Controllable story generation is a challenging task in the field of natural language processing (NLP), which has attracted increasing research interest in recent years. While most existing works generate a whole story conditioned on appointed keywords or emotions, they often ignore the psychological changes of the protagonist. To address this limitation, we propose a Psychology-guided Controllable Story Generation System (PICS) that incorporates global psychological state chains to help create more controllable and well-planned stories. Our approach models both local and global psychological state changes of the protagonist as the story progresses. Specifically, we employ psychological state trackers to memorize the protagonist's local psychological states and capture their inner temporal relationships. Additionally, we use psychological state planners to gain the protagonist's global psychological states for story planning. Finally, a psychology controller is designed to integrate both local and global psychological states into the story context representation for composing psychology-guided stories. Compared to previous methods that incorporate sentiment or emotion for story generation, our approach focuses on modeling the protagonist's need and emotion chains as central components of storytelling. We evaluate PICS through automatic and manual evaluations and show that it outperforms baselines in writing stories with more consistent psychological changes. Furthermore, our task definition involves generating a story about a given leading context while adhering to the protagonist's name and desired psychological state chains. We provide an example of our task using Table 1 where each event is controlled by given psychological state chains. In summary, our approach introduces global psychological state chains inspired by psychology theories to generate more controllable and well-planned stories. By modeling both local and global psychological states of protagonists, PICS outperforms previous methods that focus solely on sentiment or emotion for story generation.
SummaryPICS is a computer program that helps create stories with characters who have feelings. It does this by keeping track of how the character's emotions change throughout the story. PICS makes sure that the story makes sense and follows a plan. It does better than other programs because it focuses on psychology, not just emotions.
Definitions
- Controllable: able to be controlled or managed
- Story generation: creating new stories
- NLP: Natural Language Processing, which is using computers to understand and use human language
- Psychological changes: changes in a person's thoughts or feelings
- Protagonist: main character in a story
- Global psychological state chains: patterns of emotions and thoughts that happen over time for a character
- Local psychological state changes: small changes in a character's emotions and thoughts as the story progresses
- Psychology controller: part of PICS that puts together all the information about a character's emotions and thoughts to create a good story
- Sentiment: feeling or emotion
- Adhering: following or sticking to something
- Leading context: information at the beginning of a story that sets up what will happen next
Exploring Psychology-Guided Controllable Story Generation System (PICS)
Natural language processing (NLP) has become a popular research area in recent years, and one of the most challenging tasks within this field is controllable story generation. Existing works have focused on generating stories based on appointed keywords or emotions, but they often overlook the psychological changes of protagonists. To address this limitation, researchers have proposed a new approach called Psychology-guided Controllable Story Generation System (PICS). This system incorporates global psychological state chains to help create more controllable and well-planned stories.
Modeling Local and Global Psychological States
PICS models both local and global psychological states of protagonists as the story progresses. The system employs psychological state trackers to memorize the protagonist's local psychological states and capture their inner temporal relationships. Additionally, it uses psychological state planners to gain insight into the protagonist's global psychological states for story planning purposes. Finally, a psychology controller is designed to integrate both local and global psychological states into the story context representation for composing psychology-guided stories.
Task Definition
The task definition involves generating a story about a given leading context while adhering to the protagonist's name and desired psychological state chains. An example of this task can be seen in Table 1 where each event is controlled by given psychological state chains.
Evaluation Results
PICS was evaluated through automatic and manual evaluations which showed that it outperforms baselines in writing stories with more consistent psychological changes. Furthermore, compared to previous methods that incorporate sentiment or emotion for story generation, our approach focuses on modeling the protagonist's need and emotion chains as central components of storytelling which helps generate more controllable and well-planned stories.
Conclusion
In conclusion, PICS introduces global psychological state chains inspired by psychology theories to generate more controllable and well-planned stories than existing methods that focus solely on sentiment or emotion for story generation do. By modeling both local and global psychological states of protagonists, PICS has been shown to outperform these previous methods in terms of creating consistent storylines with believable characters who experience realistic emotional arcs throughout their journey
Controllable story generation is a challenging task in the field of NLP, which has attracted increasing research interest in recent years. However, most existing works generate a whole story conditioned on the appointed keywords or emotions, ignoring the psychological changes of the protagonist. Inspired by psychology theories, we introduce global psychological state chains, which include the needs and emotions of the protagonists, to help a story generation system create more controllable and well-planned stories. In this paper, we propose a Psychology-guIded Controllable Story Generation System (PICS) to generate stories that adhere to the given leading context and desired psychological state chains for the protagonist. Specifically, psychological state trackers are employed to memorize the protagonist's local psychological states to capture their inner temporal relationships. In addition, psychological state planners are adopted to gain the protagonist's global psychological states for story planning. Eventually, a psychology controller is designed to integrate the local and global psychological states into the story context representation for composing psychology-guided stories. Automatic and manual evaluations demonstrate that PICS outperforms baselines, and each part of PICS shows effectiveness for writing stories with more consistent psychological changes.
