서브메뉴
검색
Efficient and Robust Web Scale Language Model Based Retrieval, Generation, and Understanding
Efficient and Robust Web Scale Language Model Based Retrieval, Generation, and Understanding
상세정보
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20260209102846
- ISBN
- 9798291562574
- DDC
- 004
- 저자명
- Campos, Daniel.
- 서명/저자
- Efficient and Robust Web Scale Language Model Based Retrieval, Generation, and Understanding
- 발행사항
- [Sl] : University of Illinois at Urbana-Champaign, 2023
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2023
- 형태사항
- 203 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 87-03, Section: B.
- 주기사항
- Advisor: Zhai, Cheng Xiang.
- 학위논문주기
- Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 2023.
- 초록/해제
- 요약Large language models effectively generate contextualized word representations across languages, domains, and tasks. Drive by these abilities, these models have become a building staple for many researchers and engineers who use text as their medium of representation, much like concrete is a staple in the construction world. Via the broad study and implementation, problems with large models have come to light: they can be expensive, brittle to noise, and produce unwanted outputs. Their large size and computational overhead make them difficult and costly to deploy and use for inference. Minor variations in text inputs, such as typos or misspellings, can cause significant losses in model accuracy. Seeking to improve how these models can be used for real world usage and deployments, this thesis focuses on approaches for improving performance by compressing, hardening, and optimizing models' performance based on deployment needs. To explore the challenges with largescale deployments concerning robustness and inference efficiency, we explore four commonly used language workloads: textual understanding and classification, passage retrieval, and text generation. We chose these broad but connected tasks to ensure that our compression approaches broadly apply to natural language processing. First, we propose a general framework for improving model inference on broad language understanding workloads by studying how unstructured pruning, structured pruning, and quantization can be leveraged to compress models and improve inference speeds. Second, we examine how models can be deployed for usage in web-scale generation and understanding workloads. Leveraging multi-task modeling, asymmetrical pruning, knowledge distillation, and quantization allows for cost-efficient scaling to web-scale workloads. Third, we explore methods of tuning and optimizing dense retrieval methods post-training to ensure they perform well on real-world data. Our experiments yield simple and effective ways of increasing model robustness and decreasing inference costs without any need for retraining or index re-generation. Finally, we discuss future work, focusing on sequential compression approaches to sequence LLMs to allow generative workloads to reach web-scale deployments.
- 일반주제명
- Computer science
- 일반주제명
- Computer engineering
- 기타저자
- University of Illinois at Urbana-Champaign Computer Science
- 기본자료저록
- Dissertations Abstracts International. 87-03B.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008260203s2023 us c eng d■001000017365880
■00520260209102846
■006m o d
■007cr#unu||||||||
■020 ▼a9798291562574
■035 ▼a(MiAaPQ)AAI32271292
■035 ▼a(MiAaPQ)httphdlhandlenet2142121494
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a004
■1001 ▼aCampos, Daniel.
■24510▼aEfficient and Robust Web Scale Language Model Based Retrieval, Generation, and Understanding
■260 ▼a[Sl]▼bUniversity of Illinois at Urbana-Champaign▼c2023
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2023
■300 ▼a203 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 87-03, Section: B.
■500 ▼aAdvisor: Zhai, Cheng Xiang.
■5021 ▼aThesis (Ph.D.)--University of Illinois at Urbana-Champaign, 2023.
■520 ▼aLarge language models effectively generate contextualized word representations across languages, domains, and tasks. Drive by these abilities, these models have become a building staple for many researchers and engineers who use text as their medium of representation, much like concrete is a staple in the construction world. Via the broad study and implementation, problems with large models have come to light: they can be expensive, brittle to noise, and produce unwanted outputs. Their large size and computational overhead make them difficult and costly to deploy and use for inference. Minor variations in text inputs, such as typos or misspellings, can cause significant losses in model accuracy. Seeking to improve how these models can be used for real world usage and deployments, this thesis focuses on approaches for improving performance by compressing, hardening, and optimizing models' performance based on deployment needs. To explore the challenges with largescale deployments concerning robustness and inference efficiency, we explore four commonly used language workloads: textual understanding and classification, passage retrieval, and text generation. We chose these broad but connected tasks to ensure that our compression approaches broadly apply to natural language processing. First, we propose a general framework for improving model inference on broad language understanding workloads by studying how unstructured pruning, structured pruning, and quantization can be leveraged to compress models and improve inference speeds. Second, we examine how models can be deployed for usage in web-scale generation and understanding workloads. Leveraging multi-task modeling, asymmetrical pruning, knowledge distillation, and quantization allows for cost-efficient scaling to web-scale workloads. Third, we explore methods of tuning and optimizing dense retrieval methods post-training to ensure they perform well on real-world data. Our experiments yield simple and effective ways of increasing model robustness and decreasing inference costs without any need for retraining or index re-generation. Finally, we discuss future work, focusing on sequential compression approaches to sequence LLMs to allow generative workloads to reach web-scale deployments.
■590 ▼aSchool code: 0090.
■650 4▼aComputer science
■650 4▼aComputer engineering
■653 ▼aEfficient inference
■653 ▼aLarge language models
■653 ▼aSemantic retrieval
■653 ▼aWeb-scale inference
■690 ▼a0984
■690 ▼a0464
■690 ▼a0800
■71020▼aUniversity of Illinois at Urbana-Champaign▼bComputer Science.
■7730 ▼tDissertations Abstracts International▼g87-03B.
■790 ▼a0090
■791 ▼aPh.D.
■792 ▼a2023
■793 ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17365880▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.


