Abstract:It is very difficult to generate stories in Chinese language. So far, there is no effective method to generate smooth articles. Here proposed a novel approach to improve the generation of Chinese stories in artificial intelligence, in order that it can effectively control the part-of-speech structure in sentence generation to imitate the writer’s writing style. The main proposal consists of three parts. First, the pre-processing of the sentence discards the input as the summary and the output as the text. It uses the format containing < SOS > < MOS > < EOS > for processing and the detailed method is defined in session 4. The second part is for vectorization. Traditional vectorization methods include Word2vec, Fasttext, LexVec and Glove; the different vectorization methods can help data semantic or grammatical understanding. Combining different vectorization methods improves the information of the input data. Therefore, this paper proposes the multi-channel word embedding and the details defined in session 5. The last part contains the optimization of the model architecture and how to control the process of sentence generation effectively. It also rewrites the Bert model proposed by Google to be the proposed model architecture. In addition, the Softmax function had been optimizing to reduce the search time during training and increase the training speed in the model. To make the model have better performance, the necessary training of the generative adversarial network was carried out, and the GAN architecture was revised for the data set, and the detail is defined in session 6. After the model is trained, to effectively control the structure of the generated sentence. This paper proposes a complete generation flowchart. In the process, based on the concept of FP-Tree, all sentences in the data set are built into a tree structure, and the part-of-speech structure of the next sentence is restricted through model generation combined with FP-Tree and the detail is defined in session 7. In addition, the experimental results show that our proposed method can effectively control the results of Chinese story generation and generate sentences with better performance and the detail is defined in session 8.

From Quick-draw to Story: A Story Generation System for Kids' Robot.

StoryDrawer A Child-AI Collaborative Drawing System to Support Children's Creative Visual Storytelling

TaleBrush: Sketching Stories with Generative Pretrained Language Models

SmartPaint: a Co-Creative Drawing System Based on Generative Adversarial Networks

Outline to Story: Fine-grained Controllable Story Generation from Cascaded Events

StoryDrawer A Co-Creative Agent Supporting Children's Storytelling through Collaborative Drawing

Knowledge-Enriched Visual Storytelling

KGGPT: Empowering Robots with OpenAI's ChatGPT and Knowledge Graph.

Retrieval, Selection and Writing: A Three-Stage Knowledge Grounded Storytelling Model

A Multi-Modal Story Generation Framework with AI-Driven Storyline Guidance

It is AI's Turn to Ask Humans a Question: Question-Answer Pair Generation for Children's Story Books

StoryCraft AI: Exploring Generative Approaches to Story Narration through AI

LOGEN: Few-shot Logical Knowledge-Conditioned Text Generation with Self-training.

Plan, Write, and Revise: an Interactive System for Open-Domain Story Generation

Latent Memory-Augmented Graph Transformer for Visual Storytelling

Chinese story generation of sentence format control based on multi-channel word embedding and novel data format

AutoStory: Generating Diverse Storytelling Images with Minimal Human Effort

Intelligent Grimm -- Open-ended Visual Storytelling via Latent Diffusion Models

Generating coherent comic with rich story using ChatGPT and Stable Diffusion

Crafting Narrative Closures: Zero-Shot Learning with SSM Mamba for Short Story Ending Generation

StoryDiffusion: How to Support UX Storyboarding With Generative-AI