서브메뉴
검색
Customizing Large Language Models: From Synthetic Data Generation to Efficient Deployment
Customizing Large Language Models: From Synthetic Data Generation to Efficient Deployment
Detailed Information
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20260311091517.5
- ISBN
- 9798270232726
- DDC
- 005
- 저자명
- Wen, Yeming
- 서명/저자
- Customizing Large Language Models: From Synthetic Data Generation to Efficient Deployment / Yeming Wen
- 발행사항
- [Sl] : The University of Texas at Austin, 2025
- 형태사항
- 1 electronic resource (135 pages)
- 주기사항
- Source: Dissertations Abstracts International, Volume: 87-06, Section: B.
- 주기사항
- Advisors: Chaudhuri, Swarat Committee members: Durrett, Greg; Akella, Aditya; Jermaine, Chris.
- 학위논문주기
- - Ph.D. : The University of Texas at Austin, 2025.
- 초록/해제
- 요약Large Language Models (LLMs) have demonstrated remarkable capabilities in various tasks. However, generating responses that accurately meet the requirements of execution-based scenarios remains challenging. This challenge stems fundamentally from the pretraining objectives of LLMs, where next-token prediction is not inherently execution-aware, making customization in the post-training stage essential. This thesis explores a comprehensive approach to customizing LLMs for execution-driven contexts through three interconnected works.First, we investigate the potential of execution-derived feedback for enhancing LLM performance in specification-rich tasks. We propose a framework that synthesizes training data by combining natural language intents with specifications derived from execution. These specifications help to ground the model's understanding of task requirements, also addressing the ambiguity often present in natural language prompts. Our evaluations demonstrate substantial improvements in generating solutions that comply with execution specifications in data science notebook and tool automation.As synthetic data generation scales, we observe diminishing returns when fine-tuning a single model. To address this limitation, our second work explores how to make full use of abundant synthetic data by training multiple specialized model variants. We use influence functions to efficiently group synthetic data, promoting diversity in LLM outputs while maintaining quality. Experiments showed that our approach can diversify foundation model responses while maintaining high quality in the code generation domain and several tasks in the natural language understanding domain.Finally, we address the practical challenges of deploying these customized models with fLoRA (Fast Low-Rank Adaptation). fLoRA enables efficient, real-time serving of task-specific model variants through batched low-rank adaptation, making it computationally feasible to maintain and deploy multiple specialized models while ensuring that the most appropriate variant is selected for a given execution-based task. Together, these contributions form a holistic strategy for customizing LLMs in execution-driven scenarios, enhancing both their accuracy, diversity and efficiency across applications in code generation and tool automation.
- 언어주기
- English
- 일반주제명
- Computer engineering
- 일반주제명
- Computer science
- 일반주제명
- Information technology
- 키워드
- Synthetic data
- 키워드
- Tool automation
- 키워드
- Code generation
- 기타저자
- The University of Texas at Austin Computer Science
- 기본자료저록
- Dissertations Abstracts International. 87-06B.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008260311s2025 us eng d■001000017361247
■00520260311091517.5
■006m o d
■007cr|nu||||||||
■020 ▼a9798270232726
■040 ▼aMiAaPQD▼beng▼cMiAaPQD▼erda
■082 ▼a005
■1001 ▼aWen, Yeming▼eauthor.
■24510▼aCustomizing Large Language Models: From Synthetic Data Generation to Efficient Deployment ▼cYeming Wen
■260 ▼a[Sl]▼bThe University of Texas at Austin▼c2025
■264 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2025
■300 ▼a1 electronic resource (135 pages)
■336 ▼atext▼btxt▼2rdacontent
■337 ▼acomputer▼bc▼2rdamedia
■338 ▼aonline resource▼bcr▼2rdacarrier
■500 ▼aSource: Dissertations Abstracts International, Volume: 87-06, Section: B.
■500 ▼aAdvisors: Chaudhuri, Swarat Committee members: Durrett, Greg; Akella, Aditya; Jermaine, Chris.
■5021 ▼bPh.D.▼cThe University of Texas at Austin▼d2025.
■520 ▼aLarge Language Models (LLMs) have demonstrated remarkable capabilities in various tasks. However, generating responses that accurately meet the requirements of execution-based scenarios remains challenging. This challenge stems fundamentally from the pretraining objectives of LLMs, where next-token prediction is not inherently execution-aware, making customization in the post-training stage essential. This thesis explores a comprehensive approach to customizing LLMs for execution-driven contexts through three interconnected works.First, we investigate the potential of execution-derived feedback for enhancing LLM performance in specification-rich tasks. We propose a framework that synthesizes training data by combining natural language intents with specifications derived from execution. These specifications help to ground the model's understanding of task requirements, also addressing the ambiguity often present in natural language prompts. Our evaluations demonstrate substantial improvements in generating solutions that comply with execution specifications in data science notebook and tool automation.As synthetic data generation scales, we observe diminishing returns when fine-tuning a single model. To address this limitation, our second work explores how to make full use of abundant synthetic data by training multiple specialized model variants. We use influence functions to efficiently group synthetic data, promoting diversity in LLM outputs while maintaining quality. Experiments showed that our approach can diversify foundation model responses while maintaining high quality in the code generation domain and several tasks in the natural language understanding domain.Finally, we address the practical challenges of deploying these customized models with fLoRA (Fast Low-Rank Adaptation). fLoRA enables efficient, real-time serving of task-specific model variants through batched low-rank adaptation, making it computationally feasible to maintain and deploy multiple specialized models while ensuring that the most appropriate variant is selected for a given execution-based task. Together, these contributions form a holistic strategy for customizing LLMs in execution-driven scenarios, enhancing both their accuracy, diversity and efficiency across applications in code generation and tool automation.
■546 ▼aEnglish
■590 ▼aSchool code: 0227
■650 4▼aComputer engineering
■650 4▼aComputer science
■650 4▼aInformation technology
■653 ▼aLarge Language Models
■653 ▼aSynthetic data
■653 ▼aNatural language prompts
■653 ▼aTool automation
■653 ▼aCode generation
■7102 ▼aThe University of Texas at Austin▼bComputer Science.▼edegree granting institution.
■7201 ▼aChaudhuri, Swarat▼edegree supervisor.
■7730 ▼tDissertations Abstracts International▼g87-06B.
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17361247▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.
Preview
Export
ChatGPT Discussion
AI Recommended Related Books
detalle info
- Reserva
- No existe
- Mi carpeta
- Primera solicitud
- Non-Book Loan Application
- Nighttime Book Loan Application
Available after logging in.


