Reproduce “Multi-Modal Generative Adversarial Network for Short Product Title Generation in Mobile E-Commerce” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Jian-Guo Zhang, Pengcheng Zou, Zhao Li, Yao Wan, Xiuming Pan, Yu Gong, Philip S. Yu

MM-GAN generates short product titles from long titles, images, and attribute tags. Its LSTM-based generator combines these inputs, while a discriminator trained to distinguish human-written from generated titles supplies a training reward. On a 5,000-item test set, it scores 69.53 on ROUGE-1 versus 66.19 for Agree-MTL; a seven-day Taobao test reports CTR and CVR improvements, limited to that deployment.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.