Ex3: Automatic Novel Writing by Extracting, Excelsior and Expanding
This addresses the problem of poor logical coherence and appeal in AI-generated novels for applications in creative writing and entertainment, representing an incremental improvement over existing hierarchical approaches.
The paper tackles the challenge of generating long-term texts like novels with AI by proposing Ex3, a method that extracts structure information, fine-tunes LLMs, and uses tree-like expansion, resulting in higher-quality novels compared to previous methods.
Generating long-term texts such as novels using artificial intelligence has always been a challenge. A common approach is to use large language models (LLMs) to construct a hierarchical framework that first plans and then writes. Despite the fact that the generated novels reach a sufficient length, they exhibit poor logical coherence and appeal in their plots and deficiencies in character and event depiction, ultimately compromising the overall narrative quality. In this paper, we propose a method named Extracting Excelsior and Expanding. Ex3 initially extracts structure information from raw novel data. By combining this structure information with the novel data, an instruction-following dataset is meticulously crafted. This dataset is then utilized to fine-tune the LLM, aiming for excelsior generation performance. In the final stage, a tree-like expansion method is deployed to facilitate the generation of arbitrarily long novels. Evaluation against previous methods showcases Ex3's ability to produce higher-quality long-form novels.