서브메뉴
검색
Towards Intelligent Conversational Assistants: Enhancing Task-Oriented Dialogue Systems with Knowledge Integration
Towards Intelligent Conversational Assistants: Enhancing Task-Oriented Dialogue Systems with Knowledge Integration
상세정보
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20260202105529
- ISBN
- 9798263340841
- DDC
- 150
- 저자명
- Su, Ruolin.
- 서명/저자
- Towards Intelligent Conversational Assistants: Enhancing Task-Oriented Dialogue Systems with Knowledge Integration
- 발행사항
- [Sl] : Georgia Institute of Technology, 2025
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2025
- 형태사항
- 147 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 87-05, Section: B.
- 주기사항
- Advisor: Juang, Biing-Hwang.
- 학위논문주기
- Thesis (Ph.D.)--Georgia Institute of Technology, 2025.
- 초록/해제
- 요약This thesis explores the integration of knowledge into Task-Oriented Dialogue (TOD) systems within Intelligent Assistants (IA), enhancing their ability to understand user intent and execute structured tasks. While dialogue systems have evolved significantly with advancements in Natural Language Processing (NLP) and Large Language Models (LLMs), TOD systems still face challenges in knowledge integration, scalability, and generalization. This work proposes methodologies for incorporating domain-specific, dialogue-level, and cross-lingual knowledge into TOD systems to improve their adaptability and effectiveness across diverse applications.The study identifies key challenges in knowledge integration, emphasizing the need for specialized domain knowledge to improve Dialogue State Tracking (DST) and response generation. To address these challenges, this thesis introduces innovative methods such as structured slot-value transfer and schema-guided knowledge graphs, enhancing both accuracy and scalability in dialogue management. Specifically, we propose choice-fusion DST and schema graph-guided prompts for enhance state tracking and facilitate domain adaptation. Furthermore, it explores the integration of dialogue-level knowledge to improve context awareness, such as incorporating dialogue acts into slot-value prediction for enhanced comprehension and employing a soft mixture-of-experts approach to develop more efficient and scalable TOD systems.Additionally, the thesis presents a cross-lingual knowledge transfer mechanism to improve commonsense reasoning in low-resource languages, enhancing the multilingual adaptability of TOD systems. These contributions collectively advance the design of intelligent conversational agents by enabling systematic knowledge integration and improving the efficiency of dialogue systems in real-world applications. Overall, this research establishes a comprehensive framework for knowledge-enhanced TOD systems, addressing critical limitations in current approaches. By integrating structured knowledge across different levels, it enhances system performance in user intent recognition, dialogue state tracking, and response generation. The findings contribute to the broader field of conversational AI, providing scalable and adaptable solutions for intelligent assistants in diverse domains and multilingual settings. They hold great promise for refining knowledge integration techniques, expanding cross-lingual capabilities, and exploring further applications of LLMs in TOD systems.In summary, this work focuses on methodologies for integrating domain-specific, dialogue-level, and cross-lingual knowledge into dialogue systems, evaluating its impact across diverse contexts to enhance the design of more effective and intelligent conversational agents. By equipping TOD systems with mechanisms for dynamic knowledge incorporation, it fosters the development of more robust, scalable, and adaptable conversational agents. These advances contribute to improving user interaction, improving response accuracy, and broadening the applicability of TOD systems across various domains and languages.
- 일반주제명
- Success
- 일반주제명
- Communication
- 일반주제명
- Adaptation
- 일반주제명
- Design
- 일반주제명
- Multilingualism
- 일반주제명
- Large language models
- 일반주제명
- Knowledge representation
- 일반주제명
- Bilingual education
- 기본자료저록
- Dissertations Abstracts International. 87-05B.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008260126s2025 us c eng d■001000017360459
■00520260202105529
■006m o d
■007cr#unu||||||||
■020 ▼a9798263340841
■035 ▼a(MiAaPQ)AAI32309812
■035 ▼a(MiAaPQ)GeorgiaTech77838
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a150
■1001 ▼aSu, Ruolin.
■24510▼aTowards Intelligent Conversational Assistants: Enhancing Task-Oriented Dialogue Systems with Knowledge Integration
■260 ▼a[Sl]▼bGeorgia Institute of Technology▼c2025
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2025
■300 ▼a147 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 87-05, Section: B.
■500 ▼aAdvisor: Juang, Biing-Hwang.
■5021 ▼aThesis (Ph.D.)--Georgia Institute of Technology, 2025.
■520 ▼aThis thesis explores the integration of knowledge into Task-Oriented Dialogue (TOD) systems within Intelligent Assistants (IA), enhancing their ability to understand user intent and execute structured tasks. While dialogue systems have evolved significantly with advancements in Natural Language Processing (NLP) and Large Language Models (LLMs), TOD systems still face challenges in knowledge integration, scalability, and generalization. This work proposes methodologies for incorporating domain-specific, dialogue-level, and cross-lingual knowledge into TOD systems to improve their adaptability and effectiveness across diverse applications.The study identifies key challenges in knowledge integration, emphasizing the need for specialized domain knowledge to improve Dialogue State Tracking (DST) and response generation. To address these challenges, this thesis introduces innovative methods such as structured slot-value transfer and schema-guided knowledge graphs, enhancing both accuracy and scalability in dialogue management. Specifically, we propose choice-fusion DST and schema graph-guided prompts for enhance state tracking and facilitate domain adaptation. Furthermore, it explores the integration of dialogue-level knowledge to improve context awareness, such as incorporating dialogue acts into slot-value prediction for enhanced comprehension and employing a soft mixture-of-experts approach to develop more efficient and scalable TOD systems.Additionally, the thesis presents a cross-lingual knowledge transfer mechanism to improve commonsense reasoning in low-resource languages, enhancing the multilingual adaptability of TOD systems. These contributions collectively advance the design of intelligent conversational agents by enabling systematic knowledge integration and improving the efficiency of dialogue systems in real-world applications. Overall, this research establishes a comprehensive framework for knowledge-enhanced TOD systems, addressing critical limitations in current approaches. By integrating structured knowledge across different levels, it enhances system performance in user intent recognition, dialogue state tracking, and response generation. The findings contribute to the broader field of conversational AI, providing scalable and adaptable solutions for intelligent assistants in diverse domains and multilingual settings. They hold great promise for refining knowledge integration techniques, expanding cross-lingual capabilities, and exploring further applications of LLMs in TOD systems.In summary, this work focuses on methodologies for integrating domain-specific, dialogue-level, and cross-lingual knowledge into dialogue systems, evaluating its impact across diverse contexts to enhance the design of more effective and intelligent conversational agents. By equipping TOD systems with mechanisms for dynamic knowledge incorporation, it fosters the development of more robust, scalable, and adaptable conversational agents. These advances contribute to improving user interaction, improving response accuracy, and broadening the applicability of TOD systems across various domains and languages.
■590 ▼aSchool code: 0078.
■650 4▼aSuccess
■650 4▼aCommunication
■650 4▼aInteractive computer systems
■650 4▼aAdaptation
■650 4▼aDesign
■650 4▼aMultilingualism
■650 4▼aNatural language processing
■650 4▼aLarge language models
■650 4▼aKnowledge representation
■650 4▼aBilingual education
■690 ▼a0389
■690 ▼a0459
■690 ▼a0800
■690 ▼a0282
■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=T17360459▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.


