Leveraging Narrative to Generate Movie Script

Abstract

Generating a text based on a predefined guideline is an interesting but challenging problem. A series of studies have been carried out in recent years. In dialogue systems, researchers have explored driving a dialogue based on a plan, while in story generation, a storyline has also been proved to be useful. In this paper, we address a new task–generating movie scripts based on a predefined narrative. As an early exploration, we study this problem in a “retrieval-based” setting. We propose a model (ScriptWriter-CPre) to select the best response ( i.e. , next script line) among the candidates that fit the context ( i.e. , previous script lines) as well as the given narrative. Our model can keep track of what in the narrative has been said and what is to be said. Besides, it can also predict which part of the narrative should be paid more attention to when selecting the next line of script. In our study, we find the narrative plays a different role than the context. Therefore, different mechanisms are designed for deal with them. Due to the unavailability of data for this new application, we construct a new large-scale data collection GraphMovie from a movie website where end-users can upload their narratives freely when watching a movie. This new dataset is made available publicly to facilitate other studies in text generation under the guideline. Experimental results on the dataset show that our proposed approach based on narratives significantly outperforms the baselines that simply use the narrative as a kind of context.

Translating Text Synopses to Video Storyboards

“…Since it is hard to obtain original video storyboards from professional video makers, we decided to select keyframes from released movies to reconstruct a succinct storyboard that a human user can use as a shooting plan. In terms of the text inputs, previous works have collected aligned script [49], caption [32], Descriptive Video Service (DVS) [38], book [48], or synopsis [37] to movies. However, books cannot be well-aligned with adapted movies; DVS is hard to obtain and thus limited in scale; wiki plots are too coarse, while scripts and captions are too detailed to compose for most non-professional users.…”

Geochemical Characteristics and Environmental Implications of the Elements of Qingshan Group on the Zhougezhuang Section, Shandong Province, China

“…To date, these transformer-based models have achieved compelling success in various NLP tasks, such as movie script generation [57], personalized answer generation [8], and query suggestion [30]. Inspired by their remarkable performance, we utilize the transformer encoder and decoder for our text representation and generation, respectively.…”