BIPro: Zero-shot Chinese Poem Generation via Block Inverse Prompting Constrained Generation Framework

TL;DR

This paper introduces BIPro, a novel zero-shot Chinese poem generation framework that uses block inverse prompting to significantly enhance the quality of constrained poem generation with less powerful models, outperforming state-of-the-art systems.

Contribution

The paper proposes the BIPro framework with two block inverse prompting methods, revise and rewrite, to improve zero-shot constrained poem generation without additional training.

Findings

BIPro outperforms GPT-4 and domain-specific systems in human evaluations.

Poems generated by BIPro are closer to human literary quality.

The approach narrows the gap between AI and human poetry in constrained tasks.

Abstract

Recently, generative pre-trained models have made significant strides, particularly highlighted by the release of ChatGPT and GPT-4, which exhibit superior cross-domain capabilities. However, these models still face challenges on constrained writing tasks like poem generation under open-domain titles. In response to this challenge, we introduce Block Inverse Prompting (BIPro) constrained generation framework. BIPro leverages two block inverse prompting methods, revise and rewrite, that mimic the process of human text writing using block generative models. It significantly improves the zero-shot generation quality on the formidable constrained generation task of open-domain traditional-form Chinese poem generation. Based on a less powerful block generative model GLM-10B-Chinese, poems composed via BIPro without priming or additional training outperform both most advanced direct generative systems like GPT-4 or GLM-4 and best domain-specific systems such as Yusheng, Shisanbai, or Baidu Poetry Helper in human evaluation by proficient poets. Finally, BIPro considerably narrows the gap between AI-generated works and short-listed human literary arts in another human evaluation, unveiling the promising potential of block generative models in improving the quality of constrained generation.