A Character-Centric Creative Story Generation via Imagination

Creative story generation with diverse and detailed story elements is a long-standing goal for large language models. While existing methodologies generate long and coherent stories, they fall significantly short of human capabilities in terms of diversity and character detail. To address this, we introduce a novel story generation framework called CCI (Character-centric Creative story generation via Imagination). CCI features two innovative modules for creative story generation: IG (Image-Guided Imagination) and MW (Multi-Writer model). In the IG module, we utilize DALL-E 3 to create visual representations of key story elements. The IG generates more novel and concrete characters, backgrounds, and main plots than text-only methods. The MW module uses these story elements created by IG to generate multiple description candidates for the protagonist and select the best one. This method incorporates vivid and rich character descriptions into the story. We compared the stories generated by CCI and baseline models through human evaluation and statistical analysis. The results showed significant improvements in the creativity. Furthermore, by enabling interactive multi-modal story generation with users, we have opened up possibilities for human-LLM integration in cultural development.

LongStory: Coherent, Complete and Length Controlled Long story Generation

A human author can write any length of story without losing coherence. Also, they always bring the story to a proper ending, an ability that current language models lack. In this work, we present the LongStory for coherent, complete, and length-controlled long story generation. LongStory introduces two novel methodologies: (1) the long and short-term contexts weight calibrator (CWC) and (2) long story structural positions (LSP). The CWC adjusts weights for long-term context Memory and short-term context Cheating, acknowledging their distinct roles. The LSP employs discourse tokens to convey the structural positions of a long story. Trained on three datasets with varied average story lengths, LongStory outperforms other baselines, including the strong story generator Plotmachine, in coherence, completeness, relevance, and repetitiveness. We also perform zero-shot tests on each dataset to assess the model's ability to predict outcomes beyond its training data and validate our methodology by comparing its performance with variants of our model.