자료유형 | 학위논문 |
---|---|
서명/저자사항 | Neural Networks for Narrative Continuation. |
개인저자 | Roemmele, Melissa. |
단체저자명 | University of Southern California. Computer Science. |
발행사항 | [S.l.]: University of Southern California., 2018. |
발행사항 | Ann Arbor: ProQuest Dissertations & Theses, 2018. |
형태사항 | 171 p. |
기본자료 저록 | Dissertations Abstracts International 81-04B. Dissertation Abstract International |
ISBN | 9781687938145 |
학위논문주기 | Thesis (Ph.D.)--University of Southern California, 2018. |
일반주기 |
Source: Dissertations Abstracts International, Volume: 81-04, Section: B.
Advisor: Gordon, Andrew S. |
이용제한사항 | This item must not be sold to any third party vendors. |
요약 | The field of artificial intelligence has long envisioned using computers to automatically write stories. With the advent of machine learning, in particular neural networks, researchers have found new opportunities to automatically acquire narrative knowledge directly from text corpora. There is now interest in developing systems that interface with human-authored stories in order to dynamically predict 'what happens next' in a story. In this thesis, I apply a set of neural network approaches to this narrative continuation task. I examine the task within two frameworks. In the first (closed-choice prediction), the system is presented with a story and must choose the best continuing sentence from a set of provided candidates. In the second (free-text generation), there are no candidates given for the next sentence, and the system must generate a new continuation. I demonstrate some evaluation approaches and applications associated with each framework. I discuss the observed successes and challenges of these neural network techniques in order to motivate future work in this up-and-coming research domain. |
일반주제명 | Artificial intelligence. |
언어 | 영어 |
바로가기 |
: 이 자료의 원문은 한국교육학술정보원에서 제공합니다. |