자료유형  

 학위논문

Control Number  

0017160678

International Standard Book Number  

9798381975888

Dewey Decimal Classification Number  

004

Main Entry-Personal Name  

Chawla, Kushal.

Publication, Distribution, etc. (Imprint  

[S.l.] : University of Southern California., 2024

Publication, Distribution, etc. (Imprint  

Ann Arbor : ProQuest Dissertations & Theses, 2024

Physical Description  

241 p.

General Note  

Source: Dissertations Abstracts International, Volume: 85-10, Section: B.

General Note  

Advisor: Lucas, Gale.

Dissertation Note  

Thesis (Ph.D.)--University of Southern California, 2024.

Summary, Etc.  

요약Social interactions often involve a mixture of motives. People seek to maximize their own interests without undermining the needs of others. Success in these interactions, referred to as mixed-motive interactions, demands a balance between self-serving and other-serving motives. For instance, in a typical negotiation, a player must balance maximizing their own goals with the goals of their partner so as to come to an agreement. If the player asks for too much, this can push the partner to walk away without an agreement, hence hurting the outcomes for all the parties involved. Such interactions are highly prevalent in everyday life, from deciding who performs household chores to customer support and high-stakes business deals. Consequently, automated systems capable of comprehending and participating in these strategic environments with human players find broad downstream applications. This includes advancing conversational assistants and the development of tools that make everyday social interactions more effective and efficient (e.g., by acting as a content moderator or a coach). Additionally, these systems hold a huge potential to transform pedagogical practices by dramatically reducing costs and scaling up social skills training.Most efforts for automation focus on agent-agent interactions, where thousands of offers are exchanged between the players. These interactions are fundamentally different from human-agent conversations, which are much shorter and naturally involve human subjectivity, which in fact, has been a subject matter of research for decades across several disciplines, including Psychology, Affective Computing, and Economics. Hence, in order to simplify the design, most efforts in human-agent negotiations involve restrictive menu-driven communication interfaces that are based on button clicks and structured APIs for interaction between the human and the machine. This concreteness reduces the design complexity, but it comes at a cost -- such interfaces hinder the study and incorporation of several aspects of real-world negotiations, such as complex strategies and emotion expression. Going beyond such constrained designs, it is desirable to incorporate more realistic modes of communication, such as natural language, for their utility in better serving the downstream applications -- our work aims to fill this gap.In this dissertation, we present our foundational work for enabling mixed-motive human-machine dialogue, with a focus on bilateral chat-based negotiation interactions. We discuss our progress in three key areas: 1) The design of a novel task and dataset of grounded human-human negotiations that fueled our investigations into the role of emotion expression and linguistic strategies, 2) Techniques for dialogue systems capable of engaging in mixed-motive interactions by learning to strike a balance between self and partner interests, and 3) Defining a research space encompassing such strategic dialogue interactions to promote a research community for dedicated efforts and discussion in this area.

Subject Added Entry-Topical Term  

Computer science.

Subject Added Entry-Topical Term  

Psychology.

Subject Added Entry-Topical Term  

Personality psychology.

Index Term-Uncontrolled  

Chatbots

Index Term-Uncontrolled  

Dialogue systems

Index Term-Uncontrolled  

Emotion expression

Index Term-Uncontrolled  

Natural language processing

Index Term-Uncontrolled  

Negotiations

Added Entry-Corporate Name  

University of Southern California Computer Science

Host Item Entry  

Dissertations Abstracts International. 85-10B.

Electronic Location and Access  

로그인을 한후 보실 수 있는 자료입니다.

Control Number  

joongbu:658055