ability to produce questions of required types, compared to systems that directly apply constraints with type labels. Second, we investigate efficient training methods in the context of long-document summarization, where efficiency issues are most pronounced. For incorporating document structure information that commonly exists in long documents, we design learnable biases representing section-level relations. These biases add minimal overhead to model training, while improving models' structure understanding, leading to summaries of high quality. We further consider a divide-and-conquer framework to enable model training on long sequences under strict resource constraints. Summarizing divided chunks separately, the GPU memory requirement is significantly reduced. Additionally, our framework is equipped with memory mechanisms and global salient content to establish connections between chunks, achieving performance comparable to resource-demanding systems. To evaluate models with increasing context length enabled by improved efficiency, we develop an adaptive evaluation benchmark that can flexibly adjust input context length and task difficulty. Our evaluation benchmark curates diverse synthetic tasks targeting varying levels of model capabilities. By requiring complex model capabilities, our benchmark presents sufficient challenges to current and potential future models. Notably, these synthetic tasks share the same contexts within each domain, thus mitigating the confounding effects of input context variations and allowing for more controlled comparisons of model capabilities to understand model behaviors. Lastly, we address the issue of factual inaccuracies in generated summaries. Our proposed contrastive learning framework trains models to differentiate correct outputs from erroneous ones that are automatically collected with various carefully designed strategies. Models trained with our framework consistently deliver more trustworthy outputs. Besides trustworthy models, we also explore verifiable generation with fine-grained citations that enhances user confidence in the reliability of generated content. The generated citations to source documents are attached to specific spans of the outputs, allowing for quick and precise verification of generated content. We collect a new dataset to felicitate the study of this generation paradigm. By ensuring outputs are not only contextually relevant but also factually accurate and verifiable, real-world applications of automatic text generation systems can be expanded with excellent user trust.
일반주제명  
Computer science
일반주제명  
Computer engineering
키워드  
Automatic text generation
키워드  
Controllability
키워드  
Efficiency
키워드  
Trustworthiness
기타저자  
University of Michigan Computer Science & Engineering
기본자료저록  
Dissertations Abstracts International. 87-03B.
전자적 위치 및 접속  
로그인 후 원문을 볼 수 있습니다.

MARC

 008260126s2025        us                              c    eng  d
■001000017360282
■00520260202105418
■006m          o    d                
■007cr#unu||||||||
■020    ▼a9798291566022
■035    ▼a(MiAaPQ)AAI32271791
■035    ▼a(MiAaPQ)umichrackham006243
■040    ▼aMiAaPQ▼cMiAaPQ
■0820  ▼a004
■1001  ▼aCao,  Shuyang.
■24510▼aTowards  Controllability,  Efficiency,  and  Trustworthiness  of  Text  Generation  Systems
■260    ▼a[Sl]▼bUniversity  of  Michigan▼c2025
■260  1▼aAnn  Arbor▼bProQuest  Dissertations  &  Theses▼c2025
■300    ▼a204  p
■500    ▼aSource:  Dissertations  Abstracts  International,  Volume:  87-03,  Section:  B.
■500    ▼aAdvisor:  Wang,  Lu.
■5021  ▼aThesis  (Ph.D.)--University  of  Michigan,  2025.
■520    ▼aAs  reliance  on  automated  text  generation  continues  to  grow,  understanding  and  overcoming  the  limitations  of  current  systems  are  critical.  This  thesis  addresses  the  challenges  of  controllability,  efficiency,  and  trustworthiness  in  text  generation  systems,  which  can  enhance  their  practical  applicability  across  diverse  domains.  The  first  part  of  this  work  focuses  on  controllability,  where  text  outputs  must  meet  specific  user  needs  or  constraints.  We  introduce  a  type-controlled  question  generation  system  that  employs  exemplar  templates  and  optionally  generated  fine-grained  templates  to  achieve  improved  control  over  target  attributes.  The  template-based  control  signals  significantly  improve  the  system's  ability  to  produce  questions  of  required  types,  compared  to  systems  that  directly  apply  constraints  with  type  labels.  Second,  we  investigate  efficient  training  methods  in  the  context  of  long-document  summarization,  where  efficiency  issues  are  most  pronounced.  For  incorporating  document  structure  information  that  commonly  exists  in  long  documents,  we  design  learnable  biases  representing  section-level  relations.  These  biases  add  minimal  overhead  to  model  training,  while  improving  models'  structure  understanding,  leading  to  summaries  of  high  quality.  We  further  consider  a  divide-and-conquer  framework  to  enable  model  training  on  long  sequences  under  strict  resource  constraints.  Summarizing  divided  chunks  separately,  the  GPU  memory  requirement  is  significantly  reduced.  Additionally,  our  framework  is  equipped  with  memory  mechanisms  and  global  salient  content  to  establish  connections  between  chunks,  achieving  performance  comparable  to  resource-demanding  systems.  To  evaluate  models  with  increasing  context  length  enabled  by  improved  efficiency,  we  develop  an  adaptive  evaluation  benchmark  that  can  flexibly  adjust  input  context  length  and  task  difficulty.  Our  evaluation  benchmark  curates  diverse  synthetic  tasks  targeting  varying  levels  of  model  capabilities.  By  requiring  complex  model  capabilities,  our  benchmark  presents  sufficient  challenges  to  current  and  potential  future  models.  Notably,  these  synthetic  tasks  share  the  same  contexts  within  each  domain,  thus  mitigating  the  confounding  effects  of  input  context  variations  and  allowing  for  more  controlled  comparisons  of  model  capabilities  to  understand  model  behaviors.  Lastly,  we  address  the  issue  of  factual  inaccuracies  in  generated  summaries.  Our  proposed  contrastive  learning  framework  trains  models  to  differentiate  correct  outputs  from  erroneous  ones  that  are  automatically  collected  with  various  carefully  designed  strategies.  Models  trained  with  our  framework  consistently  deliver  more  trustworthy  outputs.  Besides  trustworthy  models,  we  also  explore  verifiable  generation  with  fine-grained  citations  that  enhances  user  confidence  in  the  reliability  of  generated  content.  The  generated  citations  to  source  documents  are  attached  to  specific  spans  of  the  outputs,  allowing  for  quick  and  precise  verification  of  generated  content.  We  collect  a  new  dataset  to  felicitate  the  study  of  this  generation  paradigm.  By  ensuring  outputs  are  not  only  contextually  relevant  but  also  factually  accurate  and  verifiable,  real-world  applications  of  automatic  text  generation  systems  can  be  expanded  with  excellent  user  trust.
■590    ▼aSchool  code:  0127.
■650  4▼aComputer  science
■650  4▼aComputer  engineering
■653    ▼aAutomatic  text  generation
■653    ▼aControllability
■653    ▼aEfficiency
■653    ▼aTrustworthiness
■690    ▼a0984
■690    ▼a0800
■690    ▼a0464
■71020▼aUniversity  of  Michigan▼bComputer  Science  &  Engineering.
■7730  ▼tDissertations  Abstracts  International▼g87-03B.
■790    ▼a0127
■791    ▼aPh.D.
■792    ▼a2025
■793    ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17360282▼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
    TF18736 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.